AIC_1978 Provided by University Library of Tübingen Transcribed with Tesseract

This work is protected by copyright or related property rights but available in Open Access. The right of use and especially the right to reproduction is only granted within the legal limits of copyright law or due to the consent of the copyright holder.

 Annales internationales de criminologie, 17 ANNALES INTERNATIONALES DE CRIMINOLOGIE INTERNATIONAL ANNALS OF CRIMINOLOGY INTERNACIONALES DE CRIMINOLOGIA ANALES Année 1978 (volume 17 - n° | et 2) TSH|GE 1}l SU U] Ge / /< —0 es E >aW E —‘ ; € 3 Ÿ “ (: PaR” , H|Ù - …… "À aÀ ; PC\IŸ ; … [;!\ 4 — … ps | .W N—. > W.….oh t> L d @ e 7; 04’ .-\. Cs .J\ JAN m e p d AVERTISSEMENT DU COMITE DE REDACTION Le Conseil de direction de la Société internationale de criminologie avait chargé le Centre international de criminologie comparée — que dirige Denis Szabo — de réaliser un volume spécial des Annales. Le manuscrit remis par le C.I.C.C. au comité de rédaction dépas- sait considérablement la taille d’un volume annuel. À vrai dire, il correspondait plutôt à deux volumes. Le comité de rédaction n’a pas voulu opérer des coupures — qui auraient dû être considérables — dans cet ensemble qui forme un tout. Il a résolu d’en répartir la matière en deux volumes la première livraison (volume 16) comprenait les deux premières parties ; le second (volume 17) réunit les trois dernières parties du manuscrit mis au point parle C.I.C.C. sous la direction de Denis Szabo. Pour bien marquer l’unité entre ces deux livraisons, la table des matières est intégralement incluse dans le présent volume. — 5— TABLE DES MATIERES TABLE OF CONTENTS VOLUME 16 SzAaBo (D.), Introduction . . ». …. e + — e + 6. 8 < < — . 6 — © © » e 6 » % 6 % © 6 © % % ® # ® I. — COMPARATIVE CRIMINOLOGY : GLOBAL PERSPECTIVES FLYNN (S.E.), Issues and Priorities in International Cooperation in the Field of Criminology 13 GURR (T.R.), Crime Trends in Modern Democracies Since 1945.. 41 CLIFFORD (W.), Culture and Crime . « e » + ». » » e « » « © « © » 6 6 % » 6 % # 6 # & à7 ARCHER (D.) & GARTNER (R.), Homicide in 110 Nations : The Development of the Comparative Crime Data File 109 I. — HISTORICAL AND MACRO-CRIMINOLOGICAL STUDIES Rico (J.), La violence en Amérique latine 143 MIDDENDORFF (W.), Bank Robbers and Their Victims ......…. 179 BRILLON (Y.), L’acculturation juridique en Afrique Noire et ses incidences sur l’administration de la justice criminelle . e ». « © 19 Rocx (P.), Law, Order and Power in Late Seventeenth and early Eighteenth Century England » e e » » e € « % — % © @ ® © « se à © % % 6 6 ® @ 6 # ° 233 SVIRBUL (V.K.), Potentialities of Social Process Regulations of Crime Prevention . « « * © 6 e <. » e © + e % % % * ® ® ® * e © ® © % %* ® % 6 » @ » # » 267 —7— VOLUME 17 II — COMPARATIVE ANALYSIS OF THE CRIMINAL JUSTICE SYSTEMS . e … © e e + »s 6 © e e * s * e 6 © © ® © © % =% % % © % + % © © e 6 ® ® « 9 SZABO (D.), Foreword ANTTILA (L.), Problems of Comparative Research on the Use of Discretion in Criminal Justice ° + e »s © %* + + » % +e » 0 + © e e « e » © 6 % ® 6 e BaiLey (W.C.), The Deterrent Effect of Arrest and Conviction on Crimes Rates in Indian States and Territories 21 TIEDEMANN (K.), Subjects, Aims and Methods of European Criminological Research in the Field of Economic and Business Crime e e « » e « … e e « » . « * © » © © —… e « . 6 — « « 8 » # + » » % % 6 % ® ® 51 VERIN (J.) & STRASBURG (P.), International Cooperation in Action G9 Research A French-American Example KERNER (H.J.), Fear of Crime and Attitudes Towards Crime Comparative Criminological Reflections e …. © » e © » » e % =s © #% e © »% 83 IV. — MICRO-CRIMINOLOGY GIBBENS (T.N.C.) & SooTHILL (K.L.), The Rape Problem tO JEVASSEUR (G.) & CHAMPENOIS-MARNIER (M.P.), L'application d’un sursis partiel par les tribunaux correctionnels français. . 17 VoporivEc (K.), Restitution to Victims of Criminal Offences . « e e « e s. » % % © * » 6 © 6 % » ® % » e « e 147 in Slovenia SFR of Yugoslavia 167 MARxX (G.T.) & MORTON (M.), Police and Minorities in England. . V. — METHODOLOGY WiIzKINs (L.T.), Decisions Theory and Comparative Criminology. [O| GraBosky (P.N.), The Comparative Study of Penal Severity : Some Methodological Considerations 209 ROBERT (Ph.), Consensus, Conflit et attitudes envers le crime et la loi pénale e e e « e à © 6 © % » © » s « < » © …s < © © … * © © 6 » * © * % e © # « © # —— —— 233 —8— | FOREWORD Comparative research in criminology is beginning to take form sign of a discipline’s vitality. From the time F. & S. Glueck (1964) appealed to the Hague Conference in 1960 for comparative criminology, up to the creation of the International Center for Comparative Crimi- nology in 1969, researchers began to explorc the various “ macro ” as well as « micro » avenues in criminological research. During the 17th International Course in Criminology, held in Montreal, in 1967, did M. Wolfgang not predict the burgeoning of comparative criminology when he traced the viable future of our discipline to the year 2000 ? (1968). In 1976, at the séminar organized by the I.C.C.C. in Santa Mar- gherita, Italy, about forty researchers examined the thematic and me- thodological problems and priorities entailed in comparative research. Finally, in 1977, the Scientific Commitee of the International Society of Criminologv entrusted me with the task of preparing the present volume. Given the abundance and quality of the articles sent in answer to my appeal, this book will be published in two volumes, the second to come out few months later. Ît was in the International Annals of Crimiùology (1973), that I recently explained my concept of comparative criminology ; the present contribution is an integral part of the research programme I outlined in that article. The papers presented and the discussions that took place at the Santa Margherita seminar, and summarized by Edith E. Flynn, give an idea of the methodological and theoretical attractions and dangers inherent in the comparative approach. The texts reflect the direction of the priorities tested by criminologists, all of whom, however, were from the industrialized European and North American nations. The problems of developing countries, on the other hand, are reflected in the contribution of W. Clifford, who through his long association with the United Nations, is no stranger to these problems. The chronological method used by T.R. Gurr and P. Rock and the diachronic method by D. Archer & R. Gartner, are tested respectively. Among the historical and macro-criminological studies, third world are amply described by J.M. Rico for Latin America, by Y. Brillon for Africa, and by W.C. Bailey for India : W. Middendorff, H.J. Kerner and V.K. Zvirbul describe one particular phenomenon. Comparative researchers are deeply concerned with the system of criminal justice : the example of the work done in this domain warns — 9 — against premature enthusiasm about the ease of comparing the elements of the system (legislation, courts, prisons, the police, etc.) even in culturally homogeneous regions such as Scandinavia, as described by I. Anttila, or regarding specific offenses such as white collar crime, described by K. Tiedemann. The transference of socio-legal reforms presents enormous problems can the same system be applied in Paris and London that is used in New York ? Such is the question asked by J. Vérin & P. Strasburg. Micro-criminology and methodology are barely touched on in this review of the works in progress : its qua- lity, however, holds promise that the monographs and thinking in these fields will undoubtedly produce a richer harvest in the next decade. The subtle distinctions in T.C.N. Gibbens’ description of phenomena such as rape, the analyses of the functioning of the mechanisms of justice administration, such as the use of partial stay of proceedings in France, by G. Levasseur & M.P. Champenois-Marnier, and the compensation of victims in Yugoslavia, by K. Vodopivec, are neces- sary pre-requisites for the comparative approach. The ambiguities in the role of the police facing problems of law enforcement among members of ethnic minorities in England are carefully outlined by G.T. Marx & M. Morton. With regard to methodology, it is the very substance of the researcher’s concern with regard to the scientific experience acquired : intensification of the latter will inevitably stinulate the progress of the former. The works of L.T. Wilkins, P.N. Grabosky and Ph. Robert are exemplary in this respect. I wish to thank all those authors who graciously agreed to contri- bute to the present volume. Although they are responsible for their own works, the undersigned alone is responsible for the choice of contri- butions and their distribution in diverse chapters. Mesdames Jacqueline DePlaen-Beaudry, in charge of the I.C.C.C.’s documentation centre, and Andrée Hamam, administrative assistant, have ably assisted me in the preparation of this manuscript for publication. Finally, I thank the Scientific Committee of our Society for having generously devoted two volumes of The Annals to these texts, and also its Deputy Secretary, Mr. Ph. Robert, for his efficient help with the final arrangements. GLUECK (E. & S.), Ventures in Criminology, Cambridge, Mass. : Harvard University Press, (1964). SZABO (D.), « Criminologie comparée signification et tâches », Annales Internationales de Criminologie, Vol. 12, N°1 et 2 (1973). WOLFGANG (MEÆE.), « The Viable Future of Criminology » in : Crimi- pages 109-135. nologie en action, Montréal, Presses de l’Université de Montréal, (1968), — 10 — II.- Comparative analysis of the criminal justice systems PROBLEMS OF COMPARATIVE RESEARCH ON THE USE OF DISCRETION IN CRIMINAL JUSTICE INKERI ANTTILA Research Institute of Legal Policy Helsinki - Finland lntetnatîonal Research on the Criminal Justice System Criminologists, in common with researchers in general, have strived for globality. This, of course, has not been easy. The variance in legis- lation has sometimes seemed to place insurmountable barriers to inter- national comparisons of crime control systems. Variations in criminal law, however, have come to be regarded as something more than just a drawback. Professor Nils Christie expressed this very well in stating that keeping to the « legal » concept of offences gives researchers new perspectives. Researchers can study why systems differ, instead of just trying to eliminate the differences, or limiting the field of study to one where differences in legislation possibly will not mean sc much. Comparisons between different systems may provide researchers with new horizons. This also holds true for research on the use of discretion. Problems in Comparafive Research What points of view should be considered when planning compà- rative research for example on 4the use of discretion ? First, a researcher should clarify his ferminology. The terms used in criminal law can be stabilized when used in one country ; but the more distant the country or the stranger the culture, the greater the risk of misunderstanding. For example, the term « preventive detention » can be understood in two ways. Apparently it is often the same thing as « detention before trial ». But it has least at times been used in Europe to refer to incarceration without a time limit, a « preventive measure for dangerous criminals ». One must, however, never be satisfied with merely comparing the content of written legislation, without finding out how the law is applied: This observation is so self-evident that I had my doubts about mentioning it : however, we are familiar with the requests from researchers abroad, asking for information on the contents of written law - possibly only one statute - at the same time indicating that they intend to draw far-reaching conclusions on the functioning of the system. The next step is to test whether the legislation is linked to other statutes. Are there possibly parallel statutes that may complete on the application level ? As an cxample of this we could take the laws on non-prosecution in Scandinavia the laws may contain a general non-prosecution clause, dealing with minor offences committed by anybody, regardless of age ; but there may also be a special non- prosecution clause dealing with young offenders. The non-prose- cution institution as a whole must be taken into consideration. But comparisons must not be limited to this. Every system not only contains procedures that go under the same name, but also mechanisms which can be regarded as competing, with a similar final result. Let us again think of non-prosecution : the prosecutor may often have to choose whether or not he will bring the case to court : if he does prosecute, the court also has several alternatives open. In those concrete cases where non-prosecution would be possible, it is generally also possible for the court to dismiss the case or give a conditional sentence. The prosecutor can predict with a great deal of accuracy what the result would be if he prosecutes - the offender would receive one of the two alternative sanctions mentioned, and he would thus anyway retain his liberty. In international comparisons there is reason to consider similar alternatives. For example, it is by no means certain that a country where a great deal of non-prosecution is applied is a country with a lenient criminal policy - it may be that this country does not have any alternatives, such as dismissal in the court, available ! It is also possible that non-prosecution itself leads to a harsher result than some others would. This could be 1llustrated by citing the situation in Finland thirty years ago. If the offender was between 15 and 17 years old, non-prosecution was possible for petty offenses. The purpose of this was to pass the case to the child welfare authorities : the prosecutor was supposed to notify the local child welfare board, which would then start from the beginning. During those « years of severe sanctions », it was possible that a fifteen year old petty thief could be sent to a juvenile institution for quite a long time. Had the prosecutor chosen to prosecute, the sentence in court would almost certainly have been given conditionally. In comparing criminal justice legislation in different countries, we look above all at the system of sanctions. However, the system is also regulated by ofher criminal law norms. For example the prosecutor’s discretion is in some offences limited by the provision that only the victim has a finai say upon whether charges will be brought or not. If the law is changed so that the right to decide whether charges are brought is given to the prosecutor, this shifting of power also gives him a new and possibly wider scope of discretion. Such a situation occured a few years back in Finland in regard to some offences against a minor, when the prosecutor was given the right to bring charges in cases where this was previously allowed only to the guardian or the minor himself. to One could go even further and pose the following question what degree is the use of discretion within the criminal justice system connected with the exfent and intensity of criminalization ? At least it is clear that decriminalization, a subject of much debate in recent years, is an alternative to depenalization measures - if an act is no longer punishable, official discretion is no longer needed. At times cases may also drop out of the sphere of discretion by being shifted outside criminal policy or the entire official control policy system. These are mostly very minor cases. For example in Finland an individual whose reputation has been assailed by a newspaper can bring a charge in the ordinary way, or use the alternative of requesting that the press establishment’s own honour court give the newspaper a war- ning. There are similar « disciplinary systems » in many other pro- fessions. When analysing the system, one must also consider self-regulatory mechanisms, which operate inside institutions, inside the whole criminal justice system, and in the relationship between the criminal policy system and other related systems. If, for example, the criminal law of a country has an excessively severe minimum sanction for « aggra- vated theft », courts tend to look for some ways that would enable them to apply the statute concerning « simple theft ». If the punishment for drunken driving is made more severe, the minor cases seem to disappear from the scene apparently the police are loth notify the prosecutor of minor cases, when in their opinion the sanction is exceptionally severe. The same principle works on many different levels. If the prisons are overcrowded, one looks for and usually also finds ways of reducing the number of prisoners, either through legislation or the enforcement of existing legislation. If the police are provided with good detention facilities, they are more likely to arrest a greater number of offenders than in a shortage of adequate facilities. I have so far listed quite a number of difficulties in comparative research. Even so, I think we should not stop here : 1f one wishes to compare systems, should not one have a wider knowledge of the legal culture of the countries being compared ? One knows or at least assu- mes, for example, that bribery of officials is in some countries, but not in others, unofficially accepted as part of the culture. This, of course, — 15 — must have an immediate effect on the use and actual extent of discre- tion within the criminal justice system. — It is another matter entirely that here we are involved in a sector of research on which it is almost impossible to gather data How to Lessen Difficulties in Making Comparisons Aîter all of these reservations it could very well be asked whether there are no shortcuts open to the researcher One way might be the following method of planning research instead of studying a certain institution, and taking into account the above reservations, one could make the level of research more abstract, one could study certain principles and ideologies, the law in b00ks and the law in action brought up to the level of principles. In this way, one could examine, for example, the following goals — The use of discretion connected with /he normal application .of law, especially in sentencing. This means, for example, the use of discretion over whether the act fulfils the criteria for an offence ; whether scales based on special mitigating or aggravating circumstances should be applied, and which sanction should be selected from the available scale. in manv cases the role of the court becomes similar to that of the legislator as the norms require a lot of interpretation — The use of discretion for the purpose of moderation. This is needed, for example, when a statute is poorly adaptable to a special case. At times a « back door » is specificallv left open in legislation for such cases ; for example, in Finland around 25 years ago, when the punishment for illegal abortion was a rather heavy one the court was given the right in exceptional cases to withhold punish- ment for the woman who has undergone an illegal abortion. À parallel is offered by those cases where it is felt that the statute has fallen into disuse, even though it has remamed on the law-books (so-called desue- tudo). — Discretion is used very commonly for rationalization. Even if it is not specifically allowed by law, it could be merely normal practice in the country to toss minor cases into the wastepapetr basket ! À times statute may be drawn up after such discretion has actually been used in practice for years. This is what happened in Finland when some years ago it was specifically decreed that the police had the right to refrain from reporting a minor offence (especially in traffic) to his superior, andneed only give a warning to the offender — Apparently one system that is partly for rationalization purposes and partly to moderate the force of oversevere sanctions is the « plea- bargaining » system familiar in American criminal procedure — 16 — — Discretion used for individualized sanctioning. Such a goal was very popular in Scandinavia, for example, a couple of decades ago in the discussion on reforming the crime control system. One could scar- cely conceive of a choice of individualized sanctions without a wide provision for discretion, no matter if the case was dealt with by a court or by some special administrative body. À wide measure of discretion was also naturally given as part of the procedure during the different stages of sanction enforcement, since release from an institution was dependent on how soon the prisoner was « rehabilited ». Actually, the goal of individualized sanctions demanded the use of discretion all along the line. The Role of Latent Goals In analysing, one must try to find out both manifest and latent goals influencing decision-making. If, for example, the system in one country gives the police discretion to decide whether to report a drunk lying in the street, and therefore have the drunk fined, the policeman can officially call the fine an appropriate and effective « individual warning ». In reality, he may clearly be aware of the fact that he wants most of all to «clean up the street» by carting off drunks to jail. Another example would be a decision model from Finnish criminal policy. Let us suppose that a murderer has been found to be criminally irresponsible and therefore is not punished. Since he is diagnosed to be suflering from, say, schizophrenia, he will be shut up in a mental institution against his will. According to law, his release must take place solely on the basis of two criteria : his need for care and his dangerousness, both of which in Finland are left to the discretion of the medical authorities. However, in practice the general sense of justice will also carry weight. The offender is not released as quickly as could be done on the basis of his own impro- vement and the estimation of his dangerousness in practice he will have to spend possibly five years instead of perhaps six months in an institution. Another case where the use of discretion also serves to fulfil latent goals is in the decision to detain before trial by the courts. It is known that detention before trial, which is meant by law only to serve criminal procedural needs, is in practice also used as a sanction, for example, for the unexpressed purpose that one wishes a speedy counteraction in dealing with an offence. There are also many examples of how the goals may change over the years. When the Law on Young Offenders was passed in Finland 35 years ago, there was a common belief in the efficiency OË treatment-oriented sanctions. and therefore, using as a model the — 17 — system prevailing in many other countries, a special sanction, « juvenile imprisonment », was included in the law. À young offender sentenced to imprisonment for at least 6 months could be sent by a decision of a special body to a juvenile prison, and if this occured, his term was automatically extended by 1-2 years. On the other hand immates of a juvenile prison could be released on parole earlier than inmates of normal prisons. At first, the law was followed quite automatically and stringently ; during the 1950’s the law was changed so that the exten- sion was made discretionary ; and at the end of the 1960°s the extensions were used only for the shorter terms. What is most surprising, however 1s the trend in the 1970°s Juvenile imprisonment is used only for those who are seen as needing release on parole at an early stage ! ÀA foreign researcher who does not know of this change would get completely lost, and think that young offenders sent to juvenile prison are considered to be in need of care and training there ! Much research work done in recent years has actually been connected with the use of unofficial discretion and with latent goals. The research has concentrated on grounds for the use of discretion that are « outside the system ». Researchers have dealt, for example, with the problem of discrimination between races or social classes, or the possible inequality between the sexes ; they have investigated whe- ther police are most apt to arrest younger people or adults ; they have discussed the question of whether white-collar offenders or ordinary thieves are sooner released on parole. One could analyse almost without end the actual function of the system in many different respects and on many different levels. What I have tried to do is to show that international comparisons may actually require comparisons made on a more general level. Is Discretion to be Encouraged ? It is not a habit of the conventional researcher to present recommendations to decision-makers. However, when he wants to, the researcher can certainly « go over to a criminal politician’s role » and present his own opinion. The final criminal policy goal should be to minimize criminality and the resultant social cost (property damages, suffering, injury and harm) so that costs to the victim and to society as well as to the offender are taken into consideration. Also, the general sense of justice and legal security are among the decision-making criteria. In considering his recommendations, the researcher sould try to answer for example the following questions — is the wide use of official discretion advisable ? — 18 — is it advisable to differentiate between discretion used in a mitigating direction from discretion used in a more severe direction ? are different officials in a different position, and is the court in à special position ? is there a difference in the situation before and after the trial ? how should one regard systems closely allied to the criminal law system, for example, disciplinary punishments in school or in prisons, or disciplinary sanctions for state officials ? what about compulsory treatment on the basis of an offence, either an individual act or habitual offences (for example va- grancy) ? In summary, therefore, I would like to point out that in carrying out research on the use of discretion the researcher should be aware of the function of the entire system in the country in question ; and he should attempt to study the manifest goals aimed at in passing and in applying legislation as well as unconver the latent goals. — 19 — THE DETERRENT EFFECT OF ARREST AND CONVICTION ON CRIME RATES IN INDIAN STATES AND TERRITORIES WILLIAM C. BAILEY Department of Sociology The Cleveland State University Cleveland, Ohio - U.S.A The research reported here consists of a multivariate analysis of the deterrent effect of arrest and conviction on crime rates for Indian states and federal territories for the period 1967 to 1973. The deterrent effect of arrest and conviction are examined separately for six serious offenses with soc1odemographœ factors introduced into the analysis as control variables. The last decade has witnessed a number of empirical investi- gations of the deterrent effect of arrest, conviction, and imprisonment on major felony rates for states, counties, and cities in the United States. As hypothesized, and consistent with the deterrence doctrine these investigations have rather typically revealed a substantial inverse relationship between the certainty of punishment and offense rates Contrary to the deterrence hypothesis, however, only a slight negative, and in some cases a positive, relationship has been found between the severity of legal sanctions and offense rates. These findings have brought many American Criminologists- to conclude that no longer can the threat and application of legal sanctions be dismissed as an important determinant of crime (Tittle and Rowe, 1974 ; Tittle and Logan, 1973 Sjoquist, 1973 Orsagh, 1973 Phillips and Votey, 1972 Ehrlich 1972). What remains unclear from this line of recent research, however is the extent to which the results of these studies can be generalized outside of the United States, for to our knowledge, no comparable research has been conducted in other countries. Accordingly, we can only speculate at this point whether similar results would occur if the deterrent effect of arrest, conviction and imprisonment were exammed in other than American states and cities. While it is clearly beyond the scope of the present mvest1gatmn to provide a conclusive answer to this question, by our examining the deterrence hypothesis for states and territories of India we hope to shed some light on this important issue. In addition, it is our hope that this research will provide the impetus to other investigators to further — 21 — explore the deterrence hypothesis in other western and non-western countries. Only through such efforts will a more clear understanding of the deterrence hypothesis result. Before presenting the methodology and results of the present investigation for India, it would first ssem of value to briefly review the available evidence for the United States (1). REVIEW OF THE LITERATURE Although the deterrent effect of one form of punishment — the death penalty — for one type of offense — murder — has long been of concern to American Criminologists, it was not until the late 1960°s that the deterrent effect of alternative forms of punishment for other offenses began to receive empirical attention in the criminology literature (2). The first notable effort in this direction was Gibbs’ (1968) investigation of the deterrent effect of the certainty and severity of imprisonment on homicide rates for states of the country. Utilizing crime and imprisonment figures issued by the Federal Bureaus of Investigation and Prisons, Gibbs examined the deterrence hypothesis of an inverse relationship between each measure of imprisonment and states’ homicide rates (3). As hypothesized, his rather limited chi square- based analysis showed a negative relationship between offense rates and the certainty (O —.48) and severity (= —.25) factors. His findings of a more subtantial inverse relationship for certainty of imprisonment, Gibbs concludes to be «entirely consistent with the position taken by practically every advocate of deterrence theory » (1968, 525). (1) For a good discussion of deterrence theory and a summary of the available evidence, see : Zimring and Hawkins (1973), Tittle and Logan (1973), Gibbs (1975), and Andenaes (1974). (2) For a good discussion of the theory and empirical research pertaining to the deterrent effect of the death penalty for murder, see : Sellin (1967), Bedau (1967), Bailey (1975), and Bedau and Pierce (1976). (3) Gibbs operationalized his certainty of punishment measure as the number of persons in each state sent to prison for homicide in 1960, divised by the total number of homicides reported to the police in that state for 1959-1960. The severity of punishment was operationalized as « the median number of months served on a homicide sentence by all persons in prisons on December 31, 1960 ». The dependent variable was defined as the average annual homicide rate per one hundred thousand population for each state from 1959-1961. The average homicide rate for a three-year period was used to allow for sufficient time for the deterrent effect and to provide greater stability to the rate. — 22 — A re-analysis of Gibbs' (1968) data by Gray and Martin (1969) using a more sophisticated statistical design has led to some modifi- cations of Gibbs’ findings. These researchers examined Gibbs’ data in its original ratio scale form rather than simply treating it as nominal as had Gibbs in his using of a contingency table analysis. They further examined these data by means of both a rectilinear and logarithmic (nonlinear) statistical model. The latter correlation model was utilized because some nonlinearity was apparent in the data. Using a rectilinear model, Gray and Martin found the correlation between the severity and certainty of punishment and homicide rate — — —.367 and r — — to be in the hypothesized inverse direction (r —,281, respectively). The corresponding log correlations were also of a negative sign and somewhat larger (r = —.506 and r = —.379, respectively). Where Gibbs found certainty of punishment to be the more important factor, Gray and Martin found the inverse to be true. Further, a multiple correlation combining the effects of the severity and certainty of punishment on offense rates yielded correlations (R = .442 for the linear model, and R =.599 for the logarithmic model) only slightly larger than the respective largest zero order correlations, thus only allowing an increment in explained variation of approximately .6 per cent and .9 per cent, respectively. These findings suggest, contrary to Gibbs’ (1968) assertion, that the severity and certainty of punishment are not ac!ditive in their effect on homicide. Tittle's (1969) deterrence investigation resembles quite closely the approach taken by Gibbs (1968). Utilizing police and prisoner statistics in a tabular analysis, he examined the relationship between the certainty and severity of punishment and offense rates for seven major felonies (4). The correlation between estimates of the certainty of punishment and offense rates for all seven offenses combined was —— — found to be in the predicted direction (Tau c — ,45). For the individual offenses the cœfficients are : sex offenses —— — —.17 ; assault = —.46 ; larceny = —.37 ; robbery — —.36 ; burglary = —.31 ; homicide — —.17 : and auto theft = —.08. Like the total crime index, each of these cœfficients is in the hypothesized direction, and with the exception of those for homicide and auto theft, are highly SŒtisücally significant (P< .01). - 000 (4) Tittle operationtally defined the certainty of punishment as the average number of admission to prison for each of the index offenses in each state, from 1959 to 1962, divided by the number of index offenses known to the police in each state, from 1958 to 1962. The severitv of punishment was operationalized as the mean number of months spent in prison by index crime offenders released in 1960. Tittle’s « Deviance Index » consisted of the mean annual number of index offenses reported to the police in each state per 100,000 population from 1959 to 1963. — 23 — The relationship between the severity of punishment and offense rates for all offenses combined was found to be Tau c ——utsetus —. .14. For —— — the individual offenses the correlations are sex offenses 26 ; assault —— — 18 ; larceny =.14 ; robbery =.05 ; burglary =.14 ; homicide = .45 ; and auto theft = .04. With the exception of homicide these findings are quite contrary to what deterrence theory would predict. In a more recent investigation, Chiricos and Waldo (1970) provide a further examination of the hypothesis that rates of crime are inversely related to the certainty and severity of punishment. Like Tittle (1969), they examine the deterrence hypothesis for each of the major index crimes, but for three points in time (5). Further, they examine the rcelationship between changes in the level of the certainty and severity of punishment and their effect on offense rates (6). Âs in Gibbs’ (1968) and T'ittle’s (1969) investigations, official police and prisoner statistics were used to construct punishment and rate indexes. Examination of the certainty-rate data revealed that all but one correlation was in the hypothesized negative direction. The strength and level of statistical significance of the associations varied greatly by offense and over time, however. Only for assault were the two variables consistently and significantly related for all three years. For burglary, the relationships were found to be statistically significant for two years, while the remaining correlations were either low or not consistent by offense (7). (5) Chiricos and Waldo’s operationalization of the certainty and severity of punishment is as follows 19XX certainty =— 19XX Admissions to Prison for “ X ” Offenses — Mean of “ X ” Crimes Known to Police in 19XX and 19XX Severity — Median Lenght of Sentence Served by State Prisoners Released mean of 1950, 1951, and 1952 rates ; mean of 1960, 1961, and 1962 rates ; and mean of 1963, 1964, and 1965 rates. (6) Chiricos and Waldo calculated changes in the levels of the severity, certainty and offense rates as follows 19XX-19YY % Change in Certainty = (19YY certainty) — (19XX certainty) PE 19XX certainty — — 19XX-19YY % Change in Severity (19YY severity) — (19XX severity) = — srrr 19XX severity 19XX-19YY % Change in Rate —— — (19YY rate) — (19XX rate) — 19XX rate (7) Chiricos and Waldo report the following phi cœfficients between the certainty of punishment and crime rates : homicide, .02 (1950), — 46 (1960), —— -17 (1963) ; robbery, — .26 (1950), — .33 (1960), — .25 (1963) ; assault, — 35 (1950), — .38 (1960), — .42 (1963) ; burglary, — .09 (1950), — .33 (1960), — .50 (1963) ; larceny, — .04 (1950), — .17 (1960), — 15 (1963) ; auto theft, — .16 (1950), — .25 (1960), — .16 (1963). — 24 — The severity-rate data also provided no consistent support for the deterrence hypothesis. With one exception (homicide, 1960), all cor- relations were either positive or low negative. Moreover, little consis- tency over time was exhibited within offence categories (8). Examination of changes in the levels of the severity and certainty of punishment and their effect on offense rates also produced incon- sistent findings (9). Only approximately 4 percent of the certainty-rate correlations were in the hypothesized direction and statistically signi- ficant, while none of the severity-rate correlations was as hypothesized and statistically significant. Chiricos and Waldo conclude that their data provide little evidence of deterrence (10). Utilizing a multiple regression analysis technique, Ehrlich (1972) also examined the hypothesis of a negative relationship between the severity and certainty of punishment and offense rates for the major index crimes for 1940, 1950, and 1960. Using punishment and rate indexes quite similar to those of the above investigators, he reports a consistent negative correlation (r) between severity and rate, averaging (1940-1960) as follows for each offense robbery = —.325 ; bur- glary — —.546 ; larceny = —.287 ; auto theft = —.174 ; murder = —.214 — — —.188 assault —— — b rape —.389 (11). Similarly, the association between certainty and rate was also found to be consistently in the hypothesized direction with cœfficients averaging : robbery = —.913 ; burglary = —.468 ; larceny — —.298 ; auto theft = — .247 ; — — murder — —.466 ; rape — —_,578 : assault = —.393. In addition, Ehrlich also reports that when median family income, percentage of families below one half of the median family income, and percentage nonwhite are introduced as control variables (all three factors are described by Ehrlich as « major theoretical determinants of (8) For 1960 and 1964 Chiricos and Waldo report the following phi correlations between severity of punishment and offense rates : homicide, — .33 (1960), — .03 (1964) ; robbery, .12 (1960), .00 (1964) ; assault, .32 (1960), — .08 (1964) ; burglary, .25 (1960), .08 (1964) : larceny, .14 (1960), — .09 (1964) ; auto theft, .13 (1960), .12 (1964). (9) Due to the complexity and scope of the change analysis, the statistical results of this analysis will not be reported here. The reader is referred to pages 209-210 of Chiricos and Waldo’s article. (10) Chiricos and Waldo are reluctant. however, to suggest that the severity and certainty of punishment fail to deter crime. Rather, they question whether deterrence may be adequately addressed through the use of police and prisoner statistics. (11) Ehrlich operationalized his certainty of punishment measure as the number of persons sent to prison in each state for each of the index crimes (1940, 1950, 1960), divised by the number of index crimes reported to the police for these years. It is not clear whether Ehrlich uses the mean or median length Of sentence served by released felons as his measure of severity, for he fails to reference the source of his prisoner data. He simply describes his severity measure as the « average » length of prison sentence. — 25 — criminal activity »), the relationship between punishment and rate remains generally unchanged. In spite of the shortcomings of police and prisoner statistics used to construct his punishment and rate indexes, and the somewhat stringent econometric specification of functional relationships required in the regression model used, Ehrlich (1972, 275) concludes that his data are « consistent with the hypothesis that law-enforcement acti- vity has a deterrent effect on offenders ». Furthermore, the fact that cœfficients for crimes against persons are not on the average lower than for crimes against property suggests that contrary to common belief, « law-enforcement activity may not be less effective in combating crimes of hate and passion than crimes involving material gain » (274). In an attempt to further explore the deterrent effectiveness of legal sanctions on crime, Bailey, e al. (1974 and Logan (1972) have provided additional evidence linking the certainty and severity of imprisonment to states’ felony rates. Re-analizing Tittle’s (1969) imprisonment and offense rate data for the period 1958-1962, Logan (1972) found a consistent inverse relationship between certainty and rates (homicide, r —.17 ; robbery, r —— —— —.17 ; sex offenses, r = —.36 ; assault, — — —— — —.54 ; burglary, r —.38 : larceny, r —.17). In contrast, with the exception of homicide, the severity- cœfficients were found to be either of a positive sign, or negative 3 less substantial than their certainty-rate counterparts (homicide, —.37; robbery, r = —.06 ; sex offenses, r = —.26 ; assault, r —.14 ; burglary, r —— — —.07 : larceny. r — — —.12 ; auto theft, LIl @@ HH 1 r —.19). While the results for severity were altered slightly, little additional support was found for the deterrence hvpothesis when the severity-rate relationship was examined holding constant (controlling for) the certainty of imprisonment. These findings lead Logan to con- clude that «….while certainty of imprisonment seems to have a direct and independent negative relation to crime rate, severity, if it has a deterrent effect, does not operate independently of certainty » (1972, 70). Considering the relationship between severity and certainty of imprisonment and felony rates for a more extended time period (1950, 1960, 1964) Bailey et al. (1974) reach a very similar conclusion. Like Logan (1972) and a number of previous investigators, they report a rather consistent negative relationship for each year between certainty and offense rates for each of the major felonies, and a less substantial negative, or positive, association between severity of imprisonment and offense rates for the major felonies (12). In addition, when both severity (12) To illustrate, Bailey c al. report the following zero-order r corre- lations for the severityv and certainty of punishment for 1964 ; homicide, — .165 (sev.), — .228 (cert.) ; robbery, — .068 (sev.), —.338 (cert.) ; assault, .087 (sev.), — .401 (cert.) ; burglary, .271 (sev.), — .527 (cert.) ; theft, .105 (sev.), — .450 (cert.) ; auto theft, .194 (sev.), — .528 (cert.). — 26 — and certainty were fit into a multiple correlation analysis, the results for each year were nearly identical to that found when only the certainty variable was considered in the analysis, thus again, suggesting the general lack of a deterrent effect of the severity of imprisonment (13). The same pattern of bivariate and multivariate results also generaily held when a non-linear correlation and regression model was utilized. While the above studies leave little question as to the probable importance of at least the certainty of imprisonment as a deterrent to crime, these investigations leave unaddressed the question of the deterrent effect of arrest and conviction on crime. In this regard, it is important to keep in mind that even upon arrest and conviction for a crime, all offenders are not imprisoned. Rather, many offenders, including convicted felons, receive jail sentences and/or probation, or some other form of legal sanction. In addition, as Logan (1975) has recently argued, arrest (even apart from it resulting in confinement in jail or prison) has to be looked upon as a form of punishment in itself. Arrest is inconvenient, typically embarassing and costly if one is required to provide bail to gain release upon awaiting trial. And, of course, for those who cannot afford bail, or are charged with a crime where bail is not permissable, arrest may result in confinement in jail until trial, regardless of the trial’s outcome. Recognizing these factors, the last few years have witnessed a handful of investigations examining the deterrent effect of arrest and conviction on crime. As with the earlier imprisonment studies, these investigations have typically examined the deterrence hypothesis of an inverse relationship between the certainty of arrest and/or conviction and crime rates. The first investigation of note of this sort was conducted by Orsagh (1973) for counties and large cities in California for 1960. Using both an ordinary least squares (O.L.S.), and a two stage least squares (T.S.L.S.) regression procedure in order to control for the two- way association between certainty and crime rates, he examined the deterrent effect of the certainty of conviction on total per capita felony rates, controlling for two important factors associated with crime ; percentage of population 15-30 years of age, and proportion of families living in poverty (14). The T.S.L.S. results revealed a ten percent increase in conviction rate (from the observed mean of convictions for (13) The multipile R correlations, combining the effect of severity and certainty e on offense rates for 1964 are homicide, .288 ; robberv, .372 , assault, 415 ; burglary, .529 ; theft, 459 ; auto theft, .532. (14) Orsagh defines certainty oïË conviction in his study as the number of persons convicted for all feionies relative to the total number of felonies reported by the police. —”— the counties and cities from 92 and 101 per 1000 felonies) to be associated with an 18 percent decline in the crime rate (from a mean offense rate of 10.2 to 8.4 per 1000 population). Less substantial were the O.L.S. results, with a ten percent increase in the conviction rate only being associated with a 2.6 percent decrease in total felony rate. In both analyses, risk of conviction proved to be equal to, or more important than, either age or poverty as a predictor of crime rate (15). From these findings, Orsagh concludes conviction rates to be an important determinant of crime, and a factor that cannot be ignored by future investigators. By his solely examining total felony rates, and not individual offenses, however, it remains unclear from his analysis how risk of conviction influences different crimes. In a very similar type of investigation, Tittle and Rowe (1974) examine the deterrent effect of the certainty of arrest on total felony rates for cities with populations of 2,500 or more and counties in Florida. Utilizing police figures for arrest clearance rates for 1971 (the proportion of crimes that the police report to be « solved » by arrest), they examine the deterrence hypothesis of an inverse relationship between the probability of conviction and total felony rates (16). Preli- minary analysis revealed the certainty-rate relationship to be non- linear for both cities and counties, with certainty becoming a more important factor after clearance rates exceed 30 percent. To illustrate, — the correlation between arrest and rates was found to be r — —.65 (P<.05) for counties and r = —.19 (P<.05) for cities. For counties with clearance rates above 30 percent, however, the correlation —— Wwas r — —.58, with an r value of only —.13 for those below this level. Similarly, for cities the respective cœfficients for those above and below the 30 percent level are — .48 and .19. These finding remained basically unaltered when a number of sociodemographic factors asso- ciated with crime rates were introduced into the analysis as control variables (1) percent male population, (2) percent population 15-24 years of age, (3) percent male population 15-24 years of age, (4) percent population age 65 and older, (5) size of city, or percent of county (15) Orsagh’s regression results utilizing the O.L.S. procedure are as follows, with t values in parentheses riskx of conviction — — .34 (3.1) ; age 15-30 years — .20 (1.8) ; poverty = .20 (18). The T.S.L.S. results are risk of conviction — — 2.30 (2.3) ; age 15-30 years = — .22 (18) ; poverty = 1.22 (2.3). The larger T.S.L.S. cœfficient for convictions lead Orsagh to conclude the T.S.L.S. Model to be preferable to the O.L.S. Model. (16) Tittle and Rowe do not use the ratio of total reported crimes to total arrests as their measure of certainty of arrest, but rather arrest clearance rates computed by the police. « Arrest clearance means that the police have satis- fied themselves that they have taken into custody the person who committed the crime » (1974, 456). As the writers point out, because arrest (and the reported clearance ratios by the police) is a product of police discretion, arrest data are subject to some distortion. — 28 — living in cities of 20,000 or more, (6) women in the labor force, (7) percent of population in poverty, (8) percent negro population, (9) percent non-white population, (10) median education, and (l1) inter- quartile range of income. These findings, Tittle and Rowe (1974, 458) conclude, indicate «that there is a critical level [30 percent] that certainty must reach before there is a noticeable change in the volume of crime ». They go on to argue that their data clearly indicate that legal sanctions can no longer be ignored as deterrents to crime, and that « perhaps the missing link » in criminology is the « fear of sanctions » (461). With Tittle and Rowe’s analysis, as with Orsagh’s (1973), however, it still remains unclear the relationship between the probability of arrest (and/or conviction) and offense rates for different crimes. To meet this problem, Bailey (1976) replicated Tittle and Rowe’s (1974) investigation of Florida cities and counties for 1972, but he further examined the certainty of arrest-offense rate relationship individually for the following felonies ; murder, forcible rape, robbery, assault, breaking and entering, larceny, and auto theft (17). As deter- rence theory would predict, this analysis showed the arrest-offense rate relationship to vary by type of offense. For Florida’s counties, the correlations between arrest rates and offense rates are as follows — murder, r = —.201 (P<.05) ; rape, r — 157 ; robbery, r = .136 ; assault, r — — —.500 (P< .001) ; breaking and entering, r = —.581 (P<.001) ; larceny, r = —.329 (P<.01) ; total felonies, r = —.631 (P< .001). For Florida’s cities, the certainty-felony rate results are — murder, r = .635 (P<.001) ; rape, r —— .461 (P<.001) ; robbery, r — .082 ; assault, r —— — —.165 ; breaking and entering, r — — —.165 ; larceny, r — — —.204 (P<.01) ; auto theft, r = —.040 ; — — —, total felonies, r 190 (P<.01). These findings, like Tittle and Rowe's (1974), lead Bailey to conclude the importance of certainty of arrest as a deterrent to crime. Unlike Tittle and Rowe'’s investigation, however, his analysis shows the deterrent effect of arrest to be far from uniform for all types of offenses. To the contrary, « for offenses commonly considered to be “instrumental” in nature, arrest would appear to be a much more important determinant of the level of crime than for “expressive” crimes » (Bailey, 1976, 153). In addition, this investigation shows that the tipping point (the level of certainty at which a substantial deterrent by the police (arrest clearance ratio) which may be manipulated to better reflect (17) To avoid the problems inherent in using a measure of arrest supplied upon police efficiency and performance, Bailey used the ratio of reported crimes to reported arrests in computing a certainty index for each felony. effect begins to take affect) varies by type of offense, with Tittle and Rowe’s (1974) suggested figure of 30 percent only applying for some offenses. To further explore the possible differential deterrent effect of the certainty of arrest on crime rates, Logan (197S) examined this question for states for the period 1964-1968. Drawing upon unpub- lished Federal Bureau of Investigation’s data, he intercorrelated average arrest clearance rates and offense rates for total felonies, as well as for eight individual crimes that make up the F.B.L’s Index Offense Category (18). This analysis revealed a low to moderate negative corre- lation between the certainty of arrest and rates for all but homicide : homicide., 1 — —— 25 (P<.05) ; rape. r = —.37 (P<.01) ; robbery r —— ——_—. 48 (P<.001) ; assault, r = —.09 ; burglary, r = —.32 P <.05) ; larceny,r = — .45 (P<.01) : auto theft, r = —.64 (P<.001) : total felonies, r = —.26. In contrast to Tittle and Rowe (1974) and Bailey’s (1976) findings, with one exception, Logan’s data did not reveal the certainty-rate relationship to be curvilinear, with a « tipping effect » occuring at the 30 percent level of certainty. Only for robbery was the relationship found to be nonlinear, with the deterrent effect of arrest decreasing beyond the 36 percent certainty level (for this offense, the log r value is —.64, compared to —.48 for the linear statistical model). Consistent with earlier investigations, however, Logan did find the certainty of imprisonment and offense rates to be inversely related for the period 1964-1968 ; homicide,r = —.19 ; rape,r = —.73 (P<.001) ; robbery, r = —.59 (P<.001) ; assault, r = —.65 (P<.001) ; burglary, r = —.46 (P<.001) ; larceny, r = — .40 (P<.0I) ; auto theft, r = —.12 ; total felonies, r = —.50 (P<.001). Comparison of these imprisonment figures with those for arrest, shows the former correlations to be gene- rally larger, with the exception of larceny and auto theft. For the remai- ning offenses, Logan suggests that imprisonment is probably a better de- terrent than arrest due to its greater severity, while arrest may better deter larceny and auto theft because of the extremely low certainty of imprisonment for these two offenses ; a mean probability of impri- sonment of .02 (2%) for larceny, and .01 (1%) for auto theft. In sum, the above, as well as a number of additional investigations, clearly point to the importance of arrest, conviction in court, and imprisonment as deterrents to crime in the United States. Beginning with Gibb’s (1968) early and rather elementary investigation of the (18) Like Tittle and Rowe (1974), Logan used the number of crimes cleared by arrest divided by the number of crimes known to the police for the period 1964-1968 as a measure of certainty of arrest. — 30 — deterrent effect of the severity and certainty of imprisonment on homi- cide rates, there has been a steady progression of such studies focusing upon other offenses and types of legal sanctions, and making use of more sophisticated methodological and statistical analyses. As noted earlier, however, these studies have been exclusively confined to juris- dictions (states, cities, counties) within the United States. Because of this, it thus remains unclear whether the pattern of results summarized above also hold for other countries. As a first step in addressing this question, we examine here the deterrent effect of arrest and conviction for selected serious crimes for states and territories of India. METHODS AND PROCEDURES Consistent with the deterrence doctrine and previous investigations, we hypothesize a substantial inverse relationship between the certainty of arrest and conviction and offense rates. In order to examine these questions, measures of the certainty of arrest and conviction and crime rates were constructed utilizing figures issued by the Indian government in its annual publication, Crime In India, for the period 1967-1973 (19). As with previous similar investigations in the United States, cer- tainty of arrest was operationally defined as the number of arrests for each type of crime divided by the total number of such crimes reported by the police. Similarly, certainty of conviction was measured by divi- ding the total number of convictions for each offense by the total number of such crimes reported by the police. This procedure resulted in an arrest and conviction index that can theoretically range from zero (0.0) to unity (1.00) for each of the following crimes : murder, rape, robbery, house breaking, and theft. À certainty value of zero would indi- cate that there were no arrests (or convictions) for the offense in question, while a value of one would indicate that there were an equal number of arrests (or convictions) and reported crimes (20). Also, like in previous deterrence investigations in the United States, crime rates for states and territories were computed as the number of offenses per 100,000 population. (19) Our analysis here is confined to the period 1967-1973 for complete police and sociodemographic data have yet to be received for earlier and later years. These data have been requested, but time constraints for this publication do not permit their inclusion in this investigation. (20) It should be kept clearly in mind that certainty estimates cannot be interpreted as the proportion of offenders who are arrested and convicted. Such data on individual offenders. while prefcrable. are simplv not available. — 3} — In addition to police figures for arrest, conviction, and crime rates, data were gathered for the following socioeconomic factors in order to introduce them into the analysis as control variables : population, per- cent urban population, percent illiteracy, and percent unemployment (21) These data were secured from the annual Indian Government publi- cation, Statistical Abstracts of India. By considering these factors along with the measures of arrest and conviction, it will be possible to deter- mine both the effects of each of these variables on crime rates, as well as the independent deterrent effects of arrest and conviction on crime rates controlling for these sociodemographic factors. These control variables have all been suggested as determinants of crime rates, and are roughly comparable to control factors utilized in similar investi- gations in the United States (22). To determine the deterrent effect of arrest and conviction on crime rates, both certainty measures, as well as the sociodemographic varia- bles, were fit into a step-wise multiple regression analysis for each offense and year. Examination of the unstandardized regression cœffi- cients resulting from this analysis will allow us to determine the effect of a unit change (increase or decrease) in the deterrence and socio- demographic variables on crime rates. Comparison of the standardized (beta) regression cœfficients will permit an examination of the relative deterrent effect of arrest and conviction, as well as the relative effect of the sociodemographic factors, on offense rates. In addition, by comparing the results for each offense, it will be possible to determine the possible differential deterrent effect, and the possible differential effect of the sociodemographic factors, on different types of crimes. While in many previous deterrence investigations statistical tests of significance have been used to determine the significance of simple bivariate and regression cœfficients, such tests will not be used in this investigation for we examine here complete population data, and not a random sample of India’s states and territories. In addition, the years considered in this investigation (1969-1973) were not randomly selected, but were rather, the only years where relatively complete police and sociodemographic data were immediately available for analysis. (21) In the absence of alternative data, the proportion of the states’ (or territories’) population on the livé registry of employment exchanges was used as an indicator of unemployment. (22) The control variables selected for inclusion in this analysis were chosen due to their similarity to factors used in deterrence investigations in the United States, and their felt importance to crime as reflected in the criminology litera- ture on India and other developing countries (Clinard and Abbot, 1973). — 32 — Missing and Incomplete Data While it was our hope initially to examine the deterrent effect of arrest and conviction on crime rates for eacn state and territory in India for each year 1967-1973, this ideal had to be compromised due to problems of missing and incomplete data. Turning first to the offenses examined, reasonably complete offense rate figures were found to be available for murder, robbery, house breaking, and theft for each year. Rape data are only reported in the Crime In India series for 1971- 1973, however. Accordingly, it was not possible to examine this offense for the earlier four vear ( 1967—1970)peri0d. In addition to this difficulty, figures required in order to compute an index of the certainty of conviction are not available in the Crime In India reports for 1972 and 1973. As a result, for these two years it was only possible to examine the deterrence hypothesis for the certainty of arrest. Examination of the conviction variable for the earlier five year period, however, will allow us to at least speculate about convic- tions’ deterrent effect for 1972 and 1973. Finally, in searching for data for the control variables, we found that population, urbanization, and illiteracy figures were not available for each of the years considered in the analysis. Data for these variables are only available for the census years 1961 and 1971. Due to this problem, figures for the nearest census year were used as an estimate for these three control factors for cach year, 1967-1973. In contrast, unemployment figures were found to be available for each year for the period 1967-1973, with missing data only being somewhat of a problem for a few federal territories (23). Possible Spuriousness oË Certainty-Rate Results Before presenting the results of our analysis, it would seem of value to at least briefly examine an argument raised by Chiricos and Waldo (1970) as to the possible spuriousness of findings resulting from examining the relationship between an independent variable (certainty OË arrest, and conviction) and a dependent variable (crime rates) that contain a common term. They argue that since certainty indexes such as ours (number of arrests/number of crimes, and number of convic- (23) Due to problems of missing data for one or more of the control variables, the original N size of 26 to 28 states and territories of India (the number of states and territories varies by vear) was reduced in some cases as follows by year 1973 — 24, 1972 — 28, 1971 =— 23, 1970 = 24, 1969 —26, 1968 = 24, 1967 = 22. In almost each case where data were missing, it was for one of the territories. — 33 — tions/number of crimes) have in their denominator the same term that appears in the numerator of our crime index (number of crimes/ population), correlating such measures will result in a spurious negative association, even when there is no association between certainty and offense rate. Because of this problem, Chiricos and Waldo conclude that their findings of an inverse relationship between the certainty of imprisonment and crime rates were spuriously produced and do not provide evidence of deterrence. As pointed out by Bailey, et al. (1972, 1974) ; Logan (1972, 1975) and other recent deterrence investigators, the spuriousness problem that Chiricos and Waldo speak of is only of minor importance, for it only applies under three conditions that do not hold in most deterrence investigations (1) when the number of persons arrested convicted, or imprisoned equals the number of persons in the popu- lation, (2) when the number of those arrested, convicted, or imprisoned and population are constant, or (3) when the number of persons arres- ted, convicted, or imprisoned and size of the population are unrelated None of these assumptions hold for either Chiricos and Waldo’s or our study for (l) number of arrest, convictions or prison admissions do not equal population size, (2) arrests, convictions, or imprisonment and population are not constant in the data examined, and (3) arrests convictions, imprisonment, and population are positively correlated with each other, and with the number of crimes, ie., the larger the population, the more crimes arrests convictions and admissions to prison (Logan, 1972). In addition, and also contrary to Chiricos and Waldo’s argument of spuriousness, there would appear to be consensus that when examining two ratios that are theoretically meaningful as ratios (in this case crime rates, and certainty rates), and when hypotheses are stated in terms of the relationship between these ratios, the questions of indexal artifact due to a common term being in each does not arise (Logan, 1972 ; Kuh and Meyer, 1955 McNemar, 1969 ; Ragarajan and Chatterjee, 1969). With this matter addressed, we now turn to the statistical results of the analysis FINDINGS Results for Homicide Table 1 reports the relationship between certainty of arrest and conviction (1967-1971 only), the sociodemographic factors and homi- cides rates by year. — 34 — TABLE 1 Relationship between Certainty of Arrest and Conviction, selected sociodemographic variables, and murder rates for states and territories of India, by year. Year Independent Simple Multi- Beta Variable r pie R2 Cœffi Cœff. —0s t — _— — - 1973 Pct. Urban . « * % + » « © »» © — » e 32 187 — 2.016 448 —— Cert. of Arrest . . …. ... 195 294 — .505 323 Pct. Illiterate . * — © © « à © © # e » 337 372 022 327 Population 046 382 — .000 111 Pct. Unemployment . … . …. 290 392 36.290 103 1972 Pct. Arrest . . . .. ... 353 124 — .580 341 Pct. Urban e ». . « e . % % …. < < < — À 245 156 782 185 Pct. Illiterate . . ..." 148 150 008 171 185 Population 064 000 082 Pct. Unemployment . ... 012 186 -— 5.409 017 1971 Pct. Illiterate . . . © … à . % » € % e 510 261 026 469 Cert. of Conviction e » . ... 365 345 — .034 330 Pct. Unemployment e … . ... 193 398 — 80.004 209 Cert. of Arrest e e » * …. ... 056 405 1.078 095 Population 048 405 009 1970 Pct. Illiterate . « . » . 6 + « 6 . » « 386 149 198 372 Pct. Urban 287 797 6.942 142 Population . . ... 008.0s%68 221 252 210 Pct. Unemployment . ... 161 260 — 579.316 134 Cert. of Arrest » ... 135 271 -— 21.524 182 Cert. of Conviction . . . .. 047 28 01 123 1969 Pct. Illiterate . « + * <. © * #% … #% # à 12 045 070 214 Cert. of Arrest 175 079 — 8.009 103 Pct. Urban 200 112 6.861 249 . . ... 102 127 Pct. Unemployment 397.728 178 Cert. of Conviction — .093 145 — .010 138 Population «s « . * . « —. s. … © e » e —# 062 146 000 025 1968 Cert. of Conviction 112 013 031 414 Cert. of Arrest « … - * e « » — € — — .020 039 — 19.896 335 Pct. Unemployment e e — » — — — .075 054 — 468.071 156 Pct. Urban 009 060 — 3.281 099 . » * © e 6 » e — < » « » » 094 062 000 044 Population Pct. Tlliterate . . . » …. 0..0.. 031 063 013 027 1967 Population « e » » * …e 8868066006 383 147 456 Pct. Illiterate 141 205 — .005 — .242 Cert. of Conviction . ... 038 210 — .005 — .079 Pct. Unemployment 031 210 — 3.758 — .025 Pct. Urban 043 210 — .029 — 015 Cert. of Arrest . . ... 139 210 102 023 — 35 Looking first at the results for arrest, the partial regression cœffi- cients (b and beta values) show the certainty-rate relationship to fall in the predicted negative direction for five of the seven years. Only for 1967 and 1971 are the cœfficients positive, with the latter being somewhat substantial, and the former being very slight. With the excep- tion of these two years, the regression results show a one percent increase in the certainty of arrest to be associated with the following decreases in the number of homicides per 100,000 population : —.51 (1973), —.58 (1972), — 21.52 (1970), — 9.01 (1969), and — 19.90 (1968). Examination of the raw data for arrest and homicide fails to provide any insight to account for the substantial variation in the size of the negative cœfficients for these five years, nor why the cœfficients are positive for 1967 and 1971. Turning to the findings for certainty of conviction, little apparent support is provided for the deterrence hypothesis. For the five years where conviction data are available, the unstandardized (b) regression results show the conviction-rate relationship to be negative in three cases (1971, 1969, 1967) and positive in two (1970, 1968). Regardless of the sign of the cœfficients, however, a one percent increase in the probability of conviction 1s associated with no more than approximately a three one hundreds of a person change (increase in 1968, and decrease in 1971) in the number of homicides per 100,000 population. For the remaining years, the trade-off between levels of conviction and homicides is only one third or less than this amount. Due to arrest, conviction and the sociodemographic factors varying in their units of measurement, it is not possible to directly compare the size of the unstandardized (b) regression cœfficients in order to assess the relative importance of each of these variables on homicide rates. By standardizing the regression results by the scale of measu- rement used for cach variable, however, direct comparisons of the effect of each factor can be made. Such standardized beta cœfficient are reported in the last column of Table 1. Looking first to the question of the deterrent effect of arrest vs. conviction for homicide for the five years where data are available for both measures, a year-by-year comparison shows the beta values to be larger for convictions for four years : 1971, 1969, 1968, and 1967. For 1968, the conviction cœfficient is positive, however. Only for the two ycars (1967. 1970) where the conviction results are positive are the arrest cœfficients larger in absolute size. For the only year (1969) where the arrest and conviction results are both negative, the beta value 1s larger for the latter factor. Whereas for this year the unstandardized (b) regression results show a unit change in the level of arrest to be associated with a larger reduction — 36 — in the homicide rate (— 8.01) than a unit change in the certainty of conviction (— .010), the relative deterrent effect of conviction (—.138) is slightly greater than that of arrest (—.103). Further examination of the last column of Table 1 reveals some additional interesting, and somewhat surprising results. With the excep- tion of 1968 where the conviction cœfficient is positive (but the arrest cœfficient is negative), and 1972 where the arrest results are negative, one or more of the sociodemographic variables proves to be a better predic- tor of homicide than either of the certainty variables. No single socio- demographic factor, however, proves to be consistently most strongly associated with homicide rates. Moreover, the signs of the cœfficients for these variables vary from year to year, with population, urbanization, unemployment, and illiteracy sometimes being positively, and some- times being negatively correlated with homicide. This mixed pattern of results calls into question the common argument that these socio- demographic factors exert a positive pressure on crime rates ; at least for homicide this assumption is not supported. Results for Rape Table 2 reports the relationship between certainty of arrest and conviction (1971 only), the sociodemogaphic factors and rates of forcible rape for 1971-1973. As noted earlier, rape data are not available for the earlier period 1967-1970. TABLE 2 Relationship between certainty of arrest and conviction selected sociodemographic variables, and forcible rape rates for states and territories of India, by year. — _ Multi B Beta Year Independent Simple Variable r ple R Cœfi_î . Ceffi. — 1973 Pct. Unemployment e « … ... 241 058 — 96.968 257 Cert. of Arrest . s. »* » . e % + # e 021 068 — .303 — 102 Pct. Illiterate 096 074 .004 059 Pct. Urban 098 075 .153 032 Population e s . …. s. #* + % <. e 6 © © » 089 075 .000 —— 029 1972 Pct. Urban ° . . « « o —s « « % — « »# —— 407 166 — 484 423 Cert. of Arrest . + « e — — 6 © — * 136 194 - .079 212 Pct. Illiterate 030 197 — -001 058 092 200 000 057 Population e © e « » À 047 201 — 3.522 —— 040 Pct. Unemployment 1971 Cert. of Conviction 237 056 — .014 —— 525 Pct. Illiterate . « * » » ». © € 6 ® — — 118 110 — .005 391 . . . . ». 188 181 23.942 252 Pct. Unemployment 208 — .259 218 Pct. Urban e + © © + © » — © % — e » » —— 047 Population . e . - — #% 9066 58a06% 107 215 .000 086 Cert. of _êrrest . … * — « 6 » 6 % à ; 021 ——n 21 — .038 — 030 — — 37 — Consistent with our hypotheses, the regression results show the certainty-rate relationship to be negative for both arrest and conviction. For 1973, the unstandardized (b) cœfficient for arrest shows a one percent increase in certainty to be associated with a decrease oË .30 rapes per 100,000 population. For the earlier two years, a unit increase in arrest is only associated with less than a one tenth of a person reduction in rape rates. Even a less substantial trade-off between certainty of conviction and rape rates is found for 1971. Because figures are only available for both arrest and conviction for 1971, it is only possible to compare the relative deterrent effect of these two factors for this year. Examination of the beta cœfficients for this year shows conviction to be more strongly associated with rape rates than arrest. In fact, the standardized cœfficient for conviction is Over 15 times as large as its counterpart for arrest. In the absence of conviction data for the later two years, however, it must remain an open question whether the same pattern of results would hold for 1972 and 1973. Hopefuily, arrest and conviction data will become available f£or later years (after 1973) in order to further explore this matter. In the meanwhile, all that can be done at this point is to tentatively conclude conviction to be a more important deterrent to rape than arrest. Further examination of the standardized beta cœfticients reveals a mixed pattern of results somewhat similar to that found for homicide. Compared to certainty of arrest, unemployment (1973) and percent urban population (1972) are found to be twice as good of predictors of rape rates. Similarly, for 1971, both unemployment and urbanization, as Well as illiteracy, are much better predictors of rape than arrest. The signs of the cœfficients for these three sociodemographic factors (except for unemployment for 1971) are negative, however, which 1s inconsistent with the common argument that these factors are positi- vely related to crime rates. To the contrary of this assumption, our findings for rape show either a mixed pattern (1971), or a predomina- tely negative association between these variables and offense rates. Results for Robbery Table 3 reports the relationship between certainty of arrest and conviction, the sociodemographic variables, and robbery rates. TABLE 3 Relationship between certainty of arrest and conviction, selected sociodemographic variables, and robbery rates for states and territories of India, by year —— e 2 e n Year Independent Simple M ulti- B Cm Variable Cœffi. Beta ple R2 —ecumer ce 00e 0 —— —cames — 1973 Population e . . …. 218 048 .000 178 Cert. of Arrest 202 082 — 489 — 160 Pct. Unemployment 145 091 57.144 095 Pct. Illiterate e e …. #e0e60s — 156 094 — 007 060 19'72 Pct. Illiterate .. ... 325 106 — .021 295 295 15 .GO9 252 Population Cert. of Arrest — 054 175 —.303 206 Pct. Urban 123 19 — .860 132 Pct. Unemployment 116 194 — 16.342 —— 032 i971 Population « ». . . ... 56 126 .000 392 Pct. Urban 199 189 2.672 348 Pct. Unemployment 218 81 193.373 322 Cert. of Conviction 191 295 — .0i2 128 Pct. Illiterate . ... 031 208 .006 070 Cert. of Arrest 055 302 638 073 1970 Population . . …. ... 835 697 010 956 Cert. of Conviction 304 711 -— 1.138 161 Cert. of Arrest 163 724 — 1.693 101 Pct. Urban 072 730 434.346 078 Pct. Illiterate ... 107 735 — 4.614 076 Pct. Unemployment O11 '73\ — 6339 434 013 1969 Population . . ... 820 672 .007 795 Pct. Illiterate »ecec6t:eoes 095 682 — 5.824 092 Pct. Unemployment 058 684 — 18845.716 —— 044 Cert. of Conviction 512 685 253 037 Pct. Urban « « » #* 06060es666e 127 685 — 91.762 017 Cert. of Arrest 116 685 273 014 1968 Population « e . …. ... 816 565 .005 542 Cert. of Conviction 705 737 2.682 349 Pct. Illiterate . +e %* 006066 363 TM — 17.891 234 Pct. Unemployment 026 780 — 55869.837 117 Cert. of Arrest . …. 049 “J 83 1.250 072 Pct. Urban « . ... 105 784 —— 238.584 041 1967 Pct. llliterate SO1 251 .016 302 Pct. Unemployment 376 307 49.137 136 Population . * « ». e % — e e% 038 317 009 — 415 Cert. of Arrest e ... 356 329 —- 3.209 — .347 Cert. of Conviction 163 376 025 407 Pct. Urban 009 376 067 014 =— —e Fs —— es crr e —0s r - — 39 — Examination oË the results for arrest again only provide partial support for the deterrence hypothesis. While for 1967, 1970, 1972, and 1973 the cœfficients are in the predicted negative direction, the results are positive for the remaining three years. In addition, as with homicide, the unstandardized (b) regression cœfficients are not uniform in size, with a trade-off between a one percent increase in the arrest rate being associated with as much as a 3.2 person (1967) reduction, to a 1.7 person (1970) increase in robbery rates. Similarly, the results for the conviction variable also show a mixed pattern, with the certainty-rate relationship being in the hypo- thesized negative direction for two years (1970, 1971) and positive for three ycars (1967-1969}. Comparison of the unstandardized (b) cæœflfi- cients also shows them to vary in magnitude, with the result of a unit increase in the certainty of conviction ranging from a 2.7 (1968) increase to a 1.1 (1970) decrease in the number of robberies per 100,000 popu- lation. Again, it remains unclear from our examination of the raw data why the arrest and conviction results for robbery vary so substantially by year. Looking to the question of the relative deterrent effect of arrest vs. conviction for robbery, the last column in Table 3 shows only one year (1970) where both of these factors are negatively associated with rate. Here, conviction proves to be a slightly better predictor of robbery rates than arrest. For 1971, the negative beta cœfficient is also larger for conviction than arrest, although the sign for the latter variable 1s positive. For 1967, the conviction cœfficient is also larger than that for arrest, but the results for the former variable are positive, and negative for the latter. Finally, for 1968 and 1969, the results are positive for both certainty factors, with the cœfficients being larger for convictions. Also, like the previous findings for homicide and rape, examination of the relationship between the sociodemographic factors and robbery rates reveals a mixed pattern. Whereas for some years, population, urbanization, illiteracy, and unemployment are positively associated with offense rates, the opposite occurs for other years. Further, even for the years where the sign of the cœfficients is the same for these variables, the cœfficients are generally not uniform in size. Results for House Breaking Table 4 reports the relationship between certainty of arrest and conviction, the sociodemographic variables, and rates of house breaking (burglary). — 40 -- TABLE 4 Relationship between certainty of arrest and conviction, selected sociodemographic variables, and rates of house breaking for states and territories of India, by year _— _— rrrs t cm 08 00 Year Independent Simple Maiti- B Cœffi. Variable piz R Cœffi. Beta _— 1973 Population 259 067 .000 397 Pct. Illiterate e . ... 105 132 155 260 Pct. Urban . « »* # eeses 220 176 9.024 225 Cert. of Arrest 040 179 — 1.820 039 082 130 48.936 016 Pct. Unemployment Pct. Urban . . + — « ». » » « à 365 143 13.366 389 1972 212 — 9.719 235 Cert. of Arrest 227 Population 171 254 .000 —— 192 Pct. Unemployment 273 276 — 451.273 170 Pct. Illiterate e ». . ... 090 281 — .027 070 1971 Pct. Urban 310 096 10.975 266 Cert. of Arrest 305 148 — 11.912 —— 170 112 163 000 132 Population 078 Pct. Tiliterate e . ... 009 166 037 Cert. of Conviction 275 171 —- .068 082 Pct. Unemployment 033 171 — 55.621 017 1970 Population 627 393 008 593 Pct. Urban 093 433 1833.779 247 155 476 150638.997 229 Pct. Unemployment 372 505 — 1.023 275 Cert. of Arrest Pct. Illiterate « » < » © » # » 097 542 15.469 191 Cert. of Conviction 119 544 —— .574 058 1969 Population 680 462 .008 648 Cert. of Arrest . . ... 29° 505 — .643 237 Pct. Illiterate 088 519 — 8.902 103 Cert. of Conviction 002 522 101 081 063 Pct. Unemployment 025 524 — 36594.713 Pct. Urban 143 527 — 444.163 061 1968 Population . . …. ... 594 353 007 479 Pct. Urban . » . © —. » « e à 319 460 - 307027.623 442 Cert. of Arrest 438 604 — 1.753 451 Cert. of Conviction 299 677 2.441 225 031 695 17.183 155 Pct. Tlliterate e » » e8%66668 025 Pct. Urban « . e « % » % ® » à 018 696 212.572 1967 Cert. of Arrest 625 391 —- 28.995 382 Population e e © 6 6 — — e e # 512 519 000 351 Pct. Unemployment 402 581 592.748 212 Pct. Urban « e » ... 192 594 — 4.729 — 131 Pct. Illiterate » .. ... 382 599 098 — .048 .040 A N — .068 Cert. of Conviction —- 485 599 — 4} — As with rape, but unlike for homicide and robbery, the results for burglary are quite consistent with what deterrence theory would lead us to expect. With but two exceptions (convictions, 1968 and 1969), the certainty-rate results fall in the hypothesized direction for both measures of sanctions for each year. Turning first to the findings for arrest, the unstandardized (b) regression cœfficients are generally much larger than those found for the offenses previously examined, although their size does vary by year, ranging from a low of — .64 (1969) to a high of — 29.0 (1967). For the seven year period, the mean unstandardized cœfficient for arrest is — 7.98, thus indicating on the average a one percent increase in the certainty of arrest to be associated with about an eight person reduction in the number of burglaries per 100.000 population. This is only an average estimate, however, with the trade-off between certainty of arrest and burglary ranging at the extremes from less than a one person reduction in rates (1969) to about a 29 person (1967) reduction in burglary rates per a one percent increase in the probability of arrest. Examinations of the regression (b) results for conviction again shows a somewhat mixed pattern, with three of the five cœfficients being in the hypothesized direction. For 1968 and 1969, the conviction- offense rate relationship is positive, although the cœfficients for these years (+ 2.44, +.701) are larger in absolute size than any of the negative cœfficients. In addition, with the exception of the negative finding for 1970 — .574), the regression results for the remaining years show a one unit increase in the certainty of conviction to be associated with no more than a one tenth of a person decrease in the number of robberies per 100,000 population. Comparison of the relative deterrent effect of arrest vs. conviction for burglary shows the former variable to be generally the more important of the two. With the exception of 1968, the standardized beta cœfficients are larger for arrest for the years (1967-1971) where the effect of the two types of sanctions can be compared. And even for 1968, the beta cœfficients for arrest and conviction are very comparable in size, with the later only being slightly larger (—.653 vs. —.614). These findings for burglary provide the most clear cut evidence in support of the deterrence hypothesis of the offenses yet examined. In addition, although variation exists for some years for some variables, the relationship between the sociodemographic factors and burglary rates fall generally in a positive direction as one might predict. Results for Thefït Table 5 reports the relationship between certainty of arrest and conviction, the sociodemographic factors, and theft rates. — 42 — TABLE 5 Relationship between certainty of urrest and conviction, selected sociodemographic variables, and theft rates for states and territories of India, by year —— Year Mulii B Beta Independeñt Simple Variable ple R2 Ceæffi Cœffi —— 1973 Cert. of Arrest 516 266 — 72.552 — .505 Pct. Illiterate . .. ... 301 360 595 261 Pct. Urban . « © * « © » © © — 225 365 8.410 055 Population 010 369 .000 — .066 227 372 -— 725.357 060 Pct. Unemployment 1972 Cert. of Arrest 387 149 — 77.328 433 Pct. Urban 219 238 35.011 — 255 Pct. Unemployment 267 260 —- 1449.310 137 Pct. Hliterate « . ... 079 268 123 081 Population « e « e + « — = » 081 271 000 058 1971 Cert. of Arrest —— 412 169 210.496 547 Pct. Urban . . .. ... 332 239 44.893 190 Population 149 264 .000 220 Cert. of Conviction 265 276 856 185 Pct. Illiterate . ... 171 281 — .184 077 Pct. Unemployment 002 283 -— 1052.104 056 1970 Cert. of Arrest 474 224 — 826.080 — 1.489 Cert. of Conviction 262 .366 6.135 068 Population « e * « * « » 6 % e 205 A86 .000 335 Pct. Urban 241 510 40.615 133 Pct. Illiterate 094 522 356 107 —— Pct. Unemployment 091 528 — 2388494 088 i969 Cert. of Arrest « . . © » e —— 340 lé0 — 470.810 796 Population . . ... 212 297 .000 — 439 Cert. of Conviction 233 334 2.068 321 Pct. Unemployment 152 335 — 857.138 — .033 Pct. Illiterate e e » * % — 8 —— 037 336 — .123 -— .032 Pct. Urban . ... 162 337 10.092 031 1968 Cert. of Arrest 508 258 -— 357.573 — 1.539 Cert. of Conviction 250 387 2.769 1.130 Pct. Urban » » » ® e — — e o — 276 512 35.796 300 Population ... 081 593 000 — 318 Pct. Unemployment 136 621 — 1682.681 — .172 Pct. Illiterate 236 622 .056 036 1967 Pct. Illiterate . « e © e » — — 432 187 099 123 Population « » . ». % © % © » e 369 261 000 207 209 265 298.469 055 Pct. Urban Pct. Unemployment 071 270 7.933 112 Cert. of Conviction — .077 272 683 537 Cert. of Arrest ... — .312 305 84.610 —- 632 —— — 43 — Âs with burglary, the regression results for arrest are consistent with the deterrence hypothesis for each year. Examination of the unstandardized (b) cæfficients shows the certainty of arrest-theft rate relationship to be in the predicted negative direction in each case. The cœfficients vary somewhat in size, however, ranging from a negative association of — 72.6 (1973), to a very substantial negative association for 1970 (— 826.1), 1969 (— 470.8), and 1968 (— 357.6). The mean unstandardized cœfficient for the entire seven year period is — 299.9, thus indicating on the average a one percent increase in the certainty of arrest to be associated with about a 300 person reduction in the theft rate for the years considered. In complete contrast, the regression results for certainty of conviction provide no support for the deterrence hypothesis. To the contrary, the cœfficients are positive for convictions for each year, thus indicating increases in the certainty of this factor to be associated with increases, and not a decrease, in theft rates. For none of the five offenses examined here, are the results for convictions so contrary to our hypothesis. Âs wWith our earlier findings, no consistent pattern is observed between the sociodemographic variables and theft rates. For some years these factors are positively associated with oflense rates, while for other years the signs of the cæœfficients are reversed. Without exception, however, the beta results for each year show certainty of arrest to be a better predictor of theft rates than any of the sociodemographic factors. In short, for no single offense nor year are the findings for arrest so consistent, but the results for conviction so inconsistent, with the deterrence argument. SUMMARY AND CONCLUSION In this investigation we examine the deterrence hypotheses of an inverse relationship between the certainty of arrest and conviction and crime rates for states and territories of India, 1967-1973. While recent years have witnessed a number of similar investigations in the United States, with some support found for the deterrence thesis, to our knowledge similar analyses have not been conducted in other countries. We view this as a very unfortunate situation, for as Durkheim pointed out many years ago « One cannot explain a social fact of any complexity (in this case deterrence) except by following its complete development through all social species. Comparative sociology is not a — 44 — particular branch of sociology ; it is sociology itself, in so [ar as it ceases to be purely descriptive and aspires to account for facts ». (Durkheim, 1938, 139). Reminding us of Durkheim’s word and the current state of comparative criminology, Clinard and Abbott (1975) conclude that, American criminologists have been highly ethnocentric in assuming that the claims they make based upon their work largely in one country — the United States — are applicable to other countries. Although comparative sociology has been increasing in recent years, there have been only limited developments in comparative criminology, with comparative work in the area of deterrence being a definite case in point. It is our hope that the investigation reported here will be the first in what will turn out to be a long line of continuing deterrence research in other countries. Like earlier investigations in the United States, the findings pre- sented above provide some support for the deterrence argument. Contrary to our hypotheses, however, a substantial inverse relationship was not found between the certainty of arrest and conviction and offense rates for each crime. Rather, the analysis revealed substantial variation in the relationship between the punishment and rate variables. Consistent with the deterrence hypothesis, the certainty of arrest- offense rate results were found to be in the predicted negative direction for each year for rape, theft, and breaking and entering (burglary). In addition, negative arrest cœfficients were found for five of the seven (71 percent) years for homicide, and four of the seven (57 percent) years for robberv. The size of the negative cœfficients was found to vary substantially. however, both between and within offense categories. In part, this variation is not all together unexpected for similar investigations in the United States have also shown certainty of arrest not to have a uniform deterrent effect on all types of crimes. In addition, deterrence theory itself suggests that the threat of legal sanctions Will not have the same etffect on all offenses. To the contrary, early as well as later proponents of punishment argue that « expressive acts » like murder, assault, rape, etc., are much less subject to the law than « instrumental crimes » like theft, robbery, burglary, shoplifting, etc. Our findings for arrest are generally consistent with this argument, with the size of the negative cœfficients being generally larger on the average for theft, burglary, and robbery than for murder and rape. (Again, however, over half of the results for robbery are positive). This pattern of findings suggests that in India, as in the United States, arrest provides a less effective deterrent to crimes against persons than crimes against property. — 45 — The results of the analysis for certainty of conviction reveal a very different pattern. For theft, where the arrest results are most consistent with our hypotheses, the certainty of conviction-offense rate rela- tionship is positive for each year, with the cœfficients being quite substantial in some cases. Somewhat similarly, for burglary, where the arrest resuits arc also very consistent with the deterrence hypothesis, the conviction-rate cœfficients are either positive or low-negative. On the other hand, for robbery, the arrest and conviction findings are more comparable, with only two of the conviction cœfficients being negative, and the positive cæœfficients being on the whole larger in size. Finally, for murder and rape, the majority of the conviction results are negative, as is the case with arrest. For each year, however, the negative convic- tion cœfficients are very slight, thus indicating the certainty of conviction to only have a very slight deterrent effect on rape and murder. These findings for conviction arc in striking contrast to what deterrence theory would lead us to expect, and the results of previous investigations in the United States. While deterrence investigations in the United States have generally not focused upon convictions directly, but rather upon arrest and imprisonment, our negative findings for convictions for India is indeed surprising in light of : (1) the rather consistent results found for arrest — a necessary requirement for con- viction — for this country, and (2) the rather consistent arrest and imprisonment results found in studies conducted in the United States. At least two possible explanations might account for why the findings for conviction are so inconsistent with the deterrence hypo- thesis. First, it might be speculated that the level of convictions for crime in India is not sufficient to achieve a substantial deterrent effect. As Tittle and Rowe (1974) found in their investigation of arrest for cities and counties in the state of Florida, certainty levels had to exceed a 30 percent level before a substantial deterrent effect became evident. Second, it might also be speculated that certainty of conviction fails to deter in India because even upon conviction, the sanctions that result are not severe enough to discourage crime. In this regard, it must be kept clearly in mind that the certainty and severity of legal sanctions are not independent factors. To the contrary, sanctions, regardless of their level of severity, can only result if there is certainty : i.e, legal sanctions can only be applied upon conviction. Likewise, regardless of the certainty of conviction, if the sanctions that follow are not sufficiently severe, the threat of conviction and the law will be robbed (at least to some degree) of its deterrent effect. In an attempt to explore the first of these possibilities, average certainty of conviction levels were computed for each offense by taking the ratio of reported crimes to convictions for the period 1967- — 46 — 1971. This analysis resulted in the following percentage conviction figures : homicide (35.3 percent), rape (23.8 percent), robbery (16.7 percent), burgiary (18.4 percent), theft (22.9 percent). Unfortunately, comparable estimates cannot be computed for the United States for these crimes because the F.B.I. only reports in the Uniform Crime Reports conviction figures for some cities of this country (which include only about one half of the United States population), and not the entire nation (24). Nor does Orsagh (1973), in his study of convictions in California counties and cities, provide us with conviction levels for the major felonies (25). As a result, it is simply not possible at this time to compare conviction levels in India and the United States, nor to address the question oï how levels of conviction in India may not be sufficieniiy high to adequately deter crime (26). Such an analysis will have to await conviction data that are currently not available. In order to explore the second possibility noted above, that upon conviction the sanctions received may not be severe enough to ade- quately deter crime, it would be necessary to examine sentencing patterns for each crime, Here again, such an analysis is not immediately possible for we have not been able to secure the necessary sentencing and imprisonment figures in order to explore this possibility. Às a result, this question too must be left to future investigators. Clinard and Abbott’s (1973) report that the vast majority (85 percent) of persons imprisoned in India are released within a minimum of six months, does however, suggest some promise in this line of inquiry (27). In addition 10 the possible effect of the above two factors, at least two more considerations must be taken into account as possible expla- nations of our generally negative findings for the certainty of conviction. First, and along theoretical lines, proponents of deterrence argue that if legal sanctions are to have a substantial deterrent eflect, punishment must be certain, severe, and prompt. While the former two dimensions of punishment have received much attention in studies in the United States, the celerity factor has been ignored in the empirical literature. Nor have we considered this aspect of punishment in this (24) Nor it is possible to compile from the F.B.l's report felony figures for the portion of the United States when conviction data are available in order to compute convicion rates comparable to those for India. (25) In addition, there are no nation-wide court statistics that would allow us to compute a conviction rate like that used for India. (26) The theoretical literature on deterrence also unfortunately fails to provide us with much insight in this regard. Discussions of this matter are no more specific than to suggest that punishment must be « certain » to be an ceffective deterrent (Zimring and Hawkins, 1973 ; Gibbs, 1975). (27) Unfortunatelv, they fail to provide length of imprisonment figures by type of offense. As a result, it is not possible to compare such figures with data that are available for states of the United States. ue AÉ ce investigation due to the necessary data not being available. As a result, we can only speculate at this point how our findings for conviction, as well as arrest, might have been affected by our not incorporating this variable into the analysis. This question too will have to await future investigation when the necessary celerity data become available. Second, and along methodological lines, it might be argued that the police figures for crime, arrest and conviction used in this analysis are not sufficiently refined in order to conduct a proper investigation of deterrence. As noted earlier, Chiricos and Waldo (1970) raise this very question about using available police and prisoner figures in conduc- ting deterrence investigations in the United States (28). Although it is not possible to provide a conclusive answer to this last question, this possibility must be taken very seriously. As reported earlier, in examining the hypotheses for arrest and conviction, we observed substantial ycarly variation in the size, and in many cases the sign, of thc certainty-offense rate results. This variation may well reflect error, or noise factors, in the figures being examined, with the deterrent effect of arrest and conviction being much more substantial than the data reflect. Future investigations such as our’s for later years, where police data will likely be less subject to possible error factors, will hopefully shed some additional light on this question. In conclusion, we have examined here an important, but neglected area in comparative criminology ; this being the deterrent effect of arrest and conviction on crime rates in a country other than the United States (29)., Due to a number of possible theoretical and methodological limitations of this investigation, our analysis failed to provide conclusive support for the deterrence hypothesis. The extent to which our mixed findings reflect a lack of deterrence, or the effect of the limitations noted above, however, remains unclear. It is our hope that future investigators will treat each of the possible limitations we have noted as questions worthy of further research. Only through continuing efforts along these lines will we gain a better understanding of the deterrent effect of legal sanctions. (28) Compared to other developing countries, Clinard and Abbott (1973) rank the state of crime statistics in India verv favorably. Their discussions of this topic is very brief and general, however, and doesn’t allow us to adequately assess the use India police figures in investigations such as ours. (29) It is of interest to note that in Clinard and Abbott’s (1973) recent survey of comparative criminologv and crime in the developing countries, no attention is given to the matter of deterrence, nor the void of comparative investigations in this area. — 48 — REFERENCES Punishment and deterrence, Ann Arbor. The University ANDENAES (J.), of Michigan Press, 1974. BAILEY (W. C.) : « Certainty of Arrest and Crime Rates for Major Felonies A Research Note », in Journal of Research in Crime and Delinquency, July 1976 (n° 13), pp. 145-154. « Murder and Capital Punishment : Some Further Evidence» , in American Journal of Orthopsychiatry, July 1975, (n° 45), pp. 669-688. « On Punishment and BAILEY (W.C.), GRAY (L.N.), and MARTIN (J.D.) Crime (Chiricos and Waldo) : Some Methodological commentary », in Social Problems, 1972, (n° 19), Fall, pp. 284-289. BAILEY (W.C.), MARTIN (J.D.), and GRAY (L.N.) : « Crime and deterrence : A Correlation Analysis », in Journal of Research in Crime and Delinquency, 1974, 11, (n° 2), pp. 124-143. BEDAU (H.A.) : The Death Penalty in America, Revised Edition. New York, Doubleday-Anchor Books, 1967. BEDAU (H.A.) and CHESTER (P.) (eds.) Capital Punishment in the United States, 1976, New-York, Ams Press Inc. BUREAU OF POLICE RESEARCH and DEVELOPMENT, MINISTRY OF HOME AFFAIRS. GOVERNMENT OF INDIA. Crime In India, 1973. New Delhi, India. 1975. Crime In India, 1972, New Delhi, India 1974 In India 1973 Crime India, 1971. New Delhi, Crime In India, 1970. New Delhi, India 1972 Crime In India, 1969. New Delhi, India 1971 Crime In India, 1968. New Delhi, India 1970 Crime In India, 1967. New Delhi, India 1969 CHIRICOS (T.) and WALDO (G.) : « Punishment and Crime : An Examination of some Empirical Evidence», in Social Problems, Fall 1970, (n° 18), pp. 200-217. CLINARD (B.) and ABBOTT (D)) : Crime in Developing Countries. New York, John Wiley and Sons, 1973. DURKHEIM (E.) : The Rules of Sociological Method. Chicago, University of Chicago Press. 1938. EHRLICH (I) « The Deterrent Effect of Criminal Law Enforcement », in The Journal of Legal Issues, 1972, 1 (n° 2), pp. 259-276. GIBBS (J.P.) : Crime, Punishment and Deterrence. New York, Elsevier, 1975. « Crime, Punishment and Deterrence», in Southwestern Social Science Quartely, March 1968, (n° 48), pp. 515-530. GRAY (L.N.) and MARTIN (J.D.) : « Punishment and Deterrence : Another Analysis of Gibbs’ Data » in Social Science Quarterlv, September 1969, (n° 49), pp. 289-295. — 49 — KUH (E.) and MEYER (JR) « Correlation and Regression Estimates When Data Are Ratios », in Econometrica, October 1955, (n° 23), pp. 400-416 LOGAN (C.H, « Arrest Rates and Deterrence », in Social Science Quarterly, December 1975, (n° 56), pp. 376-389 « General Deterrent Effect of Imprisonment ». in Social Forces, September 1972, (n 51), pp. 64-73 McNEMAR (Q.), Psychological Statistics 1969, 4th Ed.…. New York, John Wiley and Sons MINISTRY OF PLANNING, GOVERNMENT OF INDIA Statistical Abstract India 1973 New Delthi India 19775 Statistical Abstract India 1972 New Delhi India 1974 Statistical Abstract India 1971 New Delhi India 1973 Statistical Abstract India 1970 New Delhi India 1972 Statistical Abstract India 1969 New Delhi India 1971 Statistical Abstract India 1968 New Delhi, India 1970 Statistical Ahstract India, 1967. New Delhi, Ïnd1a 1969 ORSAGH (T) : « Crime, Sanctions and Scœnt1fic Explanation », in Journal of Criminal Law and Criminology, 1973, 64 (n° 3), pp. 354-361. « An Economic Analysis of the PHILLIPS (L.) and VOTEY (HLL. r. Deterrent Effect of Law Enforcement on Criminal Activity» in Journal of Criminal Law Criminology and Police Science 1972, 63 (n 4) pp. 330-342. RANGARAJAN (C.) and CHATTERIEE (S.) : « À Note on Comparison Between Correlation Cœfficients of Original and Transformed Variables » in American Statistician, October 1969, (n° 23), pp. 28-29. Capital Punishment New York. quper and Row SELLIN (T.), (ed.) Publishers, 1967 SJOQUIST (D. « Property Crime and Economic Behavior Some Empirical Results » in The American Economic kReview June 1973 (n 63) —— pp. 439-446. TITTLE (R « Crime Rates and Legal Sanct10ns ». in Social Problems Spring 1969, (n° 16), pp. 409-423. TITTLE (R.) and LOGAN (C) « Sanctions and Deviance Evidence and Remaining Questions », in Law and Society Review, Sprmg 1973, (n° 7), pp. 371-392. TITTLE (C.) and ROWE (A.R.) « Certainty of Arrest and Crime Rates A Further Test of the Deterrence Hypothesis », in Social Forces, June 1974 (n° 52), pp. 455-462 ZIMRING (F.E.) and HAWKINS (G.) : Deterrence : The Legal Threat in Crime Control, Chicago. The University of Chicago Press, 1973 — -50 — THE INTERNATIONAL SITUATION OF RESEARCH AND LEGAL REFORM WORK IN THE FIELD OF ECONOMIC AND BUSINESS CRIME* KLAUS TIEDEMANN University of Freiburg Freiburg i. Br. Federal Republic of Germany. Public and scientific interest in white collar crime, especially in economic and business crime, has considerably increased in the last few years. Certainly, French scholars, for example, have long since taught « droit pénal des affaires » in their university courses, and « diritto penale commerciale » or « derecho penal economico » is a well treated subject of Italian and Spanish speaking scholars (1). Àlso, periods of economic crisis usually produce a special economic law in most countries, with some of these legislations surviving the times of crisis and serving as a base for future legislation and law enforcement. But it is only in the last several years that a more or less general interest in the integrity or non-integrity of the whole world of business has spread all over the western world (2). (1) As to the French situation, see Delmas-Marty, Droit pénal des affaires (1973) ; Larguier, Droit pénal des affaires (2nd ed. 1975) ; Dupont-Delestraint, Droit pénal des affaires (1974) ; for Italy, see Conti, Diritto penale commerciale, 2 vol. (1965‘67) ; Pedrazzi/Da Costa, Tratado de Direito Penal Economico vol. 1 (1973 - also to Brasilian Law) ; for Spain, cp. Bajo Fernandez, El Derecho Penal y Ciencias Penales 1973, 91 f£. with further references. (2) Some examples — for Austria Liebscher, Die Wirtschaftsdelikte im Gsterreichischen Strafrecht, Zeitschrift für die gesamte Strafrechtswissenschaft 1976, 261 ff. and Tintner, Organisation und Arbeitsweise der Exekutive bei der Bekämpfung der Wirtschaftskriminalität in Osterreich, in : Schäfer (ed.), for Wirtschaftskriminalität/Weisse-Kragen-Kriminalität (1974) pp. 513 fi 9 Belgium Kellens, Le crime en col blanc sa place dans une criminologie économique, Revue de Science Criminelle 1974, 807 ff. ; for France : Cosson, Les délits en matière de banques et d’établissements financiers, Revue de Science Criminelle 1973, 1 ff. ; Delmas-Marty, La criminalité d’affaires, Revue de Science Criminelle 1974, 45 ff. ; Frilloux, La fraude commerciale, Revue Internationale de Criminologie et de Police Technique 1972, 36 ff. ; for the Federal Republic of Germany : Kaiser, Kriminologie (3rd ed. 1976) pp. 283 ff. ; — 51 — For the United States, Watergate seems to have been a kind of symbol for a broad inquiry into the behavior of people and institutions heretofore believed sacrosanct, and the Lockheed affair led to a wide national interest, especially in corruption and bribery. The Lockheed and Exxon scandals in Italy, the Netherlands and Japan revealed at the same time that even outside North and South America politicans are much more involved in business than it was previously believed in countries where generally politicians do not become politicans in order to earn much money. The intensive activity of the Securities and Exchange Commission and a remarkable number of civil suits against big corporations then disclosed large corporate frauds in the U.S., which go side by side with « traditional » illegal and criminal business patterns such as anti- trust offenses and tax fraud (3). In Europe, there were not such particular events like Watergate to contribute to explanations of or speculations about the increasing interest in economic and business crime. Nevertheless, the European press records no fewer business scandals than in the U.S., with the phenomena of business delinquency being much the same — tax fraud, anticompetitive behavior, corporate frauds, and, especially, securities Tiedemann, Erscheinungsformen der Wirtschaftskriminalität und Môglichkeiten ihrer strafrechtlichen Bekämpfung, Zeitschrift für die gesamte Strafrechtswissen- schaft 1976, 231 ff… and : La criminalité d’affaire dans l’économie moderne, Revue Internationale de Criminologie et de Police Technique 1975, 147 ff.; for the Netherlands : Schaffmeister, Das niederländische Wirtschaftsstrafgesetz, Zeits- chrift für die gesamte Strafrechtswissenschaft 1973, 782 ff. ; for the United Kingdom Borell and Cashinella, Crime in Britain Today 1975 pp. 77 ff. ; Mack, The Crime Industry 1975 pp. 21 ff. ; for Switzerland Burger, Wirtschaftskriminalität 1977 ; Graven, L’Economie du Droit Pénal et le Droit Pénal Economique, Revue de Droit Pénal Suisse 1976, 337 fîf. ; for the United States of America : Edelhertz, The Nature, Impact and Prosecution of White- Collar Crime 1970 ; Geis (ed.), White-Collar Criminal (1968), and the articles forming the symposium in American Criminal Law Review 1973, 817 ff. (3) See Herlihy/Levine, Corporate Crisis : The Overseas Payment Problem, Law and Policy in International Business 1976, vol. 8 pp. 547 ff. ; Hermann, Conspiracy, The Business Enterprise, White Collar Crime and Federal Prosecution, Creighton Law Review 1976 vol. 9 pp. 647 ff. ; Mathews, Criminal Prosecutions under the Federal Securities Laws and Related Statutes, George Washington Law Review 1971, vol. 39 pp. 901 ff. ; Annual Report of the S.E.C. for the Fiscal Year ended June 30, 1975 (1976) pp. 45 ff., 65 fi. ; Public Citizen Staff Report : White-collar Crime January 1973-June 1974 (1974) pp. 8 ff 16 ff. ; Blum/Kaplan/Halleck, Transnational Crime (L.E.A.A.-manu- script 1976) ; Hey/Kelley, An Empirical Survey of Price Fixing Conspiracies, Journal of Law and Economics 1974, vol. 17 pp. 13 ff. ; Stone, Where the Law Ends : The Social Control of Corporate Behavior 1975 ; United States Department of Justice (ed.). Multinational White-collar Crime 1976, pp. 44 fif. — 52 — fraud and corporate embezzlement (4). The activity of multinational corporations, especially in the Third World countries, has finally stimulated not only national interest in the Third World, but now has induced the United Nations to put the « Crimes through Misuse of Power, especially Economic Power » on their long-term program and make it a subject for the next Crime Prevention and Criminal Justice congress in Melbourne 1980. Until today, scientific research in this field has not corresponded to this vast public and political interest. Certainly there have been many articles and theories on the impact of economic and business crime though mostly these have been based on a small empirical background and directed toward a very restricted sector of business life. One can even say that since Sutherland there has been no broad empirical approach to the subject in the U.S., where the problem of white collar crime has, until about 10 years ago, been investigated only by socio- logists. For an outsider it is interesting to see how, in the past a few big cases — such as the Philadelphia Electrical equipment case or the Equity Funding Corporation case — have generated a vast literature which has never made a broad inquiry into the relation of these big cases to normal business life, and never examined the subtle practices of circumvention or the empirical effect of sanctions which, in the form of fines are often paid not by the culprit, but by the firms, and are subsequently reflected in increased prices, deducted from the income tax, and so on. Yet, as understood by European scholars for the last decade or so, American economists, by focusing on the economics of crime and punishment, begin to offer a valuable perspective on our subject. One of the factors responsible for this new found interest is that, in wide- spread opinion of economists and econometrics criminologists have not provided a systematic framework for analyzing criminal activity ; (whereas criminologists may think that economists have not managed the problems of economy). The choice theoretic model of microecono- mics, e.g., is believed, by scholars such as Becker (1968), to afford a par- ticular useful structure for such an analysis (5). The choice to commit a crime becomes thus a special case of the models of economic choice (4) Synopsis in Tiedemann, Phenomenology of Economic Crime, in : Council of Europe (ed.), Criminological Aspects of Economic Crime (1977). (S) Becker, Crime and Punishment : An Economic Approach, Journal of Political Economy 1968, vol. 76 pp. 169 ff£. ; see also Becker/Landes (ed.), Essays in the Economics of Crime and Punishment (1974). As an example for tvpical criminological studies in econometrics see Rose-Ackermann, The Economics of Corruption, Journal of Public Economics 1975, vol. 4 pp. 187 ff. A summary concerning the American point of view to that field of research is given by Hermann, Economic Analysis of Law (Book Reviews), Washington University Law Quarterly 1974, 354 ff. theory which hypothesizes that all individuals respond to incentives, and if the costs and benefits associated with an action change, then the agent’s choices are also likely to change. In this view, perhaps too rational in its general approach, the decision to commit offenses is reached by an egocentric cost-benefit analysis, with the expected benefits and costs containing both monetary and psychological elements. Noting just the recent books of Becker and Landes (1974), Posner (1975), McPheters and Strange (1976) or Kaplan and Kessler (1976), this literature includes into its economic approach economic and business delinquency, with the earlier papers of Allingham and Sandmo (1972), Kolm (1973), and Singh (1973) addressing the very questions of income tax evasion (6). In Europe, there has been, up to now, no corresponding approach as far as property and other business crimes are concerned. Particularly, nearly all German economists do not touch on the dirty subject of business crime at all (7), and even governmental agencies of trade and industries, as well as teachers of economic law, seem to consi- der business crime a neglectable quantity or a phenomenon incompara- ble with « real » crimes sanctioned with imprisonment. In Europe, mo- reover, research on the subject has been slight and only very recent, and most articles, written mainly by lawyers and sociologists, lack the benefit of practical economic experience, often replacing it with an ideological slant. The number of relevant articles does demonstrate, nevertheless, the increased interest of scholars. Thus, to give just one representative example, one-third of the 1974 volume of the well- known « Revue de science criminelle et de droit pénal comparé » (Paris) consisted of articles on economic and business crime. By the way, French criminological and law enforcement discussion on the subject was initiated by Professor Cosson’s book on tax fraud (1971), which demonstrates the techniques of VAT fraud, with its fake exports, fictitious bills and dummies — a typical modus operandi for many actual patterns of fraud, e.g., on the level of the Furopean Com- munity (E.C.) (8). Income Tax Evasion A Theoretical (6) Allingham/Sandmo, Analysis. Journal of Public Economics 1972, vol. 1 pp. 323 ff. ; Kolm, A Note on Optimum Tax Evasion, Journal of Public Economics 1973 vol. 2 pp. 265 ff. : Singh. Making Honesty the Best Policy, Journal of Public Economics 1973 vol. 2 pp. 257 ff. (7) Some of the few exceptions are Molsberger, Wirtschaftskriminalität und Wirtschaftsordnungen, Wirtschaftspolitische Chronik 1976. vol. 25 pp. 179 ff and Zybon, Wirtschaftskriminalität Aals gesamtwirtschaftliches Problem (1972) as well as the authors in number 5 of the review Betriebswirtschaftliche Forschung und Praxis 1975. (8) Cosson, Les Industriels de la Fraude Fiscale (1971) ; cf. also Cosson. L’Etat Victime du Crime Organisé, Revue de Droit Pénal et de Criminologie 1973/74, 237 f. — 54 — The E.C. criminality, mainly caused by the economic system of customs and -subsidies seperating the E.C. member states from the lower world trade levels, was then studied by a group of scholars appointed by French government and directed by Professor Delmas-Marty (9), who also acted as a reporter. on business crime at the National French Congress on Cnmmology in 1975. (10). _ There are additional recent studœs like that of Professor Delmas- Saint-Hilaire’'s Group on business crime in the Bordeaux area (11) or a new book of Professor Cosson on fraude in the field of banking (12) In Belgium Professor Kellens pushed economic criminological theo- ries forward and published a comprehenswe study on bankruptcy, including experiences of the U.S. in the field of organized (planned) bankruptcy (13). In West Germany, both practical and theoretical discussions and measures were stimulated by the German Lawyers’ Day 1972, which recommended, on the basis of my own expertise, a thorough analysis oË the economic and economic criminal law system with regard to business crime, the reorganization of law enforcement authorities, the amendment of statistical recording of the types of business crime, and other activities (14). One of the results of these recommendations was that in the same year the Federal Ministry of Justice introduced a Commission for Combatting Business Crime. This commission has, since then, in view of amending civil law, criminal law and law enfor- cement, done an outstanding job of research into many subjects of business life (law of competition, tax and subvention law, usury, bank ruptcy, credit affairs, bogus firms — and corporate law, computer crime bribery and corruption, frauds in the building industry, misuse of insider informations, frauds and abuses concerning securities, cheques bills, cheque and credit cards The first recommendations of this Commission led to the so- called First Law for Combatting Business Crime (1976), introducing (9) Université de Lille II, Institut de Criminologie et des Sciences Criminelles (ed:), La Délinquance d’Affaires dans le Cadre des Règles Communautaires - Rapport Final (1976). (10) See Revue de Science Criminelle et de Droit Pénal Comparé 1974, 465 f. (11) Université de Bordeaux I, Institut de Sciences Criminelles (ed.), La Criminalité des Affaires dans la Région Bordelaise (1975). (2) La Monnaie des Autres (1977). (13) Kellens, Banqueroute et Banqueroutiers (1974), see also Kellens, Ten- dances actuelles du droit de la banqueroute, Revue de Droit Pénal et de Criminologie 1972, 1047 fif. 14) Verhandlungen des 49. Deutschen Juristentages 1972, vol. 2 pp. M 201 ff — 55 — new civil and criminal provisions on usury, bankruptcy, credit fraud, fraudulent procurement of subsidies to business and industry (14 a). À drafît of a Second Law is planned for the end of 1977. In the academic sphere, there have long been come smaller studies, most of them being doctoral thesis, guided by Professor Geerds on subjects such as bankruptcy, fraud against insurance companies, falsi- fication of products (food and wine) (15). A rather broad investigation into all criminal procedures, handled by the permanent special prose- cutors for economic criminal affairs, introduced in Germany about six years ago and imitated in France since 1976, is being run by Professor Kaiser, of the Freiburg Max Planck Institute ; preliminary reports of this criminological research have been made public in 1976 and 1977, based on all economic criminal procedures opened in 1974 and 1975, and studying the concerned branches of trade and industry, the law form of the relevant corporations, the material damage and other aspects (16). My own studies in this field began with a study, done for the Federal Ministry of Justice and for the above mentioned Commission, on fraudulent obtaining of subsidies in all sectors of social and economic life, fom public welfare to subsidies on the E.E.C. level and to aids for the developing countries (17). Another study of mine, also done for the Ministry of Justice, was on criminal policy in the field of antitrust offenses, including questions of whether to transform the most severe regulatory offenses into criminal offenses and of what alternatives to criminal sanctions could reasonably be recommended, also considering the experiences and practices of the U.S., of France, and of the E.E.C. (14a) For an Italian translation of this Law see Lanzi, L’Indice Penale 1976, 470 ff. (15) Landry, Inverkehrbringen und Herstellen gesundheitsschädlicher Gegens- tände, Dissertation Kiel 1966 ; P. Schmidt, Die trügerische Reklame, Disser- tation Kiel 1965 Schad, Betrügereien gegen Versicherungen, Dissertation Kiel 1965 ; Hammerl, Die Bankrottdelikte, Dissertation Frankfurt 1970 : Sievers, Bestechung und Bestechlichkeit von Angestellten, Dissertation Kiel 1963 (c£. also Geerds, Einzelner und Staatsgewalt im geltenden Strafrecht, 1969, pp. 64 ff.). (16) Berckhauer, Forschungsbericht über die Bundesweite Erfassung von Wirtschaftsstraftaten nach einheitlichen Gesichtspunkten im Jahre 1974 (1976), Berckhauer/Rada, Forschungsbericht über die Bundesweite Ergassung usw. im Jahre 1975 (1977) ; see also, Kaiser, Kriminologie loc. cit., p. 301 f. (17) Tiedemann, Subventionskriminalität in der Bundesrepublik (1974) ; see also Tiedemann, Kriminologische und kriminalistische Aspekte der Subvention- serschleichung, in : Schäfer, loc. cit. (footnote 2), pp. 19 ff. ; Tiedemann, La Fraude dans le Domaine des Subventions : Criminologie et Politique Criminelle, Revue de Droit Pénal et de Criminologie 1975/76, 129 ff. — 56 — Commision (18). Besides a smaller study on credit and bank frauds (19), I have guided further studies on the deterrent effect of criminal sanctions, especially in the field of subsidy and tax frauds (20), on stock market frauds (21), on computer crime (22), boycott conspiracy (23) and wine fraud (24). This short survey on recent and current European research would be incomplete without mentioning the activities of the Council of Europe. The Division of Crime Problems organized the 1976 conference on the criminological aspects of economic crime (25) and a special subcommittee of the Council is starting a long term program to stimulate research in the field of economic and business crime, with the final aim of harmonizing the legislations of member nations against delinquent business behavior. I do not intend to give a detailed survey on the results of all these studies and especially those of my own. These works cover a rather wide and heterogenic field of subjects, the problems of which can only partly be compared with phenomena and problems of other countries. Thus by study on fraudulent procurement of subsidies, based on 850 original court and administrative authorities’ records, focused mainly on specific deficiencies of the E.C. legislation and of German admin- istrative and criminal law provisions. Nevertheless we could also make it clear that the very process of granting, especially of non-reimbursable subsidies, is a primary criminogenic factor, while the regular and natural control employed by every business, that is the need to make recom- pense, is absent. Thus, artificial means of social control have to be introduced, among which we proposed a licencing system modeled on parts of German tax legislation. As to macroeconomics, from the point of view of fighting against crime, it must be regretted that lost contri- butions are increasingly taking the place of subsidized loans, which generally are well controlled by banking agencies. (18) Tiedemann, Kartellrechtsverstôsse und Strafrecht (1976) b cf. also Tiedemann, Wettbewerb und Strafrecht (1976) ; Tiedemann, Wirtschaftsstra- frecht und Wirtschaftskriminalität 1976, vol. 2 pp. 14 ff, 225 f. (19) Tiedemann/Cosson, Straftaten und Strafrecht im deutschen und fran- zôsischen Bank - und Kreditwesen (1973). (20) Breland, Lernen und Verlernen von Kriminalität (1975), and more detailed Breland, Präventive Kriminalitätsbekämpfung, Dissertation Giessen 1974 (see also Tiedmann, in : Verhandlungen des 49. Deutschen Juristentages vol. 1 p. C 40 Îf.). (21) Scheu, Das Bôrsenstrafrecht und seine Reform, Dissertation Giessen 1974 (see also Tiedemann, Wirtschaftsstrafrecht vol. 2 pp. 137 ff.). (22) Sieber, Computerkriminalität und Strafrecht (1977) (see also Tiedemnann, Wirtschaftsstrafrecht vol. 2 pp. 148 ff). (23) Simmler, Wirtschaftsboykott und Strafrecht, Dissertation Freiburg 1977 (see also Tiedemann, Kartellrechtsversôsse pp. 124 fif. ). (24) Henssen, Weinkriminalitäit und Strafrecht, Dissertation Giessen 1976 (See also Tiedemann, Wirtschaftsstrafrecht vol. 2 pp. 158 ff.). (25) See footnote 4 — 57 — Instead of giving further details o my own research work, let me just give a short survey on those studies which might be the closest related to the actual international public and scientific discussion on deterrence, especially in the field of antitrust offenses, and then sum- marize, and illustrate by examples, my methodological approach and the aim of my work — especially because I think that I share the approach and the aim with several other European scholars such as Cosson and Delmas-Marty, Kellens and Pedrazzi (Italy). As far as deterrence is concerned, my studies have focused on general deterrence, which is particularly important regarding law enforcement, the sense of punishing and the number of undetected crimes (« dark numbers »). The heavy obstacles to research into general deterrence result from the fact that general deterrence does not look at captured, but at potential offenders, so that the relevant data cannot be found in court or prison records. À study of general deterrence, then, examines why and by what psychological factors people (business- men) do not commit (business) crimes. Our research on this question, more or less a pilot study, financed by the Deutsche Forschungsgemein- schaft (German National Science Faundation), was, in the footsteps of Sutherland and Cressey, put into the framework of the developed learning theory, saying that criminal behavior is mainly learned beha- vior, the learning process working by experience or by model. Our rather long and very detailed questionnaire contained projective models, i.e. cases taken from original court files, asking the businessmen ad- dresses to hypothetically identify themselves with business crimes, under a selected choice of variables. The questionnaire, established with the assistance of a well-reputed German polling firm, considered a multi- factorial group of adverse factors such as the probability of criminal prosecution, civil and administrative sanctions, publicity of prosecution and sentencing, attitudes of the social environs, inner resistance (con- science), and so on. The criminal sanctions were subdivided into fines of different levels and imprisonments of different length, with and without parole. The answers of about 70 % of the addressees led to the simply put result that, at least in the field of tax fraud and fraudulent pro- curement of subsidies, only probable criminal prosecution, sentences of (short term) imprisonment without parole, and closing of the firm represent effective means of general deterrence. This study could not strengthen the belief of the German Federal Cartel Office that making the prosecution and the fining of corpora- tions public produces a particularly deterrent effect. This caused us, in our Work on antitrust offenses, to reject the institutionalization of publicity and — in accordance with the statements of the new head of the U.S. Justice Department’s antitrust divisions or, more generally, with the results of Chambliss, Zimring and others, and contrary to what many American economists think — to recommend short term —— 58‘ —s imprisonment: and not only high fines for severe cases of restraint of trade. For the. Federal Republic of Germany, France, the E.F.C. and even Austria, we found that the classic concerted pratices concerning price fixing, the sharing out of production and of marketing patches and, in particular, inadmissible agreements on public tenders especially in the building and transport industry as well as in the sector of electrical equipment still exist, but that an increasing number of these cartels no longer take the traditional form of written agreements. They take the form of concerted practices which, however, under German, French and E.E.C. legislation, are scarcely less criminal than written provisions. In particular the oligopoly form of market often makes the express conclusion of agreements superfluous, since the small number of competitors allows a « community of reaction » to arise. The result of this is that an individual competitor cannot bring his fesources into play without considering his fellow competitors. In this instance, understandings on quality, service, and other competitive resources can often have greater significance than agreements on prices and quotes. This led us to recommend a general clause for future German legislation on criminal restraint of trade, instead of the description of single forms of restraint of trade which would fit better into the continental pattern of criminal law-giving. We also found that the legal possibility of consent decrees (transactions), especially in French antitrust law. weakens general deterrence and that the many exceptions in German antitrust provisions make it difficult to prove subjective fault of the culprit in individual cases at the same time preventing the general public from becoming aware of the basis ihjuriousness of limitations on competition. Before turning to the aims and methods of criminological research into economic and business crime let me just say a few words to a most disputed question : the definition of business crime (or economic crime). This question plays an important role in the discussion since Sutherland, especially as in the U.S. the notion of business crime overlaps with the notions of white collar crime, organized crime, consumer frauds and pollution offenses. It is astonishing that despite decades of research work in this field, there is no agreement either about the concept of economic delinquency or the notion of economic criminal law. The different opinions depend on different viewpoints of the respective scientists. They chiefly can be traced back to two different methodological roots One inclines towards the empirical sciences and focuses on the criminological aspects of economic offenses. Another leans towards a normative way of looking at the topic and stresses its criminal law aspects (26). (26) More detailed - as well as to the following - Tiedemann, El Concepto de Delito Economico y de Derecho Penal Economico, Nuevo Pensamiento Penal 1975, 461 ff. — 59 — Emphasizing the criminological, and especially the sociological, as- pects of economic offenses, originates from the work of Sutherland who, in 1949, described a white collar offense as a crime commited by a person of respectability and high social status in the course of his occupation (and. in 1959, as a crime committed by a person with high socio-economic status who violates the laws designed to regulate his occupational activities). After Sutherland, Clinard, Quinney and other American criminologists termed « occupational crime » those offenses of employees against their employers or of individuals for themselves in the course of their occupation. They reserved the term « corporate crime » for offenses committed by corporate officials for the corporation or by the corporation itself. Other criminological and especially criminal-scientific researchers both in the U.S. and in the Federal Republic of Germany stress the abuse of confidence in economic relationships as the necessary foun- dation of economic offenses. Some chiefly criminal-scientific orientated scholars declare that economic delinquency does not represent a sepa- rate category of offenses at all, but only refers to certain property crimes which are especially difficult to prove. This definition corresponds partly to the German law of jurisdiction (Sec. 74 c) and also to the French concept of criminalité d’affaires, which looks specifically upon business as an instrument and object of delinquent acts. If part of the French literature separates bogus firms which have been constituted with the intention of committing crimes from criminalité d'affaires, then it seems however to go too far, since even the most dangerous economic delinquents and serious delinquencies would then be excluded from the definition. Different from these criminological and criminal-scientific theories, the normative way of looking at the topic stresses its criminal law aspects and looks at the object of the crime more than the criminal. According to this approach criminal offenses are acts, the effects of which disturb or endanger economic life or the economic system above and beyond injury to individual interests. This definition offers a possibility to differentiate between a narrow and a wide concept of economic crime and shows that the various definitions of economic crime do not differ too much in their result. Both according to criminology and to criminal law, economic crimes in a narrow sense are beyond all doubt such prohibited acts which injure or endanger the protected economic interests of the community. As examples one can cite criminal actions in the domain of foreign trade law, law of market and of price regulations, law of competition, tax law and law concerning procurement of subsidies. Although there are some doubts from the criminological point of view regarding the respective delinquents, economic crime in this narrow — 60 — sense also includes fraud and breach of trust against institutions of economic life, especially banks and insurance companies. It is also possible to include the so-called special offenses both from the crimi- nological and criminal law standpoint. Special offenses, which in Italian are called with the ferminus technicus « delitti propri » and in German « Sonderdelikte » are offenses which cannot be performed by every person but only by persons who are qualified in a special manner. Thus, professional positions such as businessman, clerk or founder or member of the supervisory board of a trading compagny are prerequisites for punishment in a number of legal systems, especially concerning the criminal law relating to trading compagnies and bank- ruptcy offenses. Economic crimes in a wide sense are such prohibited acts which primarily injure individual interests but which do this by abusing instruments of economic life and thus consequently damage economic interest of the community too. Examples of this are offenses concerning insurance sheets, manipulation of bills and checks, abuse of credit card facilities and computer crimes. The ideas P’ve just covered include only the main concepts of economic crime. Especially in American sociological and crimino- logical literature, there exists a broad variety of different concepts. In the American, but also in French literature about white collar or economic crime, there is much concern about this problem. In contrast to this, we think the problem of defining the concept of economic crime is of minor significance. We believe people should take a provisionary broad workable definition of the concept and turn to a problem which is more important than delimiting the concept of eco- nomic crime or arguing if a certain form of crime even belongs within its given definition. This more important problem 1s : fighting economic crime. Hereby we arrive at the first topic I announced before to discuss here — namely, to define the aim of criminological research in the field of economic crime. No matter 1if we deal with economic and business crime as scientists, businessmen or politicians, I think we do have to agree, that the aim of each activity in the field of economic crime should be fo fight this form of criminality (27), and this means at the same time to reform the law system and the activities of law enforcement agencies. (27) Cf. Tiedemann, Zièle und Probleme wirtschatskriminologischer Fors- chung, in Festschrift für Richard Lange (1976) pp. 541 (544 ff.) with further references. — 61 — This fundamental statement has important consequences. Ît does not only mean that criminological research shouldn’t be an end in itself. It implies also that the purpose of criminological research in the field of economic crime should not primarily be to verify or falsify certain social or political theories. This probably explains why. most of the modern criminological theories do not look to be applicable in the field of economic crime till now it is not yet perceptible how they could help fighting economic crime. General sociology and general psychology are in a similar situation. In my opinion, psychology and sociology will only gain importance in the field of economic crime, ‘if they leave their general trend and turn to a direct involvement with economic and business crime such as special finance psychology in Europe, which investigates the causes of tax resistance (28), or the sociological systematic theory which can help to define social damage caused by economic crime (29). The consideration that the aim of cr1mmolog1œl research in the field of economic crime should be to fight economic crime, does not only show the direction in which criminological research should not go, but more 1mportant it is able to lead to the methods which should be applied in criminological research in the field of economic crime. As to the methods of criminological research in the field of eco- nomic and business crime, it is evident that they have to be determined by these just stated aims. In consideration of these aims, criminological research in the field of economic and business crime primarily has to explore phenomenology (patterns and forms of delinquent behavior) and etiology (the causes of this behavior). Let me explain these metho- dologœal questions through the sequence of cnmmologœal research works in the field of economic and business crime using some concrete examples |. The starting point of criminological research and of reform work in each scope of economic life is the question, if and to what extent deviant or delinquent behavior xists. This primary question only can be solved by phenomenologic research. The fact that such pheno- menologic inquiry can lead to surprising results can be illustrated by two research works The official reasoning behind the new German law conœrmng stock exchange states that delinquency in.the area of - stock Allgemeine Steuerlehre (4th ed 1965) Schmolders (28) Cf. Schmolders Das Irrationale in der ôffentlichen Finanzwirtschaft (1960), Strümpel, Steuer- Henrion, La moral und Steuerwiderstand der deutschen Selbständigen (1966) Fraude Fiscale. Revue de Droit Pénal et de Criminologie 1973/74, 183 (187 ff.) (29) See Amelung, Rechtsgüterschutz und Schutz der Gesellschaft (1972) in particular to the sociological role theory see Lange, Das Rätsel Kriminalität (1972) pp. 112 ff — 62 — market does not exist. À phenomenologic search managed by our staff in 1973 which was especially based on narrow contacts to insiders of the stock market, showed that although respective criminal proceedings are very rare, there do exist a vast variety of delinquent actions, which partly come under criminal law and partly fail within the so-called grey zone of criminality. As an important example of stock market crimi- nality, the exploitation of advanced insider information can be cited. The misuse of insider information is a crime in certain countries (like the U.S. or France), but not criminal in others (such as Germany or Italy) (30). Another phenomenological inquiry within the area of frau- dulent subventions stated that about 50 % of all scrap import subsidies payed under the EFuropean Coal and Steel Com- munity Treaty to German enterprises were obtained impro- perly. mearly one third of the scrap not existing at all. The given examples can also make the main problem of pheno- menological research in the field of economic and business crime clear, which is getting the necessary data for study. Às in most cases criminal statistics only manifest the penal code provisions infringed by the criminal action and do not describe the action itself, it is very difficult, for example, to find out cases of computer crimes among other « nor- mal » cases of embezzlement or fraud. German criminal statistics however bave been improved in this regard from 1974 on as a consequence of the German Lawyer Day’s recommendations. Further- more, as the just given example of delinquency in the field of stock market shows, analysis of criminal law proceedings is certainly not sufficient. In addition a scientist in the field of economic crime has to become and insider in his respective area, mainly by using interview techniques and by developing personal contacts to real insiders. General official secrecy and particular secrecy relating for example to tax law form additional obstacles in getting the necessary material for study. Phenomenological research with the purpose of answering the question, if and to what extent and in which forms economic and business crime exists, should in any case include the attempt to deter- mine the extent of undetected crimes, the so-called « dark field ». Traditional criminological methods to investigate the dark field do not help much in the field of economic and business crime. So new methods have to be used, taking advantage of the proceedings of the economic administration, of news of the economic press, and of trade statistics. Thus we used Austrian foreign trade statistics in order to show that — p E (30) See Will, Gli « Insiderâ» e i‘R'elativi Problemi nell’Attività delle Borse. Rivista di Diritto Civile 197S, 664 ff. — 63 — whole streams of products which had been subventioned by German authorities because of the pretended export to Austria were sham transactions and fake exports ; or with the help of the grape collection statistics of Rhenania-Paletnat being compared with the selling sta- tistics, we were able to establish approximately the dark number of producers adding sugar to the wine, which is an offense in the E.E.C. when exceeding a certain degree and certain modalities (see Art. 16 and 19 Ordonnance n° 816/70 of the E.C. Commission). 2. Aîter having stated deviant or delinquent acts as well as res- pective data in the field of inquiry specified phenomenologic research is then necessary to answer the second basic question of each inquiry in the field of economic and business crime : the question, if the pre- viously mentioned actions are actually criminal. This question can be answered by a normative and economic method. Let me explain this central methodological problem in more detail. Normative method means that the question if certain actions are crimes or merit criminal sanctions can only be answered by a judicial- normative way of looking at the topic. The decisive criteria answering this question is the social injuriousness of the respective action, that is to say, the extent of social damage caused by it in a certain system. This can be 1llustrated with the example of antitrust offenses : Some cartel agreements, as for example cartels formed to over- come structural crises or for the purposes of rationalisation or specialization, are especially wanted by the economic system. Other cartel agreements, as for example price fixing or inadmissible agreements on public tenders, on the other hand affect the economic system and therefore are forbidden in many states. Nevertheless other states allow cartels to operate or they exercise a mere control over the abuse of cartels or provide penalties only for the violation of registration duties. Thus, the description of cartel delinquency demonstrates the particularly close dependence and relationship of business offenses to the political system. The judicial method however also makes clear that there are considerable difficulties in describing severe antitrust offenses in material law and in dividing them from minor or harmless offenses (31). The economic starting point of method means, that answering this normative question about social injuriousness of certain actions in the (31) À comprehensive survey over these questions is given by Pedrazzi, Odierne Esigenze Economiche e Nuove Fattispeci Penali. Rivista Itiliana di Diritto e Procedura Penale 1975S. 1099 ff. — 64 — field of economic life requires economic standards and substantial economic knowledge to measure the harm caused by the actions concerned. For example it isn’t possible without knowledge of the eco- nomic system, its connections and of economic cartel law to distinguish between wanted or detrimental cartel agreements. Without knowledge of the stock markets mechanism, the social and economic harm produced bv misuse of insider information cannot be judged. 3. Exact knowledge of phenomenology further is a basic requi- rement for finding out causes and criminogenic factors of economic and business crimes. Especially if one considers the previous stated aims of criminological research in the field of economic and business crimes, it becomes obvious that besides the phenomenology just handled etiology in the sense of classic criminologv has to become the second point of the main effort. Causes of economic crime differ very much among the various forms of economic crime, and can’t be generalized or be generalized only on a very abstract level. Rather often shortcomings of economic law itself cause economic delinquency. This leads to the necessity of reforming z2conomic law regulations — e.g. in the case of bogus firms — or of reforming substantive criminal law. — The necessity of reforms of substantive criminal law which has been the first step of European reform tendency, can be illustrated with the example of computer crime and German penal law The most wide-spread form of computer crime in Europe is presented by the so-called computer manipulations. These are acts in which computer data is altered on purpose to get illegal financial benefits. So, for example, an employee of a German labour office manipulated his employer’s computer to transfer high amounts of children*s allowances to his private accounts. The programmer of an American software firm put an additional programme routine into the programme of a bank computer. The result was that his own overdrafts were not registered by the bank. In two German cases employees doubled their monthly income by manipulating payroll figures in the data store of the computer. Computer manipulations were also involved in the collapse of the German Herstatt-Bank, which led to a damage of more than 200 million Deutsche Mark, and in the spectacular American Equity Funding case which caused damages of about one billion dollars. — 65 — These manipulations are not grasped in a sufficient way by German criminal law because the application of the German provision against fraud supposes the deception of a person — not a computer — and because accounts are not consi- dered physical objects in the sense of the German provisions against theft and embezzlement whereas French and Swiss criminal jurisdictions consider the money on bank accounts (« electronical money ») to be a res (or chose) in the sense of the provisions against embezzlement. The legal provisions according to the English Theft Act are dubious in this regard to say the least. Whether Computer Crime is sufficiently grasped by N orth American criminal law, is quite uncertain too. À study of the Stanford Research Institute, prepared for the National Science Foundation, comes to the preliminary result that the Federal Criminal Code and the existing state provisions would, by broad interpretation, allow to consider the unprivileged copying of software, such as computer programs, a larceny or theft under some state laws (32). In addition 18 U.S.C. Par. 1001 requires no fraudulent obtaining of a « thing oË value», a false statement sufficing for this provision. However the meaning of res is principally restricted to tangible things so that, under common law, only tangible evidence of a chose in action (e.g, a deed or a check) can be the subject of larceny and embezzlement (33). Thus embezzlement or theft of bank account money in the given examples is not embezzlement or theft under North American law. As this short example of computer crime shows, the question of legal reform work of course depends much on the relevant frame work of national law. Under North American law, but also under French law criminal provisions generally seem to be sufficient for combatting economic and business crimes — although there seems to be only a rather wide-spread quantity of different provisions under which a perpetrator may be captured e.g. for bribery and corruption. It seems to me that the necessary reform work in substantive law of most European countries and of North America should first go into the direction of decriminalization in the field of economic and business offenses. Discussion on decriminalization has focused too much on the so-called victimless crimes in the U.S., whereas in Europe decriminalization has been discussed and provided for in almost all (32) Nycum, The Criminal Law Aspects of Computer Abuse (1976). (33) See for example Loewy, Criminal Law (197S) p 81 with references, also Kadish/Paulsen, Criminal Law and its Processes (3rd ed. 1975) p. 691 f. ; for the legal situation in the Federal Republic of Germany, France and Switzerland, see Tiedemann, Wirtschaftsstrafrecht vol. 2 p. 155 with further references. — 66 — fields of economic and social life. As Packer stated in his outstanding book on The Limits of the Criminal Sanction, with regard especially to antitrust offenses, the fact of penalising almost everything weakens general deterrence and hinders the public’s and the judge's awareness of many business crimes being socially harmful and morally bad (34). The necessary process of diminishing the number of criminal offenses in the future makes it at the same time necessary to create new statu- tory provisions concentrated on few and really severe offenses. The Alternative Draft of the German Criminal Code (Special Part, Chapter 2 : Crimes against Business), published in the beginning of 1977, might be a kind of model for reform work as far as criminal law is concerned in this field (35). Of course, law reform and especially reform of criminal law is not the only means of better fighting white collar crime. Improving the activities of law enforcement agencies is at least as important. But as these agencies can only enforce existing law the improvement of subs- tantive law might be an important step in amending law enforcement policy in the field of economic crime. — e (34) Packer, The Limits of the Criminal Sanction (1968) pp. 358 ff. ; for the principles see especially Kadish, Some Observations on the Use of Criminal Sanctions in Enforcing Economic Regulations, University of Chicago Law Review 1963, vol. 30 p. 432 ff. (3S) Lampe/Lenckner/Stree/Tiedemann/Weber, Alternativ-Entwurf eines Straf- 8esetzbuches, Besonderer Teil, Straftaten geger die Wirtschaft (1977). — 67 — INTERNATIONAL COOPERATION IN ACTION RESEARCH A FRENCH-AMERICAN EXAMPLE JACQUES VERIN PAUL STRASBURG Ministère de la Justice Paris Office Paris - France Vera Institute of Justice c/o Ministère de la Justice Paris - France Modification of pretrial release procedures occupies an important place on the list of American criminal justice reforms over the past two decades. Prior to 1961, pretrial release was based almost entirely on a defendant’s ability to post a bond or cash bail as a guarantee of his reappearance at trial. In that year, the Vera Institute of Justice in New York City began the Manhattan Bail Project, offering judges an alternative to bail as the basis for granting pretrial release. The alternative, known as « release on own recognizance » (or R.O.R.), related the likelihood of reappearance at trial not to a defendant’s ability to raise money but to the strength of his community ties. Within three years, verified community ties information provided to arraignment judges by the Manhattan Bail Project produced a four-fold increase in the rate of pretrial release while lowering the failure-to-appear rate below that of defendants released on bail. À 1964 National Conference on Bail and Criminal Justice, sponsored jointly by the U.S. Department of Justice and Vera, was a direct outgrowth of this success. Together, the Conference and the Project helped to stimulate the Federal Bail Reform Act of 1966, the first change in federal bail law since the Federal Judiciary Act of 1789. Similar projects have since been developed in hundreds of cities throughout America and today R.O.R. is a widely accepted alternative to bail (1). Although money bail is not a central feature of French criminal justice, how to determine who should be released before trial and how to assure reappearance at trial are issues of no less significance or difficulty than in the United States. For that reason, the French Ministry of Justice is interested in exploring whether the methods — r (1) For a brief history of the bail reform movement, see National Center for States Courts. An Evaluation of Policy Related Research on the Effec- fiveness of Pretrial Release Programs. Denver, Colorado, 1975. — 69 — developed by the Manhattan Bail Project for recommending pretrial release can be used to increase the efficiency and fairness of pretrial release in France. The Ministry has invited the Vera Institute to join French researchers in conducting an experimental pretrial release project at the fribunal de grande instance of Paris, with funding support from the French government, the German Marshall Fund of the United States and the Ford Foundation. This project, now in the planning stage, offers an example of the possibilities and complexities of international collaboration in action research. The Manhattan Bail Project Bail in the United States is descended from the early English practice of releasing prisoners to third parties during the long delays that often occured between an arrest and the arrival in town of one of the king's travelling magistrates. If the defendant failed to appear at trial, the third party (or « surety ») would himself be tried and jailed. Over the centuries, it became the practice for third parties to provide financial guarantees rather than standing in for defendants, and in the United States this practice developed further into the posting of cash bail by the defendant himself or bonds by commercial bail bondsmen in return for collateral and fees provided by defendants. Because law and practice established that all defendants except those facing a possible death penalty were entitled to the opportunity to post bail, the critical factor determining whether a defendant would be released became the amount of bail set by the court. Relating the risk of flight from justice to a particular bail level in individual cases is a difficult proposition, however. The difficulty is compounded when arraignment judges also take into account the danger posed by a defendant to the community, which they often do even though the legality of this practice under American law is not clear. As a result, standards for setting bail are vague and bail amounts for similar offenses vary widely. Moreover, the effectiveness of bail in assuring return to trial is open to question. Its principal impact, prior to R.O.R. at least, seems to have been simply to discriminate against poor defendants (2). The importance of finding a more equitable basis for pretrial release than bail was underscored by studies which showed a strong relationship between a defendant’s pretrial custody status and the ulti- _— (2) See for example Freed (DJ.) and Wald (P.M.), Bail in the United States 1964. Washington, D.C. U.S. Department of Justice and Vera Foundation, Inc… 1964. —m— mate outcome of his case. For example, a comparison of case outcomes for 374 defendants held in detention and 358 released on bail in New York City in 1962 showed that those in detention were significantly more likely to be convicted and to receive prison sentences than those free on bail. This disparity held true even when such factors as prior record, family integration, and employment stability were controlled (3). To confront these inequities and hardships, the Manhattan Bail Project sought an alternative to bail which would be fairer, allow more people to be released before trial. but not raise the risk of flight from justice. The method chosen was based on the simple and reasonable proposition that the stronger a defendant’s social ties to the community are, the greater the probability is that he will return for trial if released. Each morning project staff interviewed defendants awaiting arraign- ment. They collected information on the prisoners’ place of residence, employment status and history, contacts with family members, prior criminal record, and. possible character references, and then verified the information by telephone. Eventually a system of points was devel- oped and applied to the information so that the sole determinant of a recommendation for release without bail became the number of points received. In order to measure the impact of the Project on pretrial release, a controlled experiment was established. For half of the defendants recommended for release, written recommandation forms were pre- pared and presented to arraignment judges. The other half were put into a control group ; no recommendations were forwarded for them. The results were impressive. Fifty-nine percent of submitted recom- mendations were followed by judges while only 16 percent of the control group were released without bail. Only 1.6 percent of the R.O.R. defendants willfully failed to appear in court while 3 percent of those released on bail failed to appear. Sixty percent of defendants who were released pending trial were eventually acquitted, compared to only 23 percent of the control group. And only 16 percent of the released defendants who were convicted were sent to prison while 96 percent of those convicted in the control group received prison sen- tences (4). Sixteen years after the start of the Manhattan Bail Project, pretrial release based on community ties continues on a large scale throughout —— 1(3) Rankin (A.) « The Effect of Pretrial Detention », New York University Law Review, Vol. 39 (1964), pp. 641-655. Though such strong associations between detention and final disposition have been found in several studies, precise causal relationships have not yet been defined satisfactorily. (4) Programs in Criminal Justice : Vera Institute of Justice Ten-Year Report, 196]-1971. New York : (1972), Vera Institute, p. 31. — 71 — New York City. The Pretrial Services Agency, operated by the Vera Institute for the City, interviews each month well over 4,000 defendants arrested on felony or misdemeanor charges other than prostitution. Like its predecessor, the Agency verifies the information collected, makes release recommendations to arraignment judges, and, using sophisticated computer technology, reminds released defendants about subsequent court appearances through letters and telephone calls. Pretrial detention in France French criminal law detines three offense categories Ccrimes, délits, and contraventions, in descending order of seriousness. Pretrial detention is not permitted for contraventions or for délits with a maximum possible sentence of less than two years in prison ; for other délits and crimes, it is optional. Responsibility for ordering detention rests almost exclusively with examining magistrates, who oversee the preparation of cases for trial. The single exception occurs in cases of flagrants délits, délits in which the defendant is caught in the act or in possession of incriminating evidence. Prosecutors are authorized to order persons charged with flagrants délits detained for appearance in court the same day or, at the latest, the next day. If detention for more than 24 hours is necessary, the case must be handed over to an examining magistrate. À prosecutor’s detention order may be prolonged by the Tribunal at which the defendant appears, however, and often is, usually for periods of a week at a time. Thus prosecutor-ordered detention, though brief, covers a substantial number of cases and frequently leads to extended detention. While lacking authority to order detention in non-flagrant cases, prosecutors nevertheless exercise powers that have an important bear- ing on detention decisions. The most significant of these is their discretion to involve an examining magistrate. Cases charged as crimes must be passed to examining magistrates, but this category applies to less than two percent of all offenses. With regard to délits, prose- cutors have the option of proceeding to trial without ordering a pretrial investigation by an examining magistrate. Since circumventing the examining magistrate precludes detention, this option gives prosecutors de facto authority to release. In 1975, a new law expanded prosecutors’ release powers into the realm of flagrants délits. Formerly all such offenses had to be tried within 24 hours or turned over to examining magistrates. The new law made it possible for prosecutors to remand flagrants délits to a later court date and release defendants in the interim, a procedure known — 72 — as rendez-vous judiciaire. Rendez-vous judiciaire has obvious advan- tages for the defendant. It permits him to consult with a lawyer, prepare his defense, and perhaps most important of all appear in court as a free person who has already demonstrated his reliability by returning for trial. In its first year, rendez-vous judiciaire was granted in about ten per cent of cases charged as flagrants délits in Paris. À second source of prosecutors’ influence on detention decisions is their role in setting the initial charge level. Not infrequently, prose- cutors reduce change levels from crime to délit by choosing to ignore aggravating circumstances, such as commission of the offense at night, in groups, with a weapon or by an employee of the victim. This reduction, known as correctionalisation, not only lowers the maximum possible sentence on conviction but also lowers the amount of time that can be spent in pretrial detention. It also makes it possible for the prosecutor to forego detention altogether, if he desires, by giving him the option not to involve an examining magistrate. Prosecutors have considerable influence in the other direction as well, that is, in imposing detention when they deem it necessary or desirable. They may issue formal requests for detention when for- warding cases to examining magistrates, and they may appeal decisions that do not conform to their requests. Although data are not readily available, it is commonly believed that in the Paris Tribunal examining magistrates accept 90 percent of prosecutors’ requests for pretrial detention. Furthermore, should an examining magistrate wish to free a defendant held in detention, opposition by the prosecutors automa- tically prolongs detention until an appellate court rules, usually a matter of several weeks. The duration of pretrial detention in France is limited by law. For délits carrying a possible sentence of two years or more in prison, the initial term of detention is four months, renewable for additional four-month terms following review by the examining magistrate. Defendants charged with délits who have not previously spent at least three months in prison may not be detained for more than a total of six months, however. Persons charged with crimes may be detained indefinitely before trial, up to a period equal to the maximum sen- tence that could be imposed for the offense on conviction. The Code of Criminal Procedure creates a presumption in favor oÏ pretrial release. Detention is justified only when necessary to preserve evidence to prevent defendants from tampering with witnesses or conspiring with accomplices or co-defendants ; to preserve the « public order » disrupted by offenses ; to protect defendants ; to put an end to offenses or prevent their renewal ; to guarantee defendants’ return — 73 — to court ; and, finally, as a response when defendants willfully break conditions of pretrial release (5). This list is long enough and the conditions broad enough to permit considerable discretion. When questioned about the basis for their detention orders, examining magis- trates most often cite prior criminal record, the seriousness of the offense charged, and the strength of the defendant’s community ties — factors which are also important to American judges. But because examining magistrates have no objective means of weighing these fac- tors, their decisions stem largely from a subjective evaluation of whatever information is on hand. The role of bail in pretrial release has varied somewhat over the past 200 years, though in general it has remained minor. Bail made its first and most legislation granted an absolute right to release on bail for most offenses. It was a short-lived right, however. The first Code of Criminal Procedure, promulgated by Napoleon in 1808, abo- lished the right in favor of mandatory pretrial detention for most offenses. An exception was permitted for first offenders accused of délits, but their release was discretionary and required the posting of bail. In 1865, the Code was liberalized with the result that bail declined further in importance. First offenders with fixed residences who were accused of délits carrying prison sentences shorter than two years were released as a matter of right. Release in other cases was at the discretion of the examining magistrate, and although bail could be required, it was infrequently imposed. On several occasions since 1865, legislation has sought to encou- rage wider use of bail as a means of expanding pretrial release when it is not a matter of right. For the most part, these moves have not succeeded (6). The most recent and most important initiative occured in 1970 when the legislature authorized examining magistrates to make use of a variety of intermediate control measures lying between detention and outright release, during the pretrial period. Bail figures prominently in the package of permissible measures, known collec- tively as contrôle judiciaire. Two separate purposes of bail are defined by this law. One is to assure that the defendant fulfills all his obli- gations, both before trial and after trial (e.g., conditions of probation). The second is to cover damages that may be awarded to victims of the offense, fines, or the costs. of the judicial procedure. At the time bail is set, the examining magistrate determines the amount, the terms of payment (period and level of installments) and the proportion that (5) Code de procédure pénale, Article 144. (6) See Craig (W.L.) and Dobrovir (W.A.), « The French Experience with Preventive Detention », American Bar Association Journal, Vol. S7, (June, 1971), pp. 565-569. — 74 — is to be used for cach of these two purposes. The part that serves as a guarantee is returned to the defendant if and when all obligations are met. The other part is returned in whole if the defendant is found not guilty, or it may be returned in part if the final judgment against him imposes a fine or court costs that are less than the amount of bail required (7). The uses of bail prescribed by the 1970 law are innovative variations on Anglo-Saxon practice, but they have remained largely theoretical. Only a small proportion of defendants released from pretrial detention are required to post bail (five percent in 1973, the latest year for which figures are available) (8). Other kinds of pretrial release condi- tions permitted under contrôle judiciaire (e.g., restrictions on mobility, periodic reporting requirements, or participation in treatment) are applied more frequently, although the great majority ofË released defendants (74 percent in 1973) are still freed without any conditions. The apparent explanation for French resistance to bail is concern that it might lead quickly to economic discrimination and a belief that if detention is not essential, release should be granted outright. Abuse of bail, one of the issues that inspired the Manhattan Bail Project, is therefore not a stimulus for reform in France. Yet there is continuing concern that pretrial detention may be excessive. According to national figures, in 1973 examining magistrates handlied 70,000 new cases (some involving more than one defendant) and detained approximatively 54,000 individuals, two-thirds of whom spent tes at least a month in detention. In Paris that year, examining l handled about 11,000 new cases and issued 16,000 detention orders (9). While precise figures are not available, it appears that virtually all defendants charged with crimes in Paris are held in pretrial detention for some period, and that a minimum of two-thirds of persons charged with non-flagrant délits for which detention is possible are in fact detained (10). The modal period of detention for all these detainees is between one and four months (10 bis). (7) Code de procédure pénale, Articles 138 and 142. (8) Compte général de l’administration de la justice pour 1973, Paris, umentation française, 1976, vol. 2, p. 138. (9) Ibid., pp. 314-317. Cases carried over from 1972 are not included in these figures. (10) The detention rate for délits is estimated by dividing the number of detentions in Paris by the number of charges for detainable délits. This latter igure was calculated by first increasing the number of délits convictions (op. cit. (8), Table 31) by 15 percent to account for the hypothesized number of cases Not resulting in conviction, and then excluding offense categories for which the Maximum penalty is too low to permit detention. (10 bis) Op. cit. (8), p. 318. — 78 — These figures do not prove that pretrial detention is excessive, of course, but this and other evidence hints that it may be. While the total prison population in France declined by 10 percent between 1970 and 1975, the number in pretrial detention rose by 20 percent. In 1975, pretrial detainees were 44 percent of the prison population {11), a remarkably high figure in view of the presumption of innocence on which French criminal justice has been based since the Declaration of the Rights of Man in 1789. Furthermore, of the 54,000 detention orders in 1973, only 21,000 (39 percent) were terminated by convic- tions involving prison sentences. The remaining detainees were either relcased before trial by examining magistrates who deemed the necessity ‘Vr tor propriety of) detention to have ended (41 percent) ; acquitted or had their cases dropped before trial (2 percent) ; received non-custodial sentences on conviction (15 percent) ; or released for other reasons (3 percent) (12). In short, in a substantial majority of cases detention was not followed immediately by prison sentences, and in at least IS to 20 percent of cases no prison sentences resulted at all. Even these numbers, suggestive though they are, do not provide convincing evidence of excessive detention. What is required in order to know whether, in the words of a British commission studying bail, «those remanded in custody form an irreductible minimum » (13) is research focusing specifically on this question — research that compares the characteristics of detained and released defendants, examines judges* reasons for detaining, and studies behaviour during pretrial release. Such research might also reveal whether detention discriminates against certain classes, as it did in América before ROR and whether detained defendants receive more prison sentences and longer sentences, other things being equal, than defendants granted pretrial release. The methods used in the Manhattan Bail Project —— in particular, controlled experimentation centering on objective measurement of the strength of community ties — are a useful way to begin addressing these issues. Applying the Vera System in France Transferring the Vera methods from the American to the French context will require their adaptation to accomodate important proce- dural and structural differences. The role of the prosecutor in detention __ ns es e r 0 c r - (M Rapport Général sur l’exercice, 1974. Direction de l’administration népitentiaire. Paris, Ministère de la Justice, 1974, pp. 5-6. (12) Op. cit. (8), p. 318. (13) Bail Procedures in Magistrates’ Couris. Report of the Home Office Working Party on Bail Procedures in Magistrates’ Courts, London : HLMSS.0. (1974). decisions is one significant difference. In New York, all arrested defen- dants appear before an arraignment judge who makes this decision. Procecutors often make recommendations on the release question, and their recommendations generally have an important influence, but they have no independent release authority of their own. In France, as noted, prosecutors have an option of relcasing persons char- ged with délits instead of referring their cases to an examining magis- trate. The prosecutor, rather than the judge, should therefore be the first recipient of community ties information collected by the nrojec:. In addition, since prosecutors’ recommendations to examining magis- trates carry, if anything, even more weight than prosecutors’ recom- mendations in the U.S., their cooperation and support are vital to the success of any pretrial release plan. Procedural differences will also affect the project’s timing, which U.S. experience has shown to be an important element in successful release of large numbers of defendants. The sooner a project is able to interview a defendant after arrest, the more likely it is to be able to secure his release (14). In general, early access to defendants has not been a major problem in America. While defendants are held in police stations or local jails before arraignment (sometimes as long as 72 hours), they are allowed to contact lawyers and family or friends. Because such contacts are routine, project staffs have been able to conduct interviews during this period without disrupting normal procedures. In New York City, in fact, stafËf of the Pretrial Services Agency make the rounds of police stations throughout the night to interview newly arrested persons so that recommendations will be available to the court the following morning. Pre-arraignment procedures in France make early access to detained defendants more difficult. Police may hold suspects at sta- tion houses for questioning prior to formal charging during a period of 24 hours, renewable for another 24 hours on approval by the prosecutor. The inquiry that takes place during this time, known as garde à vue, 1s secret no one except police or prosecutors is per- mitted to have contact with the detained person, not even lawyers or relatives. Once the period of garde à vue has ended and the preli- minary police inquiry has been completed, the law requires that the detained person either be freed or presented to a prosecutor for formal charging as soon as possible. At the charge hearing, the initial detention decision is also made. In Paris, persons whom police and prosecutors decide must be held for charging are brought from police stations to a central detention center (called the Dépôr) in the Palace of Justice where they are kept — (14) Op. cit. (1)! pp. 66-69. — 77 — until their appearance before a prosecutor or examining magistrate. Of approximately 40,000 persons subjected to garde à vue each year, about 13,000 are subsequently brought to the Dépôt. The rest are either released without being charged or are ordered to appear at court on their own at a later date. Only thosce brought to the Dépôr would be interviewed by the project. The time spent at the Dépôt prior for the first appearance before a prosecutor or magistrate may range from 15 minutes to 24 hours, depending on when the defendant arrives and when prosecutors or magistrates are available. The majority of defendants arrive at the Dépôt late at night. Before the first appearances begin at 9:00 a.m. the following morning, booking, fingerprinting and other adminis- trative procedures must be completed. This leaves little time — in many cases, less than an hour — for completing interviews and verifications. The problems this poses will be compounded if, as anticipated, a large number of defendants are unable to supply veri- fication contacts who have telephones. In short, not only will the project be intervening relatively late in the post-arrest period, it will be doing so under the pressure of a tight deadline. Another factor which may reduce the project’s impact relative to the impact of the Manhattan Bail Project is the amount of communi- ty ties information already available to French judges and prose- cutors. In New York before the Manhattan Bail Project started, arraignment judges had little if any information concerning the community ties and personal characteristics (other than prior records) of defendants. Police reports were for the most part limited to facts surrounding the arrest and charges. Thus the Project was able to fill a large void, which partly accounted for its considerable impact. In France, on the other hand, police reports routinely include some information on the accused person’s residence, family and employment status, obtained during questioning in garde à vue and sometimes verified during searches of the defendant’s residence. Thus, because some community ties information is already available, the information gaps to be filled are smaller than they were in the USS. Nevertheless, according to some judges and prosecutors, even the most general background information may be missing or inaccurate if police have not had enough time to do a thorough investigation. Furthermore, important kinds of information relevant to a judgement of a defendant’s stability — such as length of employment or residence in the community — are not usually obtained, nor is verification of information with third parties standard practice. Finally, the available information is by and large not systematically evaluated before deten- tion decisions are made. One of the early tasks of the project will — 78 — be to determine whether filling these information gaps would alter the practice of pretrial detention sufficiently to justify the creation of a new court service. A final distinction to be made is that, at the broadest level, the goals of the Paris project are different from — in a sense more modest than — those of the Manhattan Bail Project. In New York, the challenge was to find a new basis for pretrial release to replace a bail system whose inadequacies were becoming apparent to all. In France, where community ties already figure prominently in release decisions, the objective is simply to learn whether the existing approach can be made to work more- efiectively. Even so, the project has encountered resistance from parts of the criminal justice system during its planning phase. To a degree, resistance is based on skepticism some people believe that pretrial release is already operating as well as it can. Certainly the distance between current reality and the ideal of an « irreducible minimum » in detention seems shorter in France today than it was in New York sixteen years ago, if for no other reason than that detention in France is permitted only for offenses punishable by more than two years in prison. In addition, current practice has apparently produced low rates of failure-to-appear (1.4 percent in 1973) and crime-on-release (1.0 percent) (15). There are also deeper reasons for resistance, which would themselves make an interesting comparative study. Mistrust of experimentation with the criminal justice system in any manner is greater in France than in the U.S. today. Yet on both sides of the Atlantic, changes, when they occur, are usually the .result . of legislative initiatives Which in turn reflect shifts in public opinion. Rarely do social experiments — especially in criminal justice — precede and encourage legislative reform. To find a bold experiment of this kind in France, for example, one has ta return to 1951, when an agreement between the Ministry of Justice and the Court of Appeals in Toulouse esta- blished pilot projects in work release and probation without pre-existing legislative authorization. The experiments worked and eventually resul- ted in major changes in the law (16). Concern has also been voiced about the fact that the project would be staffed at the outset largely by part-time volunteers. Volunteers have traditionally played a limited role in France in probation work and —— t (15) Op. cit. (8), p. 318. (16) See Marcel Reboul, « Grandeur et misère de l'expérience toulousaine de la probation », Revue de science criminelle, 1954, pp. 497 et seq. prison social work, and more recently in contrôle judiciaire. But the prospect of their presence at the Dépôt and their involvement in deci- sions about detention has raised new questions of accountability as well as anxieties about security. For historical and cultural reasons, resistance to this kind of private initiative and the participation of private citizens in the administration of a public service may be greater in France than in Anglo-Saxon countries. To some degree, however, it is a universal — and understandable — phenomenon. Resistance to change may be increased if the issue in question has been made sensitive by earlier disputes, which is certainly the case with regard to pretrial detention. Controversion have erupted in France in the past several years over the duration of garde à vue, the procedure of flagrants délits, the principle of secrecy in pretrial investigations, and and abuse of pretrial detention. Such sore points have contributed to an atmosphere of tension in relationships among magistrates, defense lawyers and the police which, to say the least, works against cooperative experimental efforts. It is worth remembering that the people who developed the Manhattan Bail Project encountered a similar reluctance to accept new ideas when they began their work in New York, and depended heavily on support from the Mayor to overcome it. Once under way, however, the Project helped unite elements of the criminal justice system whose division had been a principal cause of intertia, by giving them a com- mon ground for action and reflection. In so doing, it opened the way for further experimentation and change. The Paris project’*s research orientation should, in any event, alleviate some of this concern. The first stage will be devoted to collecting baseline data without forwarding recommendations to proce- cutors or examining magistrates. Approximately 600 persons brought to the Dépôt will be interviewed. Data on community ties, prior criminal rocord.s current charges. and detention and senencing outcomes will be obtained. Analysis will center on the relationships between defendant characteristics, detention decisions and final dispositions. If it appears that detention rates could be lowered by a more systematic evaluation of community ties, a point system will be developed and tested during the second phase of the project. In the second stage, interviewed defendants will be randomly assigned to experimental and control groups. Recommendations based on the point system will be recorded for both groups, but interview rn and recommendations will be forewarded to prosecutors and examining magistrates only for the experimental group. The two groups will be compared in terms of release rates, characteristics of released and detained defendants, and failure-to-appear and recidivism rates for defendants released during the pretrial period, Additional comparisons will be made with defendants interviewed during the first phase of the project and with those who do not come into contact with the project at all. The results should reveal whether the project’s recommendations have any effect in increasing release rates or in standardizing detention decisions. - If it turns out that despite the neutral research orientation of the project, the various obstacles discussed make intervention prior to the first appearance unworkable at this time, the focus will be shifted to the period immediately after initial detention decisions are made. Interviews will be conducted with defendants on their arrival at detention centers, not only for the purpose of studying the relationship between defendant characteristics, detention decisions and subsequent dispositions, but also to assess whether the provision of services or guarantees beyond those the defendant can provide himself might tip the balance in favor of a decision to release. Over time, a point system will be developed, based on which follow-up recommendations for release will be made to exa- mining magistrates. In addition to supplementary information and recommendations, project staff will attempt when necessary to assemble additional resources (e.g., housing or jobs) or guarantees required by judges as conditions of release. Resarch would be conducted on this procedure in much the same manner as envisioned for a project at the Dépôt. Eventually, whether the project begins in detention centers or in the Dépôt, it is hoped that continuity can be established throughout the decision-making process in order to ensure that information on defen- dants at risk of detention is collected as soon as possible, that all relevant parties are given early access to the information, and that all necessary follow-up steps are taken to secure release whenever it appears to be warranted. Even if it fails to achieve a substantial decrease in pretrial detention, the project should serve a useful function by providing concrete information which French magistrates, researchers and officials may use to judge for themselves the efficiency and fairness of pretrial procedures. Its results might also suggest other approaches to the objective of an « irreducible minimum » in detention. Conclusion It is perhaps worth emphasizing that, despite a substantial diversity in legal, administrative, and social structures, there are important similarities as well as differences in the problems facing criminal justice administrators in both France and America. We believe it is entirely possible that action research projects conducted cooperatively across National lines may reveal that differences between criminal justice Systems relate primarily to technical or secondary factors, while on — 81 — matters of primary importance they have much in common. Two objectives of the experiment described here seem to us to be central. The first is to test whether provision of more complete and more sys- tematic information to judges can produce a reduction in pretrial detention. The second is to encourage a more trusting, cooperative relationship among the various agencies that form the criminal justice system. The progress of the experiment on these fronts should tell us something about whether commonalities in the French and American systems outweigh differences. FEAR OF CRIME AND ATTITUDES TOWARDS CRIME. COMPARATIVE CRIMINOLOGICAL REFLECTIONS* HANS-JURGEN KERNER Institute of Juvenile Law and Youth Welfare University of Hamburg - Hamburg Federal Republic of Germany Introduction : The fear of becoming victimized, particularly to become the victim of crimes of violence, seems to be widespread all over the industrial nations of the Western world today. According to the official data an increase of the particularly relevant offenses and crimes can be observed everywhere. The increase differs to some extent from nation to nation, only in Japan the trend is out of line, if reports about the past years can be considered to be true. In spite of the increase at present there is no answer to the question, whether the « objective security » or the « real state of security » has become decisively worse for most citizens (outside specific areas, where crime is likely to occur) compared to former decades. This difficult problem will be neglected in this essay. À closely related question will be discussed more detailed, instead. For it is known that one has to distinguish exactly between the « objective state of security » and the so called « subjective state of security ». This last term refers to the feeling of security of the people, i.e. the everyday knowledge about the reality of crime, which is based on a variety of partly true, partly false findings, prejudice, attitudes and so on. This feeling of security of the people is somehow connected to the reality of crime, but there is not necessarily a close relation between these two phenomena. In particular cases there might even be an absolute dispa- rity of the feeling of being in danger and the real danger caused by crime. There are several reasons for that. One of the obvious ones is the fact that most people rarely get in contact with crime directly, at least not with the more threatening kinds, although it has to be taken into consideration that recent research has found quite an amount of victimization (e.g. Criminal Victimization 1977). (*) I am obliged to Michael Alex for valuable help in preparing and discussing ,the material. This research was supported by a grant from the Bundeskriminalmt. — 83 — The few experiences of the people affecting themselves form the background of the fear of crime together with the various factors of knowledge from other sources, and also together with relatively stable stereotypes that are strongly fixed to « common sense ». Just when (and just because) the dominating attitudes toward crime in general and toward certain kinds of crime in particular have so little to do with the immediate observation of oneself, they seem to be so consistent. The stronger, consistency seems to be the more apart the (questioned) phenomenon is frora one’s own limited world and daily problems. This aspect will be investigated more detailed below. How people estimate the development of crime, what their attitudes are in generai toward crime and criminals, what fear of crime means and how great this fear 1s those questions are becoming more and more important, because the answers to them (or perhaps the opi- nions about the possible answers to them) may have a strong impact on public means concerned with crime and reform programs. Therefore it is important to improve the knowledge about these problems. There are several different methods to try and reach this purpose. One of the most popular ways are public opinion polls among a more or less representative sample. This method has got its specific problems, which are presumed to be known and therefore will not be discussed in this essay. Surveys like this have been made in the United States since the middle of the 1930s also in the Federal Republic of Germany numerous attempts of this kind have been made in the past, especially after World War II. Sometimes open-ended questions were used refer- ring to what respondents thought being the most pressing problem facing the country at the time, for example. Besides, closed questions were used like, whether respondents felt more or less afraid of crime compared to some time ago, or whether they believed that crime had increased, stayed the same, or decreased during the past years. In the United States also questions of the following type were (and still are) very popular «Is there anv area around here — that is, within a mile —. where you would be afraid to walk alone at night ? ». In connection with a larger research project about the possibility to introduce an index-system as a measuring device for crime, a survey was carried out in fall 1976 in the Federal Republic of Germany among a representative sample of the total population (N = 2000) to inves- tigate the evaluation of crime. In addition to questions about the relative concern about different « classical » and « modern » crimes also several questions were asked, which permit to make statements about the real fear of crime and the estimation of the development of crime by the German people. The answers to these questions and the conclusions will be discussed below in comparison to previous surveys carried out in the Federal Republic of Germany and other countries, with special emphasis on the results obtained in the United States. Among — 84 — others the important and so far neglected question — in our view — has to be answered, whether the results usually obtained by surveys might possibly depend more on the formulation of questions than on the really existing fear of crime of the people and their evaluation of crime. I. — Crime as a General Problem vs. Crime as a Personal Threat 1l) The situation in the United States. If findings about social problems in the United States may be trusted, crime has not been regarded as an outstanding problem by the American people until the 1950s (Lauer 1976, 124). This may be concluded from the open-ended question, for example, what Americans thought was the most important problem facing the United States at the time. Since November 1935, the Gallup Poll has queried Americans annually on the most important problem facing the country in a more or less unchanged manner. Crime and delinquency were first identified as a social problem in 1951. From that time on crime has been repeat- edly named as the most important problem, however, continuously only since 1968 (Lauer 1976, 125). Referring to these results it is particularly remarkable, how much attention was paid to crime just in 1968. According to Wiedemeyer (1976, 383) crime was named by 15 % in July ?68, by 29 % in August °68, by 21 % in September ’68, by 25 % in November ’68, and only by 17 @ of the respondents in March 1969. Only the wietnam war was identified as the most pressing social problem more frequently at that time. It would be interesting for further investigation, whether and to which extent there is a correlation between these results and the « war on crime », which was particularly declared by the American Government of that time. Anyway, the frequency of mention decrased later on and was down at 5 % in August 1975 (Wiedemeyer 1976, 383). Several Gallup surveys showed similar results, when respondents were asked, what they were most afraid of to happen at present or in the future. While crime was not named at all in 1959 and 1964, 6 % of the adults regarded it as a problem in February 1965, and 5 % in 1971, after all (Erskine, 1973-1974, 140, 142). Apart from the especially high frequency of mention in 1968 it may be stated : in open-ended questions Americans used to regard crime as a minor personal and social problem, whereas the results changed, when respondents were asked to pick the problem they had been paying the most attention to lately, from a list of given problems. Thus, in a national survey conducted by the National Opinion Research Center (N.O.R.C.) for the President’s Commission on Law Enforcement and Administration of Justice in 1966 crime was the second most frequently selected from a list of six major domestic problems only race relations were picked by more people (MeIntyre — 85 — 1976, 36). This is in accordance with the results of a Gallup survey mentioned in comparison to an Australian study, in which 23 % the American respondents had named crime as the problem they had been most concerned about after race relations (32 % Wilson 1971 224) The fear to walk alone on the streets at night reveals the same problem from a slightly different aspect. When people are asked directly they tend to respond in the affirmative to a high degree. In 1965 34 and in December 1972 even 42 %, of the respondents named such fear (Erskine 1974, 137) 2) Results from the Federal Republic of Gerinany The survey carried out by us brought about some results — In so far in correspondence with other studies done in the Federal Republic of Germany —, which indicate that the situations cannot be compared off-hand First of all respondents were asked an open-ended question on national problems that deserved more attention (up to 5 answers were coded) : Question n° | « There’s hardly ever everything alright in state and society. One does not agree to everything. There are quite a few problems which the public is not concerned about sufficiently at all. Can you tell me some of the problems that — in your opinion — deserve more attention ? » Unemployment, inflation, school-problems were named more fre- quently than crime and the fight against crime, which were not men- tioned more often than child — and youth — problems and social problems in general (see Table 1). The frequency, with which crime was named (17 %) is relatively high compared to the United States where only 5 % of the respondents had considered crime to be the most important problem in August 1975 (Wiedemeyer 1976, 383). However, the frequency of mention was also higher than in surveys previously done among the population of the Federal Republic of Germany Crime lost its significance completely, however, when a second open-ended question referring to the fundamental personal problems was asked. The question was Question n° 2 « Well, everybody’s got his own personal problems. One person gets depressed by some event in life, the other one gets upset about something else. Can you tell me a few problems, which you are most worried about personally right now ? » — 86 — TABLE 1 Responses to. Questions, Fail 1976, F.R.G. Question n° 1 : Question n° 2 Question n° 3 : Problems, which dœewo much Important personal Importance of governmental responsabilities public attention problems 2000 Basis (N=) Basis (N=) Crime/fight against crime/ Little contact to other sexual offence/terrori very Ë E ; no prisons 17 % people/loneliness 2 % important important important Ë answer « Internal security » 2 % Security for the people .. 6% (Youth-) unemployment. . 31 % keep job 10 % Fight against Child-/Youth-problems/ Child-/Youth- inflation » .eee8s0esseesstese 63 32 > education e e . …. %+%e%0e6es6 17 % problems . . - . »0 000e 5se6e 8% School/College/ Less difference school-reform 19 % School/ in income s....."...... 34 41 18 College . . « 6 « + » e e © 6 % — » 6 À 10 % Problems of old people/ Equal chances old-age-pension e + ». 6806s Old-age-pension/ in school 39 41 14 ta retirement needs . e ® 660586 7% 74 21 2 medical insurance Medical œre/physicians/ Health « . . + 6 6 © % à © © 6 © ® » # 12 % Less unemployment More support Pollution “ e . .. ... 10 % Worties at work 3 % for industry « « . » 6 # . » ® # à # 33 48 13 ta Inflation/prices/taxes/une- Inflation » » » « 0. …. 8008666886s 11 % More safety 43 43 9 qual distribution of taxes . 20 % conc. traffic Home problems 3 % Home problems 5 % Better medical care 42 42 11 general econ. problems 10 % 63 29 » RV + Social problems « © « 6 © » # à 17 % Fight against crime Miscellaneous . e …. . …. 605e 6 # 11 % Miscellaneous « . . ... 6 % Strengthen efforts No answer « « ». … « * s. © » © à e 42 % for defence and No answer . © « © e » © 6 à « » » — — 18 % military forces 13 30 37 15 Total (1) . . » 6 « © © © © #% #% » # 191 % Total (1) 128 % Remove pollution 44 44 (1) More than one answer possible. Again the first 5 answers were coded. As can be seen from Table 1 the answers interesting for our purpose showed so little specification that they had to be combined with other singular mentions to worries about « internal security » (ie. state of endangering of life, freedom, property, etc. within society) at a rate of 2 %. These findings are in correspondence with the results of an Allensbach-survey from February 1972 (Noelle/Neumann 1974, 139). However, in a comparable survey done in the United States in 1971, crime was named by 5 % of the respondents among other worries about the future, whereas it had not been mentioned at all in earlier surveys (Erskine 1973/74, 140) This indicates that the Germans are more concerned about crime as a national problem than the Americans, whereas they feel less worried about it personally. when open-ended questions are asked It is very interesting, now, to compare this phenomenon to the results obtained, when crime has to be selected from a list of given problems. In the present study this was done both in respect to crime as a social and a personal problem. For this purpose ten social problems were written down on a list. Question n° 3 and n° 4 were : Question n°3 « Everyone of us has to pay taxes to enable government to fulfill the functions. However, not everything can be done at the same time, because there’s not enough money to spend. There may be different opinions about the problems, which are most pressing. Well, here’s a list of functions. If you were to give advice to government which function would you regard as very important, important, less important, not important at all ? » Question n° 4 « Well, there are many things one is worried about very much personally and is afraid they might happen to oneself in the near future. Can you tell me, whether you are afraid that any of the following events might happen to you ? Are you afraid of that all the time, often, sometimes, but sometimes not, rarely, not at all ? » As can be seen, the outlook was changed slightly in particular giving the question referring to general problems a « positive » orien- tation. Às Table 1 illustrates the fight against crime was selected as an important or even very important responsibility of government by 92 % of the respondents, hardly less frequently than the fights against inflation and unemployment (each 95 %). At first sight the high frequency of mention compared to the United States (as mentioned above) and Australia (35 % : Wilson 1971, 224) is stnkmg However, — 88 — part of the difference may be explained by the fact that in the American and Australian sample polls respondents had to rate the problems in order of seriousness, whereas in the present survey the importance of the given problems was supposed to be evaluated independent from each other. That means, to every single social problem a unique significance was to be attributed with regard to a possibly increased number of governmental responsibilities in the future. Anyway, the resukt is equivalent to the findings of a survey conducted in the Federal Republic of Germany in August 1972 in a similar way. In this survey 91 % of the respondents had selected means against the increase of crime as important or even very important. Only peace, old-age pension and medical care, measures against rising prices, and pollution had been considered equally or more important (Noelle/Neumann 1974, 132). Regarding question n* 4 the fear of victimization was selected as a permanent or frequent threat by 11 %, of the respondents. However, significantly more often the loss of relatives (32 %,), severe sickness (22 %), involvement into a road traffic-accident (17 %), and the loss of savings (16 %) were mentioned. The fear to walk alone on the street at night was not examined directly in the present study. But a survey conducted byv the Max- Planck-Institute of Freiburg in 1973 may be of considerable help in this respect. According to that survey among the population of the metropolis of Stuttgart, only 52 %, of the respondents said they felt completely or relatively safe (Stephan 1976, 86, 87). This indicates that the citizens of Stuttgart feel somewhat unsafer at night in their own neighborhood than North-American respondents previously asked. However, it has to be taken into consideration that Stuttgart is a densely populated city of 650 N00 inhabitants, whereas the American surveys had usually also included suburban and rural areas. The large difference between responses dependent on the direction and quality of questions, which were obtained by our special methodo- logical approach, illustrates fundamental problems in conducting sample polls. Those problems certainly do exist for American surveys, too, and therefore they should be carefully taken into consideration in surveys to come. Whenever people are not hit exactly on a certain problem, but are asked for their everyday worries or general national problems, crime obviously does not bother them very much. From this, it may be concluded that the public concern about crime as it is suggested by traditional surveying and reported frequently by the mass media might rather be a function of the method of conducting the survey than of the real evaluation of the problem by the people. This aspect was not paid attention to sufficiently in the past, when surveys were carried out on the perception of crime. — 89 — The general risk of getting results biased by response-styles has already been discussed in detail in the studies on the quality of interviews as a means of social science. Especially important in this respect is the social-desirability-response-set. Regarding crime an addi- tional major effect may be assumed by the social stereotype that crime Is increasing all the time, which is mediated by the mass media per- manently, too. It may be stated therefore that obviously the way of putting the question can lead to an overestimation of the concern about crime, or — expressed more clearly — that questionnaires provoke a higher response-tendency in this direction. A suggestive effect can be noticed, so that respondents not really being aware of crime as a problem simply are saying to regard it as an important problem, just because they expect the conductor of the experiment wants them to (Stephan 1976, 111). Therefore it is possible that respondents do not mention such problems they are really most concerned about, but only those. which the scientists supposed as serious at setting up the list thus confirming the conductor”s ideology. That this assumption is more than merely a speculation may be concluded especially from the fact that the respondents in our survey named problems in the open-ended questions that are very much plausible according to the present economic and social situation of the Federal Republic of Germany as well as to their significance for the personal, everyday life, as there are health, unemployment, inflation, and school-problems. However it cannot be discussed in this essay how much the responses do vary in a typical extremely plausible manner dependent on age, sex, income, education, etc. 3) Fear of Crime and Reality of Crime. Because of methodological reasons both open-ended and closed questions concerning the problem were used in the present study purposely. Also on purpose, the evaluation of the fight against crime as a general problem of society was separately inquired from the fear of victimization The results are : as an abstract matter, people tend to regard the fight against crime as a relatively important problem. But even as a possible choice from a list the respondents do not consider crime to be an outstanding personal worry. This permits the conclusion that the fear or victimization is rather low in the Federal Republic of Germany, particularly in comparison to the feeling of insecurity in the United States measured by the impression not to be safe any more on the streets as an individual (however, again the « suggestive effect » of the question has to be taken into consideration, of course). — 90 — Âs a matter of fact, the probability to become the victim of certain crimes is considerably larger in the United States than in the Federal Republic of Germany, as can be seen from Table 2, which includes only a few selected data — to avoid confusion — obtained from the official police crime statistics. TABLE 2 Rates* of certain oftençes in the US.A. and the F.R.G. Offence Nation 1971 1972 1973 1974 Murder …. ....."ee U.S.A 8.6 8.9 9.3 9.7 F.R.G 4.0 44 4.3 4.3 Robbery U.S.A 1875 180.3 182.6 208.8 25.3 30.5 29.5 30.6 26.1 Forcible rape 20.4 22.4 24.4 114 ë 10.7 11.4 11.3 Burglary (1) 1 1564 11339 1 215.1 | 429.0 F } 130.2 ! 4721 | 448.1 ! 581.0 Larceny-theft ËË 21313 | 9004 20582 2 473.0 F.R.G 1 294.3 1 288.4 } 256.0 ] 318.1 Sources : Polizeiliche Kriminalstatistik 1975 - Uniform Crime Reports 1974. (*) Number of registered cnmes per 100.000 inhabitants. —— Since the descriptions of offences partly differ to a high degree from the United States to the Federal Republic of Germany, only a cautious comparison is possible. Furthermore, the structure of the general problem of the evidence of police crime statistics is slightly different in either of the nations. However, at least the tendency of the data might be true, which is also illustrated by the findings of the inves- tigation on the rate of victimization during the past years (Arzt 1976, 32, 33). Although the research done so far has not proved it yet, never- theless the assumption is possible at present that the different responses are due to a really existing difference in the extent of fear, and that this fear is at least somehow related to the réal problems of each of the two nations. However, some authors argue that the response to the question about the fear of crime cannot be regarded as an indication or even as evidence for a concrete feeling of insecurity at all (1) Theft under aggravating circumstances is wider, includes motor vehicle theft, for example (FR.G.). — 9 — (e.g. Arzt 1976, 11). It is stated that responses to those questions only reveal an abstract fear of disaster, and that at best the responses concerning the fear to walk alone at night on the street in their own neighborhood may be regarded as an expression of the real feeling of insecurity. It must be doubted that the general question about the fear of victimization is much more « abstract » than the question about the fear of victimization on the street in the neighborhood. For the general fear of crime also includes an emotional component. This might well lead to concrete action, e.g. improving personal safety measures. In addition, a general question also includes the fear of certain offences which cannot occur on the street (e.g. burglary). Finally, the fear to walk alone at night does not necessarily have to depend on a certain fear of crime only, but possibly on a general feeling of threat, which might be caused by other factors (Stephan 1976, 97, 103 mentions the fear of « dissocial figures » and trouble in general, the fear of foreigners, the fear of drunkards, and the fear of darkness, for example) The remarkable discrepancy between the concern about personal worries and the evaluation of the general respectively the national problems noted in the present study also has been found in other German studies, although on a completely different level. The closer people are to the problem inquired, or the more they are familiar with the subject concerned, the less worried or even « paranoiac » they respond. This, for example, is illustrated by the results of the « Stuttgart victimization survey » from 1973. 20 %, of the respondents asked in that survey supposed an increase of crime in their own ne1ghborhood , assumed an increase during the past one or two years, whereas 64 % for the total city of Stuttgart, and even 83 % thought crime had increased in the Federal Republic of Germany (Stephan 1976, 126) Although there seem to be significant objective differences between the two nations, and although particularly the personal fear of victi- mization belongs to the repertory of regularly defined problems in sample polls conducted in the United States, the discrepancy between personal fear and general concern revealed by German studies has not been examined systematxcally yet in American studies, at least so far as we know. The discrepancy is probany and at least partially due to the followmg two factors : Some scholars do reter to the 1nfluenœ of the mass media, whmh are considered to overemphasize spectacular developments and single cases of crime, hereby deeply influencing (if not distorting) the ideas that people have about crime outside their own immediate environ- ment and their limited view of the world (Jubelius/Stein-Hilbers 1977 177 Ludwig Boltzmann Institut für Kriminalsoziologie 1977, esp — 92 — 15 ff, 46 ff and 77 ff : Feltes/Kerner/Wiebking 1977, 13 ff). Others additionally point out that the identification with their own neighbor- hood, which supposedly exists among respondents, may exert some influence. The identification might support the tendency to defend their environs, so that they do not want to make negative statements about their own neighborhood (Stephan 1976, 128 with further refe- rences ; similar : McIntyre 1976, 38 ; Wilson 1971, 226). On the basis of his experiences in the United States Arzt clamed that it is improper to derive the bad influence of sensational reports from the fact that the estimation of the development of crime in general is more pessimistic than in respect to the own neighborhood of the respondents. He would rather suggest the principally correct notice of the difference between abstract threat of crimes of violence and concrete danger for oneself as an individual to be responsible for this effect. This criticism seems to be confirmed by the results of the present survey at first sight. For only 43 % of the respondents regarded the strengthening of the efforts for defence and the military forces as im- portant or very important, although this subject has always been of great interest to the mass media, just like crime. When expenditures for the military forces are regarded as relatively unimportant in spite of the comparatively broad discussion in the media, this seems to indi- cate that the concern about the fight against crime as an important problem is not only or not even primarily a result of « collective awareness » created by the media, but expresses a real desire of the people. However, this explanation does not fit very well to the fact that in our survey most respondents named the fights against inflation and unemployment as the most important responsibilities of government. which are exactly those subjects that were also regarded as the most worrying personal problems right after the fear of severe sickness. T'his illustrates that usually the personal worries are also meant to be the most important government responsibilities, at least as far as their removal is regarded (supposedly) possible by means of government. When there is a significant gap between the evaluation of crime as a social problem on one hand, and as a personal concern on the other, this probably depends on some intervening variables, after all. In spite of all doubts and objections, one of those variables might be the sensational presentation of the news on crime by the media. The results of the present survey concerning military forces are not in contrary to that. For reports on military subjects are traditionally not nearly as occupied by emotion as the ones on crime. In addition great consensus throughout the population and the political parties of the Federal Republic of Germany can be assumed in that government will — 93 — try hard (enough) to save the « external security» (defence of the state, etc.) and does not need being forced by the public. However, this controversial question cannot be answered by referring only to the surveys conducted so far. II — The Real an the Estimated Occurence of Crime From the points stated above,l it may be concluded that open- ended questions are most suitable to get relatively correct information about the problems the respondents are really most concerned about. Under these circumstances it seems quite interesting to see that in our survey respondents from rural areas identified crime as a problem that should bepaid more attention to significantly more frequently than other respondents (21 %). For according to the official data this group of the population is less likely to get involved into crime itself than the inhabitants of larger cities, as can be seen from Table 3, which is a brief review of the data from the United States and the Federal Republic of Germany. TABLE 3 Crime Rates according to size of city 1974 Rate® Population group (I =fiF.R.G., Il =_U3A.) _ — — mmms mt e e e dc e dn en cities under 20,000 inhabitants 2,540.0 cities, 20,000 to 100,000 4,605.0 cities, 100,000 to 500,000 .. ... ................s 5,939.0 cities over 500,000 (1) . . à à = + . 6. 6 » © 6 « »* © 6 = 6 6 6 6 . % ® © » — mts c0c r 7,324.9 TT 3,631 cities under 10,000 . . . .. .. ... ... 3,818.2 1,30M cities, 10.000 to 25,000 . .. ................... 4,4178 519 cities, 25,000 to 50,000 » © e àn 6 %* @ € © 6 e © 6 à « © à © » » = » 5,151.6 253 cities, 50,000 to 100,000 « » » »* —. #% « € » 5 » © % ® e € < e — 5,741.4 107 cities, 100,000 to 250,000 + . « ». » à % € + € 6 » » »s % » — à ° 7,111.0 31 cities, 250,000 to 500,000 . e e #% 9 %% # » « © © » © « @ » % = 7,155.4 21 cities, 500.000 to 1,000,000 e ................. 8,029.4 . » . #% % 5 5 < à 5 % 6 © #w « © #% ® 6 à n 6 8 # 6,961.4 ——n 6 cities over ; 1,000,000 —emar - Sources see Table 2. (*) Number of crimes per 100,000 inhabitants. (1) Usually not standard metropolitan statistical area, but core-city. — 94 — Interesting for a comparison is only the similar structure in both countries. The data itself cannot be compared exactly, because the units chosen differ very much. In addition, there is only a sample available of cities in the United States, including only 6 cities with a population of more than 1,000,000 and a total population of 18,541,000, for example. The enlarged concern about crime in rural areas of the Federal Republic of Germany is not only in contrary to the crime rate noticed by the police but also to survey-results from other countries. In the Australian study mentioned above a positive relation was found between increasing urbanization and concern for crime, for example, according to the increase of crime rates (Wilson 1971, 224). Just as well surveys done in the United States in 1972 showed that the fear to walk alone at night in the neighborhood was far less distinct in rural areas (24 % resp. 28 %) than in cities, where 42 % up to 53 % respectively 40 % up to 51 % of the respondents answered, they were afraid (Hindelang 1975, 9 : Erskine 1973/74, 138). However, the contrary result of the present study possibly has quite a realistic background considered the development of crime during the past years according to official data. For crime rates increased above average in rural areas from 1974 to 1975, for example ; the increase was 8.8 % compared to 4.3 % in small cities, 1.3 % in larger cities, and 3.1 % in major cities. However, it has to be taken into consideration that the typical sources of error in sample polls might have influenced the results of this survey, too. Thus, the major city as a political object or administration unit sure is something else in the minds of people than the city as an organized, social environment of their lives they partially know quite well and are able to identify themselves with. Finally, as mentioned above, surveys conducted in Germany repeatedly demonstrated that the evaluation of crime as a social problem is largely independent from the probability of personal victimization. Further details, such as the relationship between the officially stated crime rates and the personal fear of crime depending on age and socio- economic status cannot be discussed in this essay. IM. — The Estimation of Crime Trends In previous surveys on the estimation of the development of crime respondents usually were asked, whether they felt crime was increasing. In April 1975, 75 % responded to this question affirma- tively (Noelle-Neumann 1976, 86). In a survey conducted in Bremen in 1972 even 83 % of the respondents had supposed an increase (Bremen- Report 1973, review 2). As can be seen from American surveys, in December 1972 about 51 %, and in October 1973 about 48 %, of the Americans thought that crime had increased in their neighborhood, — 95 — whereas 9 ©/, resp. 7 %expressed the opinion it had decreased (Erskine 1973/74, 132-134). Again, it has to be taken into consideration that these American surveys refer only to the immediate social environs, whereas German respondent, were inquired about the development in the country as a whole. But even the comparably lower frequency, with which an increase of crime was supposed by Americans in the quoted polls, leads to the problem, whether the formulation of the question might have suggested the response and provoked attitudes artificially. The question, whether people think crime is increasing, reinforces the social stereotype of the respondent, his everyday knowledge, and therefore acts as a strong stimulus, which under certain circumstances might even force him to respond affirmatively. The survey conducted by us does not prove this assumption, but makes it appear rather plausible. For in accor- dance with other recent studies respondents were not asked for the development of crime in general, but had to state for certain offences, whether in their opinion these offences had increased during the past years, had stayed the same, or had decreased. The result was that respondents suppose an increase significantly less often, when, what is described as crime in a questionnaire, is somehow concrete and understandable for them. In the question referring to this, ten different offences were given, as can be seen from Table 4. The question was : Question n° 6 « Well, therce are different kinds of crime, and one reads a lot about crime, or at least hears about it. Therefore you probably have an impression, what the development was like during the past years. Here’s a list of several offences. Would you please tell me, which of these — in your opinion — have increased, stayed the same, or may be even have decreased during the past years ». TABLE 4 Estimation of Crime Trendî F.R.G. 1976 —0 0= —— — Off enœ—(1)) “ Responses increased C stayed decreased « thç same p OO —— —n e — 000 Criminal homicïdë 36 % 51 % 10 % Tax evasion etc. oooooooooooooo 37 % 52 % 5 % Rape 42 % 47 % 8 % Political offences 43 % 41 % 13 % White collar crime etc. uuuuuuuu 51 % 40 % 4 % Driving under influence 53 % 32 % 13 % Robbery, mugging etc. . e © » « = « » 53 % 37 % 8 % Drug offences © . + # ». à = » * + # + — e « ® 54 % 29 % 16 % Breaking and entering (oËf private houses) 54 % 39 % 4 % Shoplifting etc. ..........240200 67 % 28 % 4 % (1) Order is different from the list given to réspoÎdents. — 96 — The structure of the list put down in Table 4 equals the officially stated crime trends according to police crime statistics, a result which was also found to be true for Great Britain in a similar survey (Banks et al. 1975, 229). Only offences relating to narcotic drugs were rela- tively often considered to have decreased, although they had increased the most from 1974 to 1975 according to the official data. Furthermore, it is striking that offences, which are broadly discussed in the mass media (in particular political and sexual offences). were relatively often regarded as not having increased. An inquiry into single offences therefore seems to be at least a more suitable procedure than the rather general questions (formerly asked) to find out the real estimation of crime trends by the people. It also illustrates that the estimation of these special development is very much based on subjective factors and, besides, varies largely dependent on age, sex, socio-economic status, and place of residence of the respondents. These problems. however. cannot be discussed in detail in this essay. Consequently, in surveys to come more attention should be paid to the fact that the estimation of the development of crime largely depends on certain interests and that fears concerning crime therefore can easily be abused for manipulation. This is most important with regard to possible consequences in matters of criminal policy, espe- cially in planning new laws. IV. — The Effects of « Counter-Provoking » In the present study we have also tried, for methodological reasons, to make respondents independent from their usual response patterns by setting up an extreme variation, and to enable them to gain new categories of thinking respectively responding by provoking inquiring. Thus, we asked a question, which was contrary to the usual stereo- type of the « criminal » as somebody completely different from the « average-citizen ». Furthermore, we set up a strong « counter-sti- mulus » by another question referring to crime trends Question n° 5 : « According to press-reports, a state secretary stated a few days ago that crime was not increasing any more in the Federal Republic of Germany, that it might even begin to decrease. What do you think of this opinion considering your own impression ? Is it absolutely right, largely right. not so easy to decide, largely wrong, absolutely wrong ? » — 97 — While up to 67 % of the respondents claimed at least certain crimes had increased (question n° 6), only 40 % were willing to say the opinion of the secretary was largely or absolutely wrong. That means 60 % did not stick to the habitual stereotype of the increase of crime, and were at least not sure, whether there might be some- thing about the statement given by the secretary. In summary, therefore, we can say from the results of our survey, and draw attention to the fact that the evaluation of the state and the development of crime seems basically resistent to change, but never- theless can be influenced deeply by external means in its specific shaping. The opportunities given to methodological research by this incidence on one hand, the dangers in regard to criminal policy on the other hand, are obvious. REFERENCES ARZT (G.) : Der Ruf nach Recht und Ordnung. Ursachen und Folgen der Kriminalitätsfurcht in den US.A. und in Deutschland, 1976, Tübingen : Mohr. BANKS (C.E.) MALONEY and WILLCOCK (H.D) « Public Attitudes to Crime and the Penal System », British Journal of Criminology, 1975. n° 15, pp. 228-240. ERSKINE (H.) : « The Pol Ho , Fears, and Regrets », The Public Opinion Quaterly, 1973-1974, n° 37, pp. 132-145. ERSKINE (H) « The Pols : Fear of Violence and Crime », The Public Opinion Quarterly, 1974, n° 38, pp. 131-145. FELTES (T.), KERNER (H.J.) and WIEBKING (N.) : « Mit Criminal gegen “ Kriminelle », Psychologie heute, 1977, n° 4, pp. 13-18. HINDELAND (M. « Public Opinion Regarding Crime », Criminal Justice and Related Topics, 1975, Washington, D.C. : G.P.O. JUBELIUS (W.) and STEIN-HILBERS (M) « Vermittlung von Informa- tionen über Kriminalität in Massenmedien. Uberlegungen zur Erklärung unterschiedlicher Bewusstseinsinsinhalte », Monatsschrift für Kriminologie und Strafretsreform, 1977, n° 60, pp. 177-185. LAUER (R.H) « Defining Social Problems Public and Professional Perspectives », Social Problems, 1976, n° 24, pp. 122-130. LUDWIG BOLTZMANN INSTITUT FUR KRIMINALSOZIOLOGIE, : ed. « Kriminalität in der Massenmedien »,Kriminalsoziologische Bibliografie, 1976, n° 4, Heft 11-13. McINTYRE (J) «Public Attitudes toward Crime and Law Enforcement », The Annals of the American Academy of Political and Social Science, 1967, n° 374, pp. 34-46. — 98 — NOELLE (E.) und NEUMANN (E.P.), (ed. Jahrbuch der offentlichen Meinung 1968-1973. Allensbach und Bonn : Verlag für Demoskopie. 1974. NOELLE (E.) und NEUMANN (E.P.), (ed.) Allenbacher Jahrbuch der Demoskopie 1974-1976, Bd. VI, Wien-München-Zürich : Verlag für Demos- kopie, 1976. STEPHAN (E.) : Die Stuttgarter Opferbefragung. Eine kriminologisch, vikti- mologische Analyse zur Erforschung des Dunkelfeldes unter besonderer Berücksichtigung der Einstellung der Bevôlkerung zur Kriminalität. 1976. Wiesbaden : Bundeskriminalamt. Forschungsreihe, Nr. 3 WIEDEMEYER (N..) : « The Polls : Do People Worry About the Future », The Public Opinion Quarterly, 1976, (n° 40), pp. 382-391. WILSON (P.R.) « Crime and the Public.», The Australian and New Zealand Journal of Criminology, 1971, n° 4, pp. 223-232. MAT ERIALS : Bremen-Report Innere Sicherheit und Polizeiorganisation. INFAS-Reprä- sentativerhebung, Land Bremen 1970-1972, Bremen not published. Criminal Victimization in the United States. À comparison of 1974 and 1975 Findings. A National Crime Survey Report. Prepared by : U.S. Department of Justice Law Enforcement Assistance Administration National Criminal Justice. Information and Statistics Service. Washington, D.C. : G.P.O., 1977. Polizeiliche Kriminalstatistik 1975 für die Bundesrepublik Deutschland. Heraus- gegeben vom Bundeskriminalamt Wiesbaden : Eigenverlag, 1976. « Crime in the United States United Crime Reports for the United States 1974 ». Issued by Kelley. C.H. Washington, D.C. : G.P.O., 1975. — 99 — IV. - Micro criminology THE RAPE PROBLEM T.C.N. GIBBENS K.L. SOOTHILL Institute of Psychiatry Department of Sociology University of London Lancaster University London - United Kingdom Bailrigg, Lancaster - United Kingdom There is no crime which has attracted so much enquiry, complaint and controversy in recent years, nationally and internationally, as rape. Three factors seem to be involved. First, a surprising, and for some people alarming, increase in complaints of (and convictions for) rape in many countries. Secondly, the rise of women’s organisations to relieve the distress of those who have been raped and sometimes to secure justice for them — such as Rape Crisis centres, Rape Aid, etc. Thirdly, increasing complaints about the ineffectiveness of the criminal law and especially its male-orientated bias which tends to treat the woman victim almost as if she were the offender by implying that she is making a false accusation or is anyway an immoral woman. It seems almost certain that these three trends are related to one another, and that it represents a shift in culture and mores which the criminal law has to adjust to as best it can. Apart from this popular interest, rape presents in a narrow compass a fascinating goup of medico-legal and socio-legal issues. It is almost the only crime in which the mental state of the offender and his victim is equally important in deciding not only what sort of offence has been committed but whether a crime has been committed at all. In non-sexual violent crime, there are many legal gradations of oflence, varying according to provocation and what sort of injury occured. But common to all is the reasonable assumption that no one wants to be injured, even if he is the assailant. When a woman shows no sign of bruising or physical resistance to intercourse, there is nothing to go on except her story, unless there is some corroboration ; though rape is quite rightly defined in law as having intercourse with a woman knowing that she does not consent, or being reckless as to whether she consents or not. Violence is emphatically not an essential element in the offence though some old police surgeon’s text-books refer to woman «having to resist to the limit of their streneth » ! Most modern young women will assume that resistance beyond a certain point is dangerous, and since virginity is not highly regarded, there is less incentive to preserve it. — 103 — In addition, in non-sexual violence there is some popular estimate of normal behaviour ; though research has shown that the range of attitudes to what constitutes violence is surprisingly wide. Aggressive men define a fight in which they got knocked unconscious as an «argument » or a « quarrel », and nothing less than an attack by someone with a knife or a lethal weapon tends to be defined as an « assault ». In relation to sexual behaviour, however, no one has the slightest idea of what the normal range of behaviour is or even the standard deviation of normal behaviour, in spite of Kinsey’s valiant efforts. At a recent meeting in South Africa the question was raised whether a white middle class magistrate was a proper person to consider whether rape had occurred between two coloured people, since many white people tended to think that this was « normal » behaviour between blacks. In Amir’s study of rape in Philadelphia (1971) the great majority of rapes occurred between very young coloured girls and coloured boys not much older, 27 per cent of the complaints being made by the girl herself, raises similar issues of cultural interpretation. Similar cultural issues arise with Anglo-Saxon juries, where there may be a generation gap between older women on the jury and a young victim whose fairly ordinary contemporary behaviour may be inter- preted by them as wanton or promiscuous. Incidence. To turn to the areas of controversy mentioned earlier, the increase has been considerable. À recent Australian survey (Newton, 1976) reports increases in several countries. In the U.S.A. Chappell (Newton, 1976) reported « during the past decade forcible rape rates have more than doubled, the pace of increase being more rapid since 1967 and in the early 1970s reaching a speed outstripping all the major categories of violent crime ». Kraus (1976) quotes reports of increasing incidence of individual and gang rape in Western Germany, the U.S.A. Czechoslovakia, Japan and England. Kraus found that in New South Wales from 1956 to 1969, in addition to a general increase, the rates for pack rape had increased two and a half times faster than the indivi- dual ones. During this period pack rape increased to no less than 61 per cent of all rapes. Rapes by adults increased by 192 per cent but by juveniles by 773 per cent. In England and Wales (Soothill, Jack and Gibbens, 1976) trials for rape or attempted rape numbered 119 in 1951, 266 in 1961 and 412 in 1974, doubling everyv ten years. Since 1973 there has been a falling off, however. Rapes recorded as known to the police and the number found guilty (in brackets) were in 1973, 998 (331) ; 1974, 1052 (343) :1975, 1040 (336) (1). Stability in statistics is always more baffling (1) The reference to these figures is Parliamentary Anwser April 7 1977. than figures which jump up and down, which one would think natural in real life. There are others even more extraordinary. While waiting for the subsequent criminal records of 1951 and 1961 rapes (which were derived from totals prosecuted, so that they included in an interesting way a follow-up of those acquitted) our team (Soothill, Jack and Gibbens, 1976) searched the newspapers for those victims who had been mentioned by name and sometimes also by address. The proportion in the three decades was respectively 24 per cent, 26 per cent and 25 per cent. Considering the absolute numbers involved this seemed to indicate a considerable increase in newspaper space (though in practice it was limited to one or two newspapers). This raises the question, as the fascinating studies of perception of violence by the public in Scandinavia suggest (Council of Europe, 1974) to what extent our anxicties and attitudes are virtually created by the press. The second, even more surprising, fact was that the proportion of those prosecuted who were acquitted, in spite of 100 per cent increase in numbers in each decade, was constant — 19 per cent of 119 cases in 1951, 22 per cent of 266 in 1961, and 25 per cent of 412 in 1974. What balance of social perceptions, police practice, sexual mores or other factors could lead to such constant results ? The Law. Whatever the real changes have been, the law has felt a need to adapt itself rather better to the needs of society. There is little doubt that the process of conviction of a rapist has most unfairly increased the burdens on the unfortunate victim. As if it were not enough to be raped, the girl or woman, if she reports it to the police, finds that she is asked a number of searching questions about what happened, which easily leave her with the impression that she is not really being believed, that she is exaggerating or bending the truth. À certain number of women are sufficiently hurt or annoyed to cease cooperation at this point. Soon afterwards she has an unpleasantly detailed physical examination by a police doctor who also takes a medical history about her physical and gynaecological complaints and who makes an assessment of how much her statement agrees with her physical state. She expects to find, in the doctor at least, someone who will sympathise and help, and she may pour out the confidences of a life time. These will and must be recorded as well as injuries and scratches etc. However, unless the official report is limited to strictly relevant matters, injuries, demeanour, and signs of intercourse, it may, when handed to the prosecution, be sent on «in fairness » to the defence as an objective report, and at the trial an unmarried victim may be asked suddenly whether it is true that some years before she had had two lovers in succession and an abortion a few years after- wards etc, information given originally on the assumption that it was in medical confidence. Very difficult issues of medical ethics are raised — 105 — if the report is not carefully limited to objective physical findings Nevertheless, careful note should be kept of all aspects of the exami- nat10n for a great deal may depend on them Durmg the trial she is liable to be questioned about her sexual behaviour not only with the accused but with others. The more honest and truthful she is the more likely it is that she can be shocked or confused into appearing unreliable or untruthful. Worst of all, if the accused is acquitted, either on the instructions of the judge that the evidence is not sufficient, or after his warning (which some American observers have sought to prohibit) that it is unsafe to convict a person on -uncorroborated testimony, she is made to appear as a liar. Subse- gently. she will be left to get over it as best she can, a situation which Rape Aid has strived to improve. It is of some interest that in South Africa the Rape Aid workers have secured agreement from the police to attend the initial examination and support the victim throughout The criticisms or suggestions for reform raised in many coun- tries therefore include (1) that the processes of trial expose the victim to embarrassing and traumatic questions about her present and previous behaviour : (2) that she receives an unnecessary and damaging publicity : (3) that the scope of permitted cross-examination makes it too easy to raise doubts about consent and secure an acquittal : (4) that husbands should not be excluded from the ambit of the law ; or, if they are, that regular cohabitees should be included also : (5) that clumsy police interrogation could be avoided if specially trained « rape squads » were formed (6) that mere lack of consent, or reckless- ness about it, refers to such a wide range of behaviour in terms of severity (excluding statutory rape or rape of the mentally disordered) that the law should define gradations of offence and penalty, reserving rape for the extreme forms of violence by strangers Many countries had considered these issues with varying degrees -of urgency. Much depends upon the particular pattern of the law whether e.g. guilt is decided by a judge or jury : whether prosecution 1s'decided after private investigation by juge d’instruction, as in France the extent of the powers of the court to regulate the trial procedure etc. In Italy, Denmark, the Netherlands, Germany and New Zealand for example there is legal provision, freely used, to exclude the public from trials or parts of trials, when sensitive evidence is being given to forbid publication of names or even press reports of any kind In these countries controversy seems less vociferous. It is mainly in the Anglo-Saxon countries with common law tradition — Australia Canada, the United States, the United Kingdom — that discussion has been most urgent. Would it be true to say that in these countries the movement for equal rights for women in all respects has been most strongly urged ? — 106 — The circumstances in which rape can be alleged have been so varied that attempts are sometimes made at very comprehensive defi- nitions. As Howard (1970) says of. the different states of Australia «the codes of Queensland and Western Australia have a definition which is more elaborate because it makes express reference to some of the difficulties which have arisen at common law in delineating the limits of the offence » « Any person who has carnal knowledge of a woman, or girl, not his wife, without her consent, or with her consent, if the consent is obtainable by force, or by means of threats or intimidation of any kind, or by fear of bodily harm, or by means of false or fraudulent representations as to the nature of the act, or in the case of a married woman, by impersonating her husband, is guilty of a crime, which is called rape». Most people will find it rather extraordinary that a married woman should not be able to recognize her husband in the dark or that some should innocently think that sexual intercourse with a doctor is not different from a routine gynae- cological examination, but there seem to be no limits to the possibilities. Indeed when the circumstances of rape trials are considered in detail it is sometimes impossible to formulate a coherent idea of what took place. The American Model Penal Code goes even further in trying to define all the possible varieties of rape in order to achieve some differentiation of the penalties to be attached to the varieties. They thus arrive at what are virtually degrees of offence, « Rape by force or its equivalent » is divided into first degree felony, if there was serious physical injury to a victim who is not « a voluntary social companion » and had not « previously permitted him sexual liberties» ; second degree felony if she was compelled to submit by force or threats of injury to herself or family, given drugs or drink without her consent, was unconscious, or was a child under the age of ten ; third degree felony if « intercourse without legally effective consent », if intimidated, incapable of understanding the behaviour, unaware of the sexual act etc. And it 1s a misdemeamour to abuse authority or have intercourse with false promise of marriage. The object of this exercice 1s siated explicitly to be the achievement of uniformity in sentencing in the many American jurisdictions. In England the Heilbron Committee on the Law of Rape decided against gradations, for the circumstances are different. England has a cohesive judiciary with a relatively uniform conception of the sentence appro- priate to particular types of offence. À very interesting Australian study of sentencing policy in rape (1976) shows as a comparable study in England probably would that sentences can be seen to vary according to the behaviour of the victim, the surroundings of the offence and the character of the accused. À sliding scale was drawn up as follows : — 107 — A COMPOUND GUIDE TO RAPE SENTENCING | The Offence r r e rr 0 rr e p Ï —-— —— _— —— es . r e e rr — The F)fiender _— Violence Victim°s Meeting Co—operation Circpm_ÿanœs __ No violence Unambiguously Casual hotel meeting Respectable, no prior consents initially convictions Threats only Signs interpreted as Street pick-up Good reputation, but consent specific personal problem Pushed around Encourages sexual Meets victim at social l — 2 prior convictions only intimacy function, takes her home Moderate Provocative, aware Considerable number of struggle of effect Fraudulent pick-up, but girl goes willingly priors Forcible Provocative, unaware Meets victim at social Extensive delinquency restraint of effect function, lures her away and/or sexual priors Violent Unsuspecting, social Legitimate, pre-arranged Full rape prior struggle companion outing Punching, Totally non-provoking Attack without warning Institutionalised beating psychopathic —— FIGURE I Almost certainly an added factor in not having « gradations » of offence is a perception that the criminal law is a very clumsy and often ineffective means of judging or controlling sexual behaviour. No one doubts that it must intervene in cases where physical force or inti- midation is used or for the protection of the young and weak, or to prevent gross insults to the concept of public decency, but a long list of lesser punishable offences would lead to a great increase in in- conclusive litigation, as well as the possibility that whenever a complaint of any Kkind is made there will be a tendency to convict the offender of at least a lesser offence rather than to decide that a serious offence has been committed or not. The controversy which led to the appointment of the Advisory Group on the Law of Rape in England (Heilbron Committee) arose from an apparently recondite but nevertheless very important legal decision made by the highest court of appeal (the House of Lords) dealing solely with one question if an accused maintains that he thought the woman was consenting, does the jury decide this on the basis of what a « reasonable » man would have thought or what the accused actually appears to have believed even if it is unreasonable. They decided upon the latter criterion, in line with the modern conception of the guilty mind. This was hailed by some as a « rapists charter », implying that a man is entitled to be believed however ridiculous his explanation appeared. The Group’s recommendation on this and other matters was incorporated quite rapidly in the Sexual Offences (Amendment) Act of 1976. This re-defines rape as unlawful sexual intercourse with a woman who at the time of intercourse does not consent to it ; and, that the accused knows at the time that she does not consent or is reckless as to whether she consents to it. It re-affirmed that in deciding whether a man believed that the woman was consenting the presence or absence of reasonable grounds for such a belief is a matter to which the jury is to have regard. in conjunction with any other relevant matters. The Act, drafted with exemplary care, also provided that evidence may not be adduced or questions asked about matters relating to the sexual experience of a complainant with a person other than the accused, or her disposition or reputation in sexual matters without the leave of the court. This will not be granted unless the judge decides (in the absence of the jury) that the evidence presented to him is of such relevance that it would be unfair to the defendant to exclude them, and will onlyv be allowed when the court is satisfied that there is a «striking relationship » between these matters and the complainant’s alleged behaviour on the occasion in question. — 109 — The Group recommended that the complainant should remain anonymous, and that the jury should be composed of a statutory pro- portion of men and women. But Parliament, while accepting the first added that the accused should also remain anonymous until and unless convicted (perhaps in deference to recent legislation with regard to the quality between the sexes) and also rejected the idea of a statutory balance of sexes on the jury, perhaps because of practœal comphcat10ns and the very large precedent it would cause. Criminological. Our own concern has been rather with the question of what sort of behaviour leads to a prosecution for rape, and especially whether rapists are often reconvicted subsequently of this sort of offence, for the long sentences often imposed seem to imply at least partly that they are viewed as persistently dangerous. Most studies (e.g. Christiansen) suggest that sex offenders are less often reconvicted and at longer intervals than property offenders. We, therefore, studied the previous and subsequent criminal records of all cases prosecuted for rape in 1951 and 1961 (including therefore those who were acquitted). So far as types of behaviour were concerned they fell into three broad groups (1977). About a third (30 per cent) were, rather surprisingly, pedophiliac rapes of girls under 14. They pleaded, or were found guilty much more often than others, no doubt because there was no defence or consent ; moreover such men are often remor- seful and also know that the sentence may be lessened if they do not add to the damage by forcing a child to give evidence against them 1n court. About half of these men were charged with attempted rape which may not be very clearly defined. À second group of 20 per cent were «agressive » rapists, with convictions for non-sexual violence in their record before or after the offence. These include some men who are generally intolerant of frustration and assault their friends, enemies or their wives as well as any other girl who refuses them, and since violent offences usually « mean » heavy drinking probably included some who became drunken offenders later. About half pleaded not guilty and they were fairly often found not guilty. It seems possible that apart from greater criminal sophistication which would lead them always to plead not guilty, such men have such an aggressive view of normal sexuality that they do not think they are guilty. We found one man Who had three times, at considerable intervals, been acquitted of rape and who said to reporters « [ do not know why these women to do that you would have to understand how a said I raped them woman’s mind works ! ». The third group of 50 per cent were « others » mainly characterised by men with few if any previous or subsequent offences apart from very mmor ones. About a third of these were acquitted. Hali Williams (l977) has extended these groups from the much fuller details available in the parole dossiers of 78 cases seen (1970/72) — 110 — as a member of the Parole Board (which deals with those sentenced to 18 months imprisonment or over among the 70 per cent of those convicted who are sent to prison). He described four groups — 16 per cent who are pedophiliac offenders, 18 per cent « extremely aggres- sive», of the most violent and dangerous kind, 47 per cent mildly aggressive individuals or group rapists, and 18 per cent « sexual attacks committed by individuals as a means of resoiving personal problems or as a response to extreme stress or tension ». Problems with drink or drugs were never less than 43 per cent, in the mildly aggressive applied to one half of the others, and to 65 per cent of the extremely aggressive group. Severe personality disorders were present in the vast. majority and were least (45 per cent) in the mildly aggressive. Over half (57 per cent) of the « personal stress and tension » group had no previous convictions but nearly all had had serious personality disor ders and 57 per cent a problem with drink or drugs. For the whole sample 23 per cent of the victims were known prevnously to the accused ; only 10 per cent of the offences were committed in the home e1thcr of the victim or the accused. Pack rape, which is exceptionally common in Australia and to a lesser extent in Germany, applied to 35 individuals or ten of Hall Williams’ 79 «cases ». In our 1951 and 1961 sample, a charge with «two or more others » was brought in 17 per cent of aggressives 20 per cent of «others», and a surprising 17 per cent of the pedophiliac rapes. These latter, however, referred often to rapes of (presumably) rather forward 13 year old girls by three or four boys of 16 or 17. It may be difficult to apportion responsibility among these boys and they would usually not reach prison. Among adults however, there are a number of cases in which two or three men in a car, usually the worse for drink, offer a lift to two girls (Who think that this makes them safe), throw one out and all rape the other As other studies have shown, rapists tend to have fewer previous sexual offences than other sex offenders (Ellis and Brancale, 1956 Karpman, 1954 ; Christiansen, 1965) but often a considerable number of other offences. The large Cambridge study (1957) showed that only 9 per cent of rapists had previous sexual convictions. Amir’s large Philadelphia study (1971) reported 4 per cent with previous other sex offences, but « only» 9 pér cent. (which seems astonishingly high) With a previous « rapè arrest ».‘In oùr two cohorts only 22 pet cent of pedophiliac or aggressive rapists had no prévious convictions ; previous sex offences were higher in the pedophiliacs (37 per cent) than m the aggressives (17 per cent). Subsequent offences were common in both groups but only 20 per cent of the 1961 pedophiliac and aggressive offenders had subsequent sex offences and 15 per cent of.the 1951 group. On the main question — 111 — of whether they ever committed rape again, the findings were interesting. Out of 86 convicted 22 years before five were reconvicted of another rape at intervals (after release from the previous imprisonment) of two years, four years, eleven years, and fifteen years after the first offence and one on two occasions — at three and sixteen years. Of the 200 convicted in 1961 there were four reconvictions for rape or attem- pted rape, twelve for indecent assault of different sorts. and eight of indecent exposure. One of the advantages of studying prosecutions was that we were able to follow those acquitted (though not of course subsequent acquittals). Of the 82 acquittals in 1951 and 1961, three were sub- sequently convicted of rape or attempted rape — a slightly higher proportion than the five out of 86 convicted in 1951. Subsequent sex and violent offences were similar in the aoquitted and convicted, which might support the contention popular with criminologists that apart from the period of neutralisation in prison there is not much evidence that the courts affect the behaviour of offenders, even if they affect non-offenders. Subsequent violent offences needed close examination. Fairly often men who were rebuffed in a sexual approach to a stranger snatch the girl’s handbag or punch her. and are thus convicted of robbery or common assault. Sociological. There remains the problem of the possible reasons for increased rape or the increased definition and reporting of rape. Few can doubt that the changes in women’s self image and conception of their status has been affected by physical, educational and economic factors. The physical factor has been the availability of safe contra- ception which enables women to lead a free sexual life without crippling consequences. Much has been written about impressions that those girls who decide for marriage and children tend to marry by the age of 20 to 22, while the career girl decides not to marry till she has reached the peak of her professional competence at about 30, permitting herself short term sexual relationships meanwhile, and only marrying when she decides that it is time to have a family. The traditional male, it is said, who after some dallying decides he is in an economic position to settle down and marry at 26 to 28, finds that there are no eligible girls. The young ones are married, the older not yet interested in marriage, or even meeting the maladjustment of doubting whether any one male will really be suitable. Much no doubt depends upon the sex balance in the population, far too many girls 10 years ago, too few girls now, and later to be a surplus of girls again. The educational factor is related to the economic but since discrimination still ensures that university or grammar school girls are more intelligent than men, one has the impression that men in — 112 — universities must feel that their physical strength is the only indestruc- tible advantage that they still possess. Economically the ever-increasing employment of women, its value and the resulting abandonment of time consuming domestic tasks leads to conflict with those men who were brought up in the traditional view of male dominance, though they do not appear in such traditional patterns of female employment as fishwives and cotton-spinners. It is the wife’s earnings that determine, to the exasperation of the male, whether the family holiday will be in Spain or somewhere else. Clinically, examples of male irritation and devaluation are quite apparent if the change takes place rapidly. In the U.S.A. Kanin (1969) obtained 341 completed questionnaires from 400 male unmarried undergraduates reporting sexually aggressive episodes since they had entered the university. To be accepted as aggressive the man had to affirm having made (1) a « forceful attempt for coitus to the point of being disagreeable or offensive to the girl », and (2) « had to perceive the female responding with offended reac- tions, e.g. fighting, screaming, crying etc. ». Eighty seven (25.5 per cent) reported at least one incident — 181 episodes with 142 females. Às one would suppose, the response of « screaming and fighting » was commonest with pick-ups and first dates, quarreling and «turning cold» among occasional dates, crying and pleading with regular dates. In upper class males the incidents were most often with regular dates, pick-ups and the first dates with working class men and middle class intermediate. In middle and upper class males 77 per cent of the incidents followed fairly regular genital petting on previous occasions, but this was also in the preliminary of 67 per cent of the less frequent pick-up and first date situations. There was much indication of cross-sex communication failure e.g. in the significance of provocative or teasing behaviour, and even more of cross-class failure of communication in assessing the significance of behaviour, where aggression tended to be partly punitive towards what was regarded as gold-digging or provo- cative behaviour. Males with double standards (e.g. believing that 1t is wrong for a girl but not for a man to have sexual intercourse before marriage) described the girl more often as inferior in intelligence and no more than average in attractiveness while the vast majority of « single standard » men spoke of the girl as being much above average in attractiveness and equal in intelligence. The study provided no surprises but did provide systematic confirmation of what most men would expect to be the general pattern of relationships. Schultz and De Savage (1975) obtained questionnaires about the definition of rape and methods of communication of willingness for sexual intercourse from 20 men and 30 girls aged 19 to 24 (who were active daters), and attending «a human sexuality course at a state — 113 — university ». The numbers are very small and the self selection and sampling frame very inadequately described, but the method interesting. They found very inadequate information about what constituted rape legally,- but out of 15 (half) of the girls subjected to. aggression four had been raped,. and 11- considered they had been subjected ‘tô attempted rape. It was of some interest that four of-the twenty meñ reported having been hômosexually raped and two- others of violent homosexual attempts, and none of these had complained to anyone Among the « assaulted » girls there was a hint that those raped more often felt they should not or could not resist, seven thought you ihdicated -consent by'not offering any resistance to increasing intimacy while the non-aggressed girls- had clear ‘ideas “of how consent was communiçated and when a man should stop. The popular conception of male sexuality and how controllable it is clearly confuses the young and inexperienced. Their reaction to the rape tended to be to consult the university medical service about pregnancy and V.D. but other- wnse not to ment10n it. - Such studies clearly suggest that thé criminal law can intervene in such incidents‘only with great caution, and suggest that changing patterns of social life are likely to lead to mcreas…g errors of commumcatmn Carol Smart (1976) in her « femm13t cr1t1que » of cnmmology, emphasises feminine education in passivity and subservience — « the game that women are.-obliged.to, play and-the circumstances in which we participate, namely from a-predominantly vulnerable position with only one valid asset, a basically sexual availability ». It will probably always be a problem that. a girl. who has been forced or cajoled into intercourse against her will feels and has in fact been raped ; but it takes two to-make-a rape, and in law -the man’s belief cannot be excluded. Perhaps men. will have to become much less « reckless ». RÆF. ERENCES : AMIR (M.) : Patterns in Forcible Rape 1971 Ch1cago Umversnty Press CAMBR1DGE UNIVERSITY, DEPARTMENT OF CR1MINAL SCIENCE. « Sexual Offences », Englzsh Studtes in Criminal Sczence, 1957 Vol 9, - London, Macmillan: CHRISTIANSEN (K.0) et al. « Rec1dmsm among sexual ofienders », in Scandinavian Studies in Criminology, London, Tavistock, 1965. COUNCIL - OF EUROPE : Colléctèd Studies in : Chriminological Research ‘Violence in- Society, Vol XI. Strasbourg, 1974. ELLIS (A.) and BRANCALE (R)) : The Psychology of Sex Offenders Thomas Springfield. TIT. 1956. — 114 — GIBBENS (T.C.N.), WAY (C.) and SOOTHILL (K.L.) : « Behavioural Types of Rape », British Journal of Psychiatry, 1977 (n° 130), pp. 32-42. HALL WILLIAMS (J.Æ.) : Serious Heterosexual Attack, » Medicine, Science and the Law, 1977, Vol. 17 (n°2), pp. 140-146. HOWARD (C.) : Australian Criminal Law, 1970. KANIN (E.J.) : Journal of Sex Research, N.Y. Februarv 1969, Vol. S, (n° 1). KARPMAN (B.) : The Sexual Offenders and his Offences, The Julian Press. London, 1954. KRAUS : Depariment of Child Welfare, New South Wales. 1976. See Newton. NEWTON (J.E.) Factors affecting sentencing decisions in rape cases. Australian Institute of Criminology. 1976, Canberra. REPORT OF THE ADVISORY GROUP ON THE LAW OF RAPE. 1975, Cmnd 6352. H.MS.0. SCHULTZ (L.G.) and DE SAVAGE (J.). In Rape Victimology, Thomas Springfield. IIT. 1975. SMART (C.) : Women, Crime and Criminology, Routledge London, 1976. SOOTHILL (K.L.) and JACK (A.) : « The Rape Victim and the Press », New Society, 19 June 1975. SOOTHILL (K.L.), JACK, ANTHEA and GIBBENS (T.N.C.) « Rape a 22 year Cohort Study », Medicine, Science and the Law, 1976, Vol. 16 (n° 1), pp. 62-69. — 115 — L’APPLICATION DU SURSIS PARTIEL PAR LES TRIBUNAUX CORRECTIONNELS FRANCAIS * G. LEVASSEUR M.-P. CHAMPENOIS-MARNIER Laboratoire de sociologie criminelle Université de Paris II - 75005 PARIS France I. — L’INSTITUTION ETUDIEE Le sursis partiel est une institution récemment introduite dans le droit français et qui n’a guère d’équivalents dans les autres législations. Depuis 1884, le droit français connaît le sursis simple (acuellement art. 734-1 à 737 du Code de procédure pénale) : le juge prononce une condamnation à l’exécution de laquelle il sera sursis si le condamné ne subit pas une nouvelle condamnation dans les cinq années suivantes. Depuis 1959 existe, en outre, le sursis avec mise à l’épreuve (actuel- lement art. 738 à 747 du Code de procédure pénale) : l’intéressé est soumis à une mise sous probation pendant une durée de trois à cinq ans au cours de laquelle l’exécution de la peine privative de liberté est suspendue, C’est une loi du 17 juillet 1970 qui autorise le juge prononçant une peine d’emprisonnement à décider qu’il sera sursis, mais seulement de façon partielle, à son exécution, et cela qu’il s’agisse de sursis simple ou du sursis avec mise à l’épreuve (sursis probatoire). Le sursis partiel a, pour le condamné, l’inconvénient de le sou- mettre à l’incarcération pendant une partie de la peine prononcée. En revanche, il présente l’avantage de ne laisser figurer la condamnation qui en est affectée que sur les extraits du casier judiciaire accessibles aux seules autorités de justice. (*) La recherche entreprise depuis quelques années par le Laboratoire de socio logie criminelle (université de Paris II) s’insère dans le cadre des études de critri nologie de la réaction sociale concernant le fonctionnement de la justice. le serten cing et le traitement des délinquants. — 117 — II — OBJET DE LA RECHERCHE La recherche se propose d’abord d’approcher l’importance de la place prise par le sursis partiel (dit aussi « peine mixte») dans l’application concrète du système répressif, et l’évolution éventuelle de cette place. Quel usage les tribunaux font-ils de la nouvelle insti- tution ? Quelles considérations incitent le juge à en user plus ou moins largement ? Elle tend ensuite à établir le profil du délinquant bénéficiaire d’une telle mesure Elle se propose enfin de rechercher les motivations des juges qui octroient le sursis partiel. L’hypothèse la plus couramment avancée par les spécialistes est le souci de « couvrir la prévention » en faisant coincider la partie ferme de la peine mixte avec le temps de détention provisoire subi par l’individu avant sa comparution. Parmi les autres hypothèses figurent, chez certains magistrats, la préoccupation d’assurer chez une exemplarité suffisante par une peine partiellement ferme certains autres, celle de permettre au condamné de mieux se rendre compte du régime pénitentiaire qui serait le sien en l’absence de sursis partiel et qu’il retrouverait en cas de révocation de celui-ci (dans ce cas l’idée de renforcement de l’effet préventif du sursis n’est pas absente). À ce sujet il faut ajouter que l’application du sursis partiel avec les avantages qu’il comporte pour l’intéressé est peut-être de nature à freiner la tendance jurisprudentielle précédente qui, cas d’octroi du sursis simple, augmentait sensiblement la gravité de la peine prononcée (ce dont le législateur amnistiant a été amené à tenir compte). D'aucuns ont pu également supposer que le sursis partiel procède d’un esprit de transaction au sein d’une juridiction collégiale ou constitue un moyen terme entre les réquisitions du ministère public et les conclusions de la défense I. — AMPLEUR DE LA RECHERCHE La recherche a été entreprise en 1974, alors que la loi du 17 juillet 1970 avait plus de trois ans d’existence, en sorte que l’on pouvait penser que la pratique judiciaire avait atteint son rythme de croisière et que le sens de l’évolution était déjà perceptible. Il a fallu néanmoins en proportionner l’ampleur aux moyens réduits dont disposait le Laboratoire de sociologie criminelle, ce qui s’est traduit par une double limitation D’une part, la recherche ne porte que sur l’application de la loi du 17 juillet 1970 par les tribunaux correctionnels ; elle exclut par là — 118 — - même celle qui. est le fait des cours d’assises (qui peuvent appliquer le sursis partiel lorsque par l’effet des circonstances atténuantæ, elles prononœnt des. pemœ d’emprisonnement).… . - D’autre part, l’étude ne retient comme champ d’investigation que quelques tribunaux correctionnels, ‘situés dans des régions diffé- rentes et non un échantillonnage représentatif de toutes les juridictions .Concernées (au total 767 affaires relevées dans six tribunaux : Paris, Lyon Lille, Pau, Bayonne, Tarbes) (1). En fonction des moyens dispo- ‘nibles, ilest en effet apparu préférable.de mettre l’accent sur le qualitatif plutôt que sur le quantitatif, encore que le souci de repré— sentativité ait fortement marqué le ch01x des dossiers dans les juri- dictions approchées. IV. — METHODES a) Collecte des données La collecte des donn«m porte sur deux pomts recherche ‘des éléments objectifs résultant des décisions de justice (étude sur dossiers) ; rœheœhe des données subjectives. intervenant dans cette décision (entret1ens avec lœ mag15trats ayant rendu les décisions étud16œ) l) Recherche des eléments ob ]ectz]‘s a) Recensement exhaustif de toutes les donnees retenues dans les dossiers de 1ugement qui font mtervemr une ou plus1eurs condamnaüons assort1es du sursis parüel ! Cette collecte des données a été menée selon un code de dépoufl lement préétabh — à Pau, Bayonne et Tarbes dans tous les dossiers ayant fait - Pobjet d’un jugement avec sursis partiel en 1971, 1972 et 1973 ; — à Paris, dans toutes les affanes 1ugées en juin 1972 et 1mn 1973 . (1) 314 à Paris (409 Yo), 164 à Lille (21,4 ), 150 à Lyon (19,6 “u), 57 à Bayonne (7,4 ©%), 54 à Pau (7 %), 28 à Tarbes (3,7 % Les grandes aggloméra tions, notamment Paris, sont donc surreprésentées {2) L'’échantillon parisien est apparu suffisamment reprœcntat1f en cifet, la place du sursis simple partiel à Paris au cours de l’année 1972 a été de 1345 °/( des sursis simples, or ‘au mois de juin choisi les dossiers retenus représentent 13,6 © és sursis simples. En 1973, la proportion . pour l’ensemble de l’année a cté dç 18.30 % et celle du mois de juin de 15,65 %. S’agissant de sursis probat01re partiel, les dœsxers de juin 1972 représentent 27,4 © des sursis probatoires alors que la proportion de l’ensemble de l’année est de 34,83 % ; en 1973. la correspondance est moins satisfaisante (30 85 % en 1um au lœu de 2217 @ pour l’ensemble dr l’année). — 119 — — à Lille, dans un dossier sur deux établi durant le premier semestre des années 1971, 1972 et 1973 : — à Lyon, dans les affaires jugées en 1972 et 1973 à raison d’une proportion retenant les variations mensuelles et le mode de saisine du tribunal (3). b) Double codage des fiches de dépouillement et établissement des cartes perforées ; les 80 colonnes des cartes mécanographiques comportent de multiples indications concernant notamment l’individu intéressé, l’infraction commise et ses circonstances, la procédure suivie, la peine prononcée et les recours exercés. c) Mise en place du programme d’exploitation. 2) Recherche des éléments subjectifs. Elle se proposait à l’origine de mieux faire apparaître les moti- vations de la sentence et de pénétrer quelque peu la psychologie dis magistrats ayant rendu les décisions analysées. Elle prévoyait, en conséquence, l’entretien, dans chacun des quatre centres (les trois tribunaux du Sud-Ouest, Pau, Bayonne et Tarbes étant considérés comme un centre unique), avec une dizaine de praticiens (magistrats et avocats) relevant du tribunal intéressé. L’effet envisagé était de confronter les positions idéologiques et pratiques affirmées par les magistrats avec les décisions qui avaient été rendues. Diverses circonstances ont malheureusement retardé la mise en route de cette partie du programme, en sorte que la quasi-simultanéité recherchée n’a pu être réalisée. Les interviews n’ont pu avoir lieu qu’en 1976-1977, alors que l’ tiel des conclusions de la recherche sur dossiers avait déjà pu être dégagé (4). Mais ce retard, malencontreux à plus d’un égard, a toutefois permis de mettre l’accent sur certaines questions précices. (3) Pour 1972, sur 200 sursis partiels relevés (dont 172 sursis simples), 79 dossiers ont été retenus (dont 68 sursis simples, donc mêmes proportions) ; en 1973, sur 158 sursis partiels relevés (dont 129 sursis simples), 68 dossiers ont été rete- nus (dont 58 sursis simples, proportions très voisines) ; les procédures de flagrant délit retenues ont été de 20 sur 34 en 1972 et de 10 sur 21 en 1973. (4) De toute façon, la fréquence des mutations judiciaires et la règle du « rou- lement » risquaient de biaiser le résultat cherché. Notons à ce propos que des re- cherches ont eu lieu d’une part pour comparer les chiffres par année judiciaire et non pas par année civile, d’autre part pour comparer l’activité respective en notre domaine des diverses chambres des grands tribunaux. Les indications qui ont pu en être tirées n’ont pas paru, jusqu’à présent, très significatives. — 120 — L’enquête s’est déroulée par interviews à partir de question- naires (5) comportant à la fois des questions fermées, semi-ouvertes et ouvertes, eux-mêmes établis à partir d’entretiens non directifs faisant fréquemment état de la peine mixte (6). Elle a été menée dans les quatre centres par un collaborateur qualifié différent. La synthèse des réponses recueillies est actuellement en cours d’établissement. b) Exploitation des données. 1) Enquête sur dossiers. Le plan d’exploitation prend évidemment pour base le tri à plat de toutes les données recencées par codage. En fonction de l’un des principaux buts de la recherche, il vise à déterminer les variables qui peuvent expliquer l’octroi par le juge d’un sursis partiel de telle ou telle durée. L’hypothèse de départ, qui tendait à considérer que la partie ferme de la peine prononcée (Pp — Sp) (6 bis) devait, le plus souvent, couvrir la détention (Dp) (6 bis), fut testée. Le processus utilisé aboutit à établir une relation entre ces deux variables, mais ne permit pas de conclure à une relation de cause à effet entre la détention provisoire et la partie ferme de la peine prononcée. En conséquence, d’autres variables explicatives, notamment la gravité des agissements du condamné, furent introduites dans le calcul des X2. Les résultats n’autorisèrent pas à poursuivre dans cette voie — la détention provisoire et la gravité des agissements ne justi- fiaient aucune primauté dans l’explication de l’octroi du sursis partiel. Ils permirent toutefois d’acquérir une vue plus large des questions posées et surtout d’établir une liste des variables explicatives suffi- samment exhaustive pour ne pas encourir la contradiction. Le schéma analytique, finalement utilisé pour mener quinze analyses en composantes principales retient donc trois variables à expliquer (sursis partiel : Sp), mais aussi peine prononcée : (Pp), et peine ferme (Pp — Sp) et dix-huit variables susceptibles d’être dotées de pouvoir explicatif : six sont relatives à la personnalité du condamné —s (3) Des questionnaires distincts, quoique convergents, ont été établis respecti- vement pour les magistrats du siège, ceux du parquet, les juges d’instruction, les juges de l’application des peines et les avocats. (6) Enquête de E. FALQUE auprès de plus de cinquante magistrats. (6 bis) Pp = peine prononcée ; Sp = sursis partiel ; Dp = détention provisoire. — 121 — (âge, nationalité, situation matrimoniale, nombre d’enfants, catégorie socio-professionnelle et sources de revenus) ; huit sont relatives au profil de l’affaire (circonstances favorables, quatre :groupes de: circonstanoes défavorables, antécédents du prévenu, octroi d’un sursis probatoire durée de la. procédure) ; quatre sont relatives aux conditions dans lesquelles-la décision est intervenue (existence et durée d’une détention provisoire, présence ou non d’un avocat, degré de sévérité de la juridiction qui a statué (7) 2) Enquête sur interviews Elle a été menée dans les quatre centres- choisis auprès de cinq praticiens : — 24 mag13ttats du snege (8) ; — 8 membres du parquet (9) — 6 juges d’instruction (10) ; . —— 4 juges de l’apphcatnon des pemæ (11) ; —— 9 avocats (12) ‘Les rcponses ‘obtenues ont été réahseœ synthéüsées et comparéæ analyse&s par fonctions judiciaires . ou. parawd1c1mres respectives et par région. 2 — PREMIERS RESULTATÊ 8 ? » a) Résultats de l’enquête sur dossners, resultats decoulant du tri à plat. 1) Dannées tenant 2 l’infraction et. à sa poursuite. a) Les infractions ayant donné lieu au sursis partiel sont dans la très. grande majorité des infractions contre les biens ; près des 3/4 t A7). Pamn les autres cléments recueillis dans les tiches dc dépouilluinent, ont Cu I“als sés de côté pour l’analyse en composantes principales : le sexc du détinquant (la”Proportion de femmes bénéficiaires du sursis partiel n’est que de 55 “o'au ‘rotal), le Heu de naxssance le domicile, la catégorie socio-professionnelle, la nature de L'infraction jugée, le. lieu de sa commission, le mode de saisine du tribunal, la K1u;ee de Ja procédure, le caractère contradictoire ou pat défaut de la décision, la ° ° mise’sous contrôle judiciaire, le caractère simple ou probatoire du sursis, les moda- lités du sursis probatoxre les peines complémentaires, l’existence d’une partie civile. les voies de recours, etc. La situation résultant des examens en question est néan moins retracce dans le rapport general 8) 4+ à Paris, 6 à Lyon, 5 à Lille, 9 au centre Pau-Bayonne-Tarbes (9) Î mr Paris, 2 à Lyon, 2 à Lille, 3 au centre Pau-Bayonne-Tarbes (10) Pans 2 à Lyon, 1 à Lille, d ” au centre Pau-Bayonne-Tarbes (11) Lyon 9 2 à Lille, 1 au centre Pau-Bayonne-Tarbes (12) ÈÊË Paris p à Lyon, 2 à Lille au centre Pau-Bavonna-Tarb, — 122 — (73,9 %) si l’on s’en tient aux dossiers comportant une seule infraction, 93,8 % si l’on y ajoute ceux dans lesquels la poursuite vise à la fois deux infractions (39,8 % des cas), voire trois (14,6 %). Les atteintes contre l’intégrité corporelle représentent cependant 10,4 % (dont 8,7 atteintes volontaires). La délinquance contre les biens est particu- lièrement forte (78,1 %) à Paris ; au contraire elle n’est que de 56,1 % à Lille, où les atteintes à la police de la route représentent 18,9 % (au lieu de 4 % en moyenne) à raison d’une jurisprudence particulière de ce tribunal. En effet, la chambre de ce tribunal qui juge les délits de la route inflige volontiers une courte peine de prison assortie du sursis pour la majeure partie de la peine (ne laissant à exécuter que huit à dix jours). b) Le tribunal correctionnel a été saisi dans 18,6 % des cas (26,8 % à Paris, 21,3 % à Lyon) par la procédure de flagrant délit, et seulement dans © des cas par citation directe du ministère public (31,1 % Lille où le sursis partiel intervient systématiquement pour de nombreux délits routiers ou accidents de la circulation). C’est dire que la grande majorité des poursuites (66,5 %) a lieu par la voie de l’information (proportion nettement plus élevée dans le Sud-Ouest et à Lyon), ce qui expliquera la présence fréquente de détention provisoire c) En effet, 84,5 % des prévenus ont comparu en état de détention (y compns ceux des jugés en flagrant délit) surtout à Paris (92.7 % e) ; (57,3 ©e) à Lille pour les raisons déjà indiquées. Cette detenüon est inférieure à six mois dans 88,5 % des cas (21,8 % moins de quinze jours, à raison notamment des flagrants déhts) elle ne dépasse un an que dans 1,5 % des cas. Remarquons à ce propos qu’un tiers des jugements (32,7 %) interviennent moins de trois mois après l’infraction (52 et 53 % dans le Sud-Est) et que les durées les plus longues sont enregistrées à Paris (19,4 % supérieures à un an) et à Lyon (25,3 %) Le contrôle judiciaire, qui était d’ailleurs à ses débuts à l’époque (il est contemporain du sursis partiel), se rencontre dans 8,7 % des cas (sauf à Lyon où il atteint 21,3 %) d) Les décisions sont, dans la quasi-totalité des cas, contradictoires (93,5 %) ; dans 77,2 % des cas, les prévenus étaient assistés d’un avocat (90 % dans le Sud-Ouest ; 60,4 % à Lille). 2) Données tenant à la condamnation Par hypothèse même, les jugements ont prononcé une peine d'emprisonnement pour une au moins des infractions, objets de Ja poursmte (4,6 % de relaxes partielles sont intervenues, presque exclu- sivement dans le Sud-Ouest et à Lyon). Cet emprisonnement est 123 — prononcé seul dans 78 % des cas, est accompagné d’une amende dans 11,9 % des cas (notamment à Paris et Pau) et de peines complémen taires (retrait de permis, interdiction de séjour, confiscation) dans 4,7 Ÿ des cas seulement a) L’emprisonnement prononcé est, dans la majorité des cas (54,1 %) inférieur à neuf mois (un peu plus souvent à Lyon et moins souvent à Paris) il ne dépasse dix-huit mois que dans 12,2 % des cas b) Par hypothèse également ces condamnations comportent un sursis partiel ; sa durée s’échelonne (95,7 % des cas) entre huit jours et quinze mois (65,4 % au-dessous de quatre mois, 68,1 % au-dessous de six mois, c’est le cas de la totalité des dossiers de Pau et de Tarbes au contraire 1/4 des dossiers de Bayonne dépassent cette durée) Le sursis partiel accordé est, dans 75,7 % des cas un sursis simple (86 % à Lyon, 84,4 Ÿ à Paris) (13). Dans le cas d’un sursis avec mise à l’épreuve, la durée de la probation est généralement de 3 ans (74,1 %, sauf à Lyon) c) Un des points les plus importants est celui de savoir si la période d’emprisonnement ferme correspond approximativement à l’incarcé- ration d’ores et déjà subie. On sait en effet que l’objectif de « couvrir la prévention » est l’hypothèse la plus couramment admise concernant l’utilité du sursis partiel La recherche permet de constater que cette tendance n’est pas néghgeable mais qu’elle n’ést pas majoritaire, et que son influence paraît même avoir tendance à diminuer. Le « surplomb» dû à la partie ferme du sursis part1el ne correspond (avec une marge de 15 jours) à la détention provisoire que dans 40,7 % des'cas (il est cepen- dant nul dans 24.9 “c) par contre, il excède 3 mois dans 19,6 % des cas Une exploitation manuelle a montré qu’à Paris, en juin 1972 43 % des décisions correspondaient à la détention provisoire, que 23 % (dont pour les 4/5 des flagrants délits) intervenaient sans détention provisoire appréciable, et que 34 % dépassaient notablement celle-ci En juin 1973, les proportions étaient respectivement de 29 %, 25 % (dont 2/3 de flagrants délits) et 42 %, Il semble donc établi que le souci de « couvrir la détention » n’est pas l’objectif essentiel de l’application du sursis partiel. Faut-il se pencher alors sur la personnalité du délinquant ? (13) Un tri croisé doit permettre de savoir si les sursis partiels avec m15e à l’épreuve aurait pu être accordés comme sursis simples. 124 — 3) Données tenant à la personnalité du délinquant. a) CARACTERISTIQUES INDIVIDUELLES. Aucune conclusion ne paraît pouvoir être tirée du sexe du prévenu ; la proportion des femmes dans l’ensemble approché est de 5,5 %, moins élevée dans le Sud-Ouest, plus élevée dans certaines chambres du tribunal de Paris. Cette proportion n’excède pas celle que l’on re- trouve dans les peines d’emprisonnement en général et notamment les L pein es de prison ferme (14). — Par contre, la jeunesse de la population concernée est assez remarquable. La grande majorité se situe entre 18 et 35 ans (81,1 %) (15), et même entre 18 et 30 ans (71,8 %). Une importante fraction relève de la tranchie de 18 à 21 ans (16). Il est vrai que, si la crimi- nalité générale est forte dans ces tranches d’âge, le passé judiciaire est généralement moins' chargé. — Assez remarquable également est la proportion d’étrangers plus du quart de l’ensemble de l’échantillon (25,6 %) (17) ; plus élevée, naturellement, dans les grandes villes, telles Paris (34,10 %) et Lyon (27,3 %), elle est cependant plus modérée dans le Sud-Ouest et à Lille en dépit de la proximité d’une frontière. La grande majorité des prévenusî étaient nés dans la région où ils ont été condamnés, et y étaient également domiciliés. À ce sujet, on remarque cependant la forte proportion de personnes sans domicile fixe (12,9 %) parmi les bénéficiaires du sursis partiel, proportion dont les variations locales sont cependant très sensibles (18). Lorsque le domicile est connu et que son indication a pu être relevée (elle manque dans 15 % des dossiers), la dominante urbaine est frappante (35,7 % domiciliés dans des villes de plus de 100 000 ha- bitants, 14,9 Ÿ dans des villes de moins de 10 000 habitants, ou en Zzone rurale) avec toutefois certaines différences régionales. Ainsi, en schématisant ces données, l’intéressé concerné par l’enquête répond approximativement au portrait-robot suivant : individu (14) Cf. LUVASSEUR. « La femme et le droit pénal », (rapport au congrès de VIDEF.) Indénendance ei coopération, 1974, p. 1217. (15) 89,1 7 à Paris, 82 % à Lvyon, 100 % à Tarbes. (161 233 G, atteignant parfois 29,4 % dans certaines chambres de Lyon et même 35.7 “ dans certaines chambres de Lille, pour la raison déjà aperçue ; la proportion est blus forte encore dans le Sud-Ouest de 35,2 % à 429 ®. (17) En 1973, la proportion des étrangers jugés par les tribunaux correctionnels était de 15 Co. (18) 222 % à Pau. 17.5 © à Bavonne, 15,6 % à Paris, 9,1 ©o à Lille, 8,7 % à *yon, O % à Tarbes. — 125 — de sexe masculin, en général jeune, le plus souvent français, ne mani festant pas, sauf s’il est étranger, une forte propension à s’éloigner du lieu de naissance et domicilié surtout en zone urbaine b) CARACTERISTIQUES FAMILIALES Les indications relatives à la situation matrimoniale sont fragmen- taires et peu sûres, Les célibataires sont nettement en majorité (58,7 % contre 29,6 % de mariés et peu de divorcés ou veufs), ce qui est en relation évidente avec l’âge des prévenus mæs comporte de larges diffémences régmnales (19). En ce qui concerne le nombre des enfants chez les bénéficiaires du sursis partiel (il faut noter qu’une lacune affecte sur ce point 27,1 % des dossiers), le cas le plus fréquent est celui où il n’y a pas d’enfant (53 % des dossiers contenant l’information) ce qui correspond assez naturellement à l’âge et à la situation matrimoniale (20) c) CARACTERISTIQÜES soc10—mom—:ssmmm.æs Sur le niveau d’instruction, les informations manquent dans plus d’1/3 des dossiers examinés (et même pour la majorité des dossiers de Paris). À Lyon (où la lacuhe ne touche que 12,7 % des dossiers) le niveau primaire est majoritaire (51,3 %) de même à Lille et à Pau. Le -niveau infraprimaite est relevé dans 11,6 % de l’ensemble, soit 1/6 de la population pour laquelle l’information est donnée. De même, la catégorie socio-professionnelle n’a pas été mentionnée dans plus du tiers des dossiers. En s’en ‘tenant aux dossiers contenant cette information, on constate la surreprésentation de la catégorie des employés (55 %, de la population connue sur ce plan) (21). On est frappé par la sous-représentation des ouvriers (7,2 % de l’échantillon, 10 %, des catégories socio-professionnelles connues) qui ne sont pas davantage représentés que les cadres supérieurs et patrons de l’industrie ou du commerce (au total 13,5 % de l’échantillon, soit 19 % de la (19) 857 © de célibataires à Tarbes, 68,7 “ à Pau, 65,6 % à Paris avec de larges variations entre les chambres, 56,1 © à Bayonne, 52,3 % à Lvon, 451 à Lil! (20) La p…sc cu d'un enfant se relève dans 12,4 o de l’échantillon (correspon dant à 17 © tillon 1 — I75 c, » de la population concernée), celle de 2 enfants dans 9,1 % de l’échan & de la populanon) celle de 3 enfants dans 6,4 © de l’échantillon celle de 4 enfants. 3.7 Go celle de 5 enfants et plus. 2,6 %%. L'absence d’enfants est particulièrement élevée à Tarbes, à Patis et dans le Sud-Ouest elle est beaucoup plus rare à Lil! (21) Il v a peu de variations locales (57 © à Paris et à Lille ; 55 o à Pau sauf à Tarbes ct à Bavonne (80 ©) — 126 — population connue . professionnellement) sur ce point, les variations locales ne sont pas très sensibles (22). Si la majorité des. condamnés occupait un emploi lors de. l’m£rpo— tion (54,6 %), une fraction importante était alors en chômage (19,4 % pour l’ensemble, allant de 10,4 % à Paris à 30,7 % à Lyon) ; il. faut noter en outre que 9,5 % étaient sans professxon (avec des d1fiéœnœs locales appréciables). - - La recherche d’md1œüons sur les revenus nænSuels s’est révél& déœvante et stat13üquement mexplontablc ) 9 CARACTERISTIQUES CONCERNANT LE COMPORTEMENT. — Les informations qui se rapportent tant à la oondmtç qu la reputaüon la moralité ou l’éthylisme sont peu significatives, . en raison d’une part de leur généralité, d’autre part de l’absence de mention sur ce point dans la :moitié des -dossiers (dont plus des 2/3 des dossiers de Pans) Tout au plus, peut—on observer que le sur&s partiel a été aœordé à 3ù % des prévenus connus pour leur mauvaise conduite, à 26 % de ceux réputés de mauvaise réputation, 23 % de ceux indiqués cernme étant’ de mauvaise moralité, et 20 % des alcooliques. — Le casier md1c1ære est curieusement absent de 8,7 % des dossiers.-Dans le cas où il figure, l’absence de condamnation est la règle pour les 2/3 et parfois les 3/4.des cas (73,2 % à Paris)-sauf à Tarbes et. Pau (50 %), ce qui est normali, mais non nécæsaxre légale— ment, pour bénéficier d’un sursis. Parmi les 27,2- % des intéressés qui ont un :casier 1ud1amre, l3 6..% de l’échantillon n’ont qu’une. seule condamnation, 27 % ont plus de 3 (ce cas se relève principalement à Lyon). — Quant aux antécédents judiciaires résultant d’autres. élflmmts que le casier (condamnations amnistiées signalées par la police ou les intéressés, mesures de rééducation, intervention des organismes de contrôle .social), on -observe- des mesures éducatives dans 8,6.% de l’échantillon (soit 16 % des intéressés connus) ; 7,7 % (soit 14,3 % des intéréssés) sont s1gnalés par la police comme connus de ses S€I'Vle«b il.y aurait 4 % , de condamnations amnistiées, et 2,3: % de sanctions admxmstraüves Dans un nombre limité de cas, un examen psych0bglque ou psychlatnque est mtervenu Il en résulte que près de 45 Œ de la populatmn p0ur laquelle on dispose d’informations est déjà connue pour une certaine forme de délinquance, à tout le moins de déviance, er 15 (22) Les ouvriers xqvnscmcnt 9 ‘o à Lille, 11 % à Paris et Pau 13 5 à l‘a°‘lx< a [von. de ! ‘a ponulation profe<smnndlemmfl connue. — 1%7:— 4) Données tenant aux circonstances de l’infraction En dépit de lacunes regrettables dans les dossiers, il apparait que les éléments tirés de la personnalité du délinquant sont moins éclairants que ceux qui tiennent à l’infraction et à sa poursuite, pour expliquer l’usage du sursis partiel par les juges correctionnels Le plan de recherche attendait beaucoup de l’analyse des moti- vations de l’octroi de ce sursis — il avait été recommandé, lors du dépouillement des fiches, de noter intégralement et littéralement tout ce qui motivait la décision — il faisait preuve hélas d’une grande naïveté. En effet la motivation est, dans la plupart des cas inexistante, à tout le moins, extrêmement sommaire. Sur ce plan, les données recensées se réduisent, à défaut du néant (dans 19,4 % des dossiers le texte du jugement n’est pas connu) à des indications tout à fait succinctes l’existence des conditions légales d’octroi du sursis partiel n’est accompagnée d’aucune autre motivation dans 43,7 % des cas Le jugement comporte une nuance de bienveillance dans 25,9 % des dossiers, une motivation bienveillante détaillée dans 3,7 % seu- lement Les décisions avec nuance de sévérité sommaire visent 2 % échantillon, avec motivation sévère détaillée : 09 % Face à cette situation bien dégevante, il s’avérait alors nécessaire de rechercher dans l’ensemble des dossiers et, notamment dans les circonstances de l’infraction, les éléments qui avaient pu impressionner favorablement ou défavorablement le magistrat (23) Un affinement des circonstances favorables et surtout défavorables fut ainsi mis au point à partir des d1fiérents éléments présents dans les dossiers. Aux circonstances favorables qui pouvment tenir à l’âge, au sexe, la situation de famille, à la catégorie socio-professionnelle, aux ressour- ces, au casier judiciaire, aux antécédents psychiatriques, au début de sanction résultant de la détention provisoire, furent ajoutées celles qui pouvaient être tirées, notamment de la faiblesse ou de l’absence de pré- judice, de la réparation de celui-ci, de l’attitude au moment de la constatation des faits, ainsi que de certaines déficiences mentales Quant aux circonstances défavorables qui pouvaient être relevées dans l’ensemble du dossier, elles furent regroupées en quatre catégories (23) Par hypothèse meme, l'affaire devait présenter certaines circonstances favo rables, sinon le juge eût prononcé une peine de prison ferme mais elle devait aussi présenter des circonstances défavorables sinon le juge eût prononcé le sursis total (simple ou probatoire) — La première (catégorie A) retient, à titre unique ou joint à d'autres éléments, le port d'armes et la correctionnalisation d’une affaire criminelle à l’origine. — La seconde (catégorie B) retient dans les mêmes conditions, le concours réel et la réitération. — La troisième (catégorie C) consiste dans la participation de plusieurs personnes au fait délictueux. — La quatrième (catégorie D) retient l’importance du dommage infligé et les facilités à commettre l’infraction qui tiennent soit à la personnalité de la victime, soit aux commodités d’accès. C’est en fonction de cette nouvelle approche des circonstances de l’espèce que l’on s’est efforcé de discerner les raisons qui avaient pu déterminer le tribunal, par un savant dosage d’indulgence et de sévé- rité, à prononcer une peine assortie du sursis partiel. Contrairement aux attentes les circonstances favorables, tirées d’ailleurs, à titre essentiel, de la faiblesse du préjudice ou des défi- ciences mentales, n’apparaissent que dans 1/4 des dossiers (26,1 %). Cette constatation n’exclut pas le rôle de ces circonstances dans l’octroi du sursis partiel par préférence à une peine ferme. Mais celles-ci se présentent avec un caractère si parcellaire qu’il est diffi- cile de leur reconnaître un pouvoir significatif. En revanche, il n’en est pas de même en ce qui concerne les circonstances défavorables. Certes, en fonction du nouveau codage retenu, deux catégories de ces dernières peuvent être relevées dans un même dossier (environ dans un cas sur deux). Mais, même en tenant compte de cette remarque et à ce niveau purement descriptif, l’importance des circonstances aggravantes paraît justifier le choix du sursis partiel par préférence au sursis total. Le tri à plat se conformant au nouveau codage fait en effet apparaître l’existence de circonstances aggravantes : — De type B (concours réel ou réitération) dans près des 2/3 des dossiers (63,1 %) e - — De type D (importance du dommage ou facilités de l’in- fraction) dans plus de la moitié (52,8 %). — De type C (participation) dans plus de 1/3 (37,7 %). — De type À (port d’armes ou correctionnalisation) dans 16,8 % €s cas. b) Résultats découlant des analyses en composantes principales. L'analyse en composantes principales présente évidemment un intérêt beaucoup plus grand que l’analyse purement descriptive précé- — 129 — dente tirée du tri à plat, puisqu’elle tend à établir des relaüons cau- sales entre variables explicatives et variables à expliquer. Parmi les 15 analyæs en composantes prmmpalœ qui ont été menées pour expliquer la peine prononcée (Pp), la peine ferme (Pp—Sp) le sursis partiel (Sp), à partir de 18 variables considérées comme explicatives (24), 10 seulement ont été retenues, en raison des effectifs très faibles mis en cause dans les autres a) En ce qui concerne l'importance de la partie ferme de la peine seuls sont retenus ici Jes résultats de la seconde A.C.P. dont l’échantillon comprend uniquement les dossiers dans lesquels la détention provisoire est supérieure à 15 jours (25). (On peut, en effet, supposer que pour les dossiers qui comportent une détention provisoire inférieure à 15 jours, la ‘peine ferme, variable à expliquer, se confond avec la peme pro- noncée, autre variable à expliquer.) - Cette A .C.P. fait apparaitre une forte corrélation positive de la peine ferme avec la détention provzsozre, à un moindre degré avec l’importance du dommage causé. Brevitatis causa, plus la détention provisoire est longue, plus le préjudicé de l’infraction est lourd, plus la part1e ferme de la peine prononcée s’allonge. En ‘revanche, cette dernière est corrélée négativement avec la catégorie socio-professionnelle plus le condamné se situe dans les mxheux dits défavorisés, plus la durée de la peine ferme diminue. Les œrräaüons de la peine ferme avec le miveau d’instruction et.les antécédents autres que ceux figurant au casier judiciaire sont établies également, mais elles revêtent une nature plus complexe (26) b) En ce qui concerne la peine prononcée, quatre analyses en oomposantes prmapalcs tentent de l’expliquer. Les deux premières (A.C.P. 1 ct A.C.P. 2) portent sur l’ensemble des dossters. Les deux dernières (A.C.P. 3 et A.C.P. 4) retiennent seulement ceux dans lesquels la détention provisoire dépasse 15 ÿ Jours. -(24) Cf supra.… (25) La première A.C.P. concernait l’ensemble des dossiers. (26) — Plus le niveau d’instruction baisse, plus la durée de la peine ferme a tendance à se situer soit dans les temps très courts, soit dans ceux qui dépassent la movenne, sans toutefois jamais atteindre les extrêmes supétieurs ; — plus les antécédents autres que ceux figurant au casier 1udœ1a1re sont nom- breux, plus la- durée de la peine ferme tend à se situer, soit aux alentours de la moyenne, soit dans les -durées les .plus longues 130 — La distinction entre A.C.P. 1 et À.C.P. 2 d’une part, entre A.C.P. 3 et A.C.P. 4 d’autre part, tient à l’introduction d’une variable explicative supplémentaire, savoir la durée du sursis partiel. Il est remarquable, quant aux résultats, que cette dernière adjonc- tion ne modifie en rien la structure des autres variables pouvant expli- quer la peine prononcée. En d’autres termes, la durée du sursis partiel est munie d’un fort pouvoir explicatif dans À.C.P. 2 et A.C.P. 4, mais les variables qui expliquaient la peine prononcée quand le sursis partiel ÿ’f:ntrait pas en jeu, dans A.C.P. 1 et A.C.P. 3, restent inchangées. Il suffit, en conséquence, de présenter les résultats des A.C.P. 2 et A.C.P. 4. |. L’analyse en composantes principales qui porte sur l’ensemble des dossiers et retient parmi les variables explicatives le sursis partiel (A.C.P. 2), fait apparaître une forte corrélation positive entre la peine prononcée, variable à expliquer, et les variables explicatives suivantes : — circonstances défavorables de type D ; — présence de l’avocat ; — sursis avec mise à l’épreuve ; — durée de la détention provisoire ; — durée du sursis partiel. Elle établit également une forte corrélation négative entre la peine prononcée et les variables : catégorie socio-professionnelle et Œœnstæmœs favorables. En schématisant, plus l’importance du dommage causé s’accroît : — plus la modalité du sursis partiel accordé est un sursis probatoire ; — plus la détention provisoire est longue ; — plus la durée du sursis partiel s’allonge ; — plus la peine prononcée a des chances d’être de longue durée. En outre, la présence de l’avocat, liée peut-être à une affaire importante, est corrélée positivement avec une peine forte. En revanche, plus la catégorie socio-professionnelle du condamné Se situe dans les niveaux bas : — plus les circonstances favorables sont nombreuses ; — plus la durée de la peine prononcée tend à se raccourcir. 2. L’A.C.P., qui éliminc les dossiers comportant un détention provi- Soire inférieure à 15 jours mais conserve le sursis partiel comme variable explicative supplémentaire (A.C.P. 4), retrouve les corrélations posi- — 131 — tives de la peine prononcée avec les circonstances aggravantes tenant à l'importance du préjudice, avec la durée de la détention provisoire et avec celle du sursis partiel. Dans cette A.C.P., le pouvoir explicatif de la présence de l’avocat disparaiît, sans doute parce que dans les affaires graves, où la détention provisoire est importante, la variable qu’il représente devient neutre. En revanche, nouveauté par rapport à l’A.C.P. précédente, la régularité de la situation (quant au travail et quant aux revenus) paraît corrélée positivement avec la peine prononcée plus les intéressés ont une «situation régulière », plus la peine prononcée a tendance à s’allonger… Si la catégorie socio-professionnelle et l’importance des circons- tances favorables ont perdu leur pouvoir explicatif, le niveau d’instruc- tion en paraît muni dans cette A.C.P. : plus celui-ci baisse (27) vers niveau primaire, infra-primaire, plus la peine prononcée tend à s’établir dans les courtes durées. c) En ce qui concerne la durée du sursis accordé, une remarque du même ordre que celle qui a été formulée au sujet de la peine prononcée doit être faite dès l’abord. L'intégration de la variable explicative supplémentaire peine prononcée, qui seule distingue les deux premières A.C.P. portant sur l’ensemble des dossiers et les deux dernières ne retenant que cux qui comportent une détention provisoire supérieure à 15 jours, ne modifie pas la structure des autres variables. Il suffit, en conséquence, de rapporter les résultats des A.C.P. 2 et A.C.P. 4. [. L’analyse en composantes principales qui retient l’ensemble des dossiers et ajoute aux autres variables explicatives la peine pro- noncée (considérée sur le plan général comme variable à expliquer) établit, d’une part de fortes corrélations négatives du sursis accordé avec les circonstances favorables, le niveau d’instruction et la q d’étranger : qalité — plus les circonstances favorables sont nombreuses, plus la durée du sursis tend à se raccourcir ; — plus le niveau d’instruction baisse, plus cette durée est courte. En outre, la durée du sursis paraît d’autant plus restreinte que celui-ci est accordé à un étranger. (27) Les variations sur ce plan étaient codées en partant du niveau d’instruction le plus haut. La corrélation est donc négative. — 132 — Quant aux corrélations positives, elles sont nombreuses. Elles concernent, notamment, le lien causal entre sursis accordé et peine prononcée : — plus cette dernière est longue, plus elle entraîne l’octroi d’un sursis de longue durée. L’A.C.P. concernée établit également le pouvoir explicatif impor- tant — de la détention provisoire plus celle-ci est longue, plus la durée du sursis accordé l’est aussi, ce qui implique que le souci de « couvrir la prévention » n’a qu’un domaine limité ; — des circonstances aggravantes, faisant intervenir l’importance du dommage (type D), la participation d’autres personnes (type C) et la correctionnalisation (type A) : plus ces circonstances sont nombreuses, plus la durée du sursis s’allonge. Cette constatation, surprenante à première vue, s’explique peut- être par l’importance des affaires en cause qui, par hypothèse, en- traînent une peine lourde, mais en valeur absolue aussi, un sursis de longue durée. La présence de l’avocat, quant à elle, paraît favoriser l’octroi d’un sursis important. 2. L'analyse en composantes principales qui élimine les dossiers mentionnant une détention provisoire inférieure à 15 jours, mais retient comme variable explicative supplémentaire la peine prononcée, fait apparaître des résultats identiques à la précédente, en ce qui concerne la relation causale : — du niveau d’instruction ; — de la durée de la détention provisoire ; — de la peine prononcée ; — des circonstances aggravantes de type D et À avec le sursis accordé. En revanche, la corrélation positive de ce dernier avec la présence de l’avocat et les circonstances aggravantes de type C disparaît. Mais cette A.C.P. fait apparaître un certain rôle du statut matri- monial. Les célibataires, en particulier, paraissent tendre à bénéficier d’un sursis de courte durée. À ce sujet, il convient de noter que l’importance des célibataires dans l’échantillon produit peut-être un effet de masse qui annulerait le_ résultat. La même remarque vaut en ce qui concerne le niveau ‘instruction qui est surreprésenté aux bas niveaux. — 133 — Les conclusions à tirer du tri à plat et des analyses en composantes principales appliquées à l’enquête sur dossiers paraissent être essentiel- lement les suivantes Le rôle de la détention provisoire est indéniable ; elle est explica- tive à la fois pour la partie ferme à exécuter (Pp—Sp), pour la durée de la peine prononcée (Pp) et pour la durée du sursis accordé (Sp), même si le sursis partiel ne « couvre » la détention provisoire accomplie que dans moins de la moitié des cas (supra p. 122). Plus la détention provisoire est longue, plus la peine prononcée est forte, plus le sursis accordé est étendu, plus la partie ferme à exécuter est importante De même la circonstance défavorable du dommage causé (qui, jointe aux facilités pour commettre l’infraction, paraît dans 528 % des affaires) est fortement corrélée, elle aussi, avec chacune des trois variables à expliquer (Pp — Sp ; Pp ; Sp) Enfin, la présence de l’avocat, qui se relève dans la majorité des cas, présente une corrélation positive avec chacune des variables à expliquer En outre, le degré d’instruction et son corollaire, la catégorie socio-professionnelle, ont un certain pouvoir explicatif (mais à raison d’une corrélation négative) avec chacune des variables à expliquer ; cependant leur incidence est plus complexe (28) Enfin, quoique les étrangers semblent en proportion anormale parmi les bénéficiaires du sursis partiel (supra p. 123), la nationalité n’apparaît pas dans les A.C.P. relatives à la peine prononcée (Pp) ; en revanche, elle a un pouvoir explicatif en ce qui concerne la durée du sursis partiel (Sp) et la partie ferme de la peine (Pp — Sp). Elle tend à réduire la durée de l’un aussi bien que celle de l’autre Il faut noter, en outre, la relation qui apparaît entre la peîne prononcée et la durée du sursis partiel l’une croît toujours en fonction de l’autre et réciproquement (29) (28) Plus la catégorie socio-professionnelle se situe dans les classes défavorisées plus la peine prononcée et plus la peine ferme sont faibles pour le sursis partic ces catégories tendent à se situer dans la moyenne (29) On verra que cela paraît correspondre à un réflexe spontané des magistrats qui semblent proportionner la durée du sursis accordé à la longueur de la peinc prononcée (infra). — 134 — Les autres données, telles l’âge, la situation matrimoniale, le nombre d'enfants ne jouent que de manière très ponflueÿe (3Q). ._ b) Résnitats de l’enquête sur entretiens. Les questionnaires utilisés pour l’enquëtè menée dans chacùn des quatre centres respectivement auprès des magistrats du siège, des membres du ministère public, des juges d’instruction, des juges de l’application des peines et des avocats comportaient plusieurs parties. Seules seront analysées ici les réponses aux guÿçstions œsen‘tîcllæ. 1) Le principe même de la peine mixte est accueilli de façon unanimement favorable et parfois enthousiaste par les magistrats du siège et ceux du parquet. Les juges d’instruction sont à peine plus réservés, encore que l’un deux observe, justement, que cette peine mixte équivaut à une courte peine de prison. Il en est de même chez les juges de l’application des peines ; l’un d’eux perçoit le sursis partiel comme un empiètement sur la libération conditionnelle, Néanmoins, pour les uns et les autres, les avantages de cette nouvelle modalité de sanction l’emportent nettement sur les inconvénients. L’institution recueille aussi une très large approbation auprès des avocats. 2) L’utilisation de la peine mixte est signalée comme fréquente par toutes les catégories interrogées. Toutefois, lorsque la question est posée particulièrement pour le sursis avec mise à l’épreuve, la réponse est légèrement en retrait (31). La division et l’embarras l’emportent, quelle que soit la catégorie des praticiens interrogés, lorsque est demandé si le sursis partiel est plus indiqué ou plus efficace pour le sursis simple ou pour le sursis probatoire (32). (30) L'’âge apparaît comme variable explicative simplement en ce qui concerne à peine prononcée (Pp) ; plus les intéressés sont âgés, plus ils ont tendance à être condamnés, soit très fortement, soit très faiblement. La situation matrimoniale n’apparaît pas dans les A.C.P. relatives à la peine ferme (Pp — Sp) ; par contre, la qualité de célibataire paraît entraîner une peine (Pp) de durée moyenne, et un sursis (Sp) de durée faible ou moyenne. Le nombre d’enfants ne présente de cor- rélation qu’avec le variable Sp, et semble conduire à un sursis partiel se situant aux extrêmes ; on verra (infra) que cette constatation est malaisément conciliable avec les intentions des magistrats. {31) Il est à noter cependant que le bénéfice du sursis probatoire total est accordé « souvent » et même plus fréquemment depuis trois ans. Magistrats et avo- cats s’accordent pour approuver totalement les lois récentes élargissant le domaine du sursis probatoire. (32) Le parquet et les juges d’instruction pencheraient plutôt vers le sursis probatoire ; mais au sein du même tribunal, les magistrats du siège ou du parquet Optent différemment. — 135 — Il en est de même de la réponse à la question « le sursis partiel sera-t-il plus exposé qu’un sursis total à la révocation par une condam- nation ultérieure ? » (33). 3) Les raisons qui amènent le juge à prononcer une peine mixte ont fait l’objet de diverses questions, dailleurs combinées et. imbibées en vue de corroborations réciproques : a) C'’est incontestablement la personnalité du prévenu (sous ses divers aspects) qui, aux yeux de la grande majorité des magistrats, constitue le critère fondamental qui les guide (34). Au contraire, la nature de l’infraction, contrairement à ce que l’on eût pu penser, est loin d'être déterminante (35) même pour l’octroi d’un sursis total. Mais quels éléments, dans la personnalité du prévenu, apparaissent assez favorables pour éviter à celui-ci une peine ferme, assez défa- vorables pour lui refuser un sursis total ? Les magistrats du siège interrogés sur ce point (35 bis) font essentiellement état de cinq éléments déterminants : — les charges familiales (36) ; — le jeune âge (37) ; (33) Il faut noter que depuis une loi du 11 février 1975, le tribunal correctionne! peut décider que la peine d’emprisonnement qu’il prononce ne révoquera pas un sursis antérieur. (34) Un seul juge du siège estime cette considération inopportune, ct un seu! avocat pense que les juges n’en tiennent pas compte. Au contraire, les parquetiers, les juges d’instruction, ct les juges de l’application des peines sont unanimes à v voir la clé du problème, tant pour lc sursis en général que pour le sursis partie! ui nous occupe, mais avec moins de fermeté lorsqu’il s’agit du sursis avec mise à l'épreuve. (35) Les réponses sont partagées, dans toutes les catégories de praticiens intær- rogés. Cette considération n’est admise que par une faible majorité des juges du siège ; seuls les parquetiers (presque unanimement) et les juges d’instruction Jui attribuent une place plus marquée. Les avocats pensent que les tribunaux en tiennent assez peu compte. Sur une question plus précise, les magistrats du siège, partisans de tenir compte de la nature de l’infraction, ont estimé le sursis partiel indiqué dans le cas d'infrac- tions « graves » ou correctionnalisées (5 cas), d’infractions violentes (4 cas) ou au contraire astucieuses contre les biens, notamment les affaires financières (3 cas), et (chez les magistrats de Lille) les délits routiers (2 cas). Lorsqu’il s’agit d’un sursis probatoire partiel, on ajoute l’abandon de famille (3 cas). (35 his) L’interrogation s’est faite en deux temps on a énuméré tout d’abord divers éléments dont on demandait s’il fallait tenir compte de façon fréquente. modérée ou nulle ; puis les personnes interrogées ont été invitées à faire connaître quels étaient à leurs yeux, parmi ces éléments, les trois ou quatre plus importants. {36) 13 réponses dont 3 en première ligne ; une seule réponse estimait cette considération inopportune. (37) 12 réponses dont 4 en première ligne ; il y avait unanimité pour faire entrer cette considération en ligne de compte. — 136 — — un passé sans tache (38) ; — les chances de reclassement (39) ; — le fait d’occuper, lors de la comparution, un emploi qu’une incarcération ferait perdre (40). D’autres éléments de la persônnafité, quoique jugés. dignes d’in- térêt, sont estimés secondaires (41) ; certains font même l’objet. d’un refus de prise en considération (42). Les membres du parquet interviewés, dans les mêmes conditions, sur le comportement des magistrats du siège, rendent un témoignage qui corrobore ces déclarations (sauf en ce qui concerne les chances de reclassement) (43). Le témoignage des juges d’instruction (plus impliqués peut-être) diffère précisément sur ce dernier point ; il place en effet les chances de reclassement en tête, suivies du passé sans tache, du main- tien de l’emploi et enfin des charges familiales (44). Quant à œlui des avocats, il concorde en général sur les éléments pris en considération (45), mais ne considère pas que les chances de reclassement soient très déterminantes (46). b) En dehors de la personnalité du prévenu, les modalités de la procédure occupent un rang important dans les motivations des juges. Il paraissait notamment indispensable de connaître la place que le souci de « couvrir » la détention provisoire occupait dans l’esprit des magis- (38) 12 réponses dont 4 en première ligne ; une seule exception pour la pri'se cn considération de cet élément. Inversement, le « passé judiciaire chargé » est considéré comme déterminant. au 3° rang (après l’arme et la vio | À encel pour unec décision sévère. {39) C'’est là un élément qui est difficile à saisir objectivement et que 1'«:anèïc sur les dossiers n’éclaire guère (cf. infra). c0} 10 réponses, dont 4 en première ligne ; il v à une quasi-unantinité pour sa prise en considération. (41) Tel est le cas du grand âge (quasi unanimement retenu, mais cité seulement 3 fois comme déterminant), de la fragilité psychique (unanimement retenu, mais citée seulement 6 fois comme déterminant), du niveau socio-culturel (majorié pour le prendre en considération. mais citée seulement une fois comme déterminant). (42) Tel est le cas du sexe, rejeté par la majorité si l’on tient compte que plu- sieurs magistrats n’acceptent de le faire entrer en ligne de compte que s’il est lié à des charges de famille. Tel est le cas également de la nationalité, rejetée à une torte majorité. (43) Le jeune âge vient en tête, suiv! immédiatement du passé sans tache, des charges familiales et du souci du maintien de l’emploi ; étaient estimés également pris en considération le grand âge et le niveau socio-culturel, puis, avec plus de réserve, la fragilité psychique et le sexe ; la nationalité était exclue. (44) Le jeune âge, unanimement pris en ligne de compte, n’est cité qu’une fois comme déterminant : il en est d'ailleurs de même du niveau socio-culturel ; le sexe ct la nationalité ne sont pris que modérément en considération. (45) Ts font cependant une place plus importante au sexe, mais non à la natio- nalité. (46) Pour eux, le jeune âge tient la première place, suivi des charges familiales du maintien de l’emploi puis des chances de reclassement et du passé sans tache. — 137 — trats du siège. En conséquence, deux questions ont été posées la première mélait l’incidence de la détention provisoire à d’autres modalités de la procédure (47) la seconde posait le problème de façon plus directe et plus personnelle. La différence de perspective conditionne des opinions qui divergent — La détention provisoire, lorsqu’elle n’est pas envisagée à elle seule apparaît au premier rang des éléments qu’il est opportun de retenir (10 réponses), mais elle occupe aussi une large place (6 réponses), parmi ceux qui, au contraire, doivent être écartés (48). — En revanche, lorsque le problème est posé de façon directe et brutale, le rôle de la détention provisoire est reconnu déterminant par la quasi-unanimité des magistrats du siège, qui justifient d’ailleurs 10urs positions par des explications détaillées. “Des remarques assez voisines peuvent être faites en consultant le témmgnage des parquetiers (49) et celui des juges d’instruction (50). 1I est intéressant de constater que pour les avocats, qui se prononcent de façon quasi unanime, d’une part la détention préventive 10ue un rôle déterminant dans l’octroi du sursis partiel, d’autre part les 1ugœ, dans le dosage de la peine mixte, manifestent le souci de « couvrir la prévention ». Les juges de l’application des peines, à qui la première quesüon avait seule été posée, ont été unanimes à répondre par l’affir- maüvo c) Sur un autre plan la motivation du juge peut également être recherchée dans l'existence de certaines circonstances ayant accompagné la commission de l’infraction et ayant déterminé le juge à accorder un sursis partiel, à la place d’une peine ferme (si elles étaient favorables) ou à la place d’un sursis total (si elles étaient défavorables). Les circonstances peuvent concerner le modus operandi, la personnalité de la victime, le dommage causé ou l’inspiration du passage à l’acte Parmi ces dernières, il semble qu’il faille éliminer l’influence de l’alcool ou des stupéfiants, écartée par la majorité des juges du siège (51). Le fait que l’infraction soit restée au stade de la tentative est (47) Poursuite par la procédure de flagrant délit, contrôle judiciaire, existence d’une partie civile, réquisitions conformes du ministère public (48) Parmi ces derniers, une très forte majorité (18) écarte l’existence d’une partie civile et les réquisitions conformes du ministère public une majorité nette (16) écarte également le fait que l’on a employé la procédure de flagrant délit (49) Mèême division que celle des juges du siège sur l’opportunité de faire entrer la détention préventive en ligne de compte ; majorité nette pour constater son carac tère déterminant auprès des juges (50) Deux d’entre eux n’estiment pas opportun d’en tenir compte ; une majorité considère souhaitable de « couvrir » la détention prov1501re (51) Elle ne figure pas parmi les éléments estimés déterminants, et les parque- tiers pensent aussi au l faut la négliger ; les juges d’instruction inclinent dans lc même sens ; ce qui n’est pas l’avis des juges de l’application des peines ; mais le témoignage des avocats confirme ce peu d’influence. — 138 — considéré lui aussi par une large majorité comme n’ayant pas d’influ- ence (52). Le fait que la victime soit un représentant de l’autorité est également écarté, quoique moins nettement (53) et la participation de plusieurs personnes à l’infraction n’est retenue que par une légère majorité (54). De même certains magistrats ont fait appel de façon ponctuelle à l’ancienneté de l’infraction et à l’attitude provocatrice de la victime (cette attitude en revanche, est soulignée par la majorité des juges d’instruction, des juges de l’application des peines, ainsi que par les avocats). A l’opposé, une majorité massive fait entrer en considération dans un sens favorable la réparation du dommage (55), et à un moindre degré le fait que l’auteur se soit laissé entraîner (56). Dans un sens défavorable, une majorité également massive souligne, dans le modus operandi, l’usage d’une arme (57) et celui de la violence (58) (souvent lié d’ailleurs au premier). Une majorité presque aussi forte s’affirme influencée par le fait que la victime soit sans défense et pitoyable (59), enfin par les facilités dont le coupable a abusé pour commettre l’infrac- tion (60). (52) 15 réponses négatives, 4 «très pèu », 5 affirmatives (notamment dans le Sud-Ouest). Les parquetiers sont encore plus divisés (3,3,2) ; les juges d’instruction penchent pour la négative ; les juges de l’application des peines se divisent par moitié. Le témoignage des avocats est légèrement en faveur de la négative. (53) 11 réponses négatives, 5 «très peu», 9 affirmatives (chaque centre est représenté) ; même réaction chez les parquetiers {3,3,2) ; au contraire, les juges d’ins- truction estiment qu’il faut grandement en tenir compte, et les avocats pensent qu’on e fait. (54) 16 réponses affirmatives, 2 «très peu », 6 négatives (chaque centre est représenté). Il en est de même chez les parquetiers (5,2,1). Les juges d’instruction en tiendraient davantage compte (3,3,0) et les avocats estiment que c’est ce qui se passe (5,2,1). Les juges de l’application des peines se partagent par moitié. (55) 21 réponses affirmatives, 3 « très peu », aucune négative. Les parquetiers sont tout aussi nets, et les juges de l’application des peines unanimes. Le témoignage des avocats confirme le comportement des juges du siège. On notera que le dom- mage minime, mais non réparé, est apprécié de façon variable (10 affirmatives, 9 « très peu », 5 négatives) alors que les parquetiers sont plus favorables (7 affirma- tives, Î « très peu », Ô négative) et que les juges d’instruction le sont moins (3,2,1). La majorité des avocats croit en un rôle positif. (56) 14 réponses positives, 7 « très peu », 1 négative. (57) 24 réponses affirmatives, 1 « très peu », aucune négative. Les parquetiers sont unanimes en ce sens, ainsi que les juges d’instruction et les juges de l’appli- cation des peines. Le témoignage des avocats est également unanime. (58) 24 réponses toutes affirmatives. Il y a de même unanimité chez les parque- tiers, et les juges de l’application des peines. Les juges d’instruction vont en l’affirmative. grande majorité dans le même sens ; le témoignage des avocats est unanime pour (59) 23 réponses affirmatives, 2 « très peu », aucune négative. Une seule excep- tion à l’unanimité des parquetiers ; unanimité des juges d’instruction, des juges de l’application des peines et des avocats. 60) 21 réponses affirmatives, 3 « très peu », aucune négative. Au contraire, les parquetiers sont beaucoup plus nuancés (3,3,2), de même que les juges d’instruction (2,2,1) et les juges de l’application des peines (2,1,1). La majorité des avocats esti- me que les juges en tiennent compte. — 139 — Priés de-souligner les éléments les plus déterminants, les magis- trats du siège citent dans l’ordre la : violence (61), l’usage d’une arme (62) et la victime pitoyable (63). Le classement est le même chez les juges de l’application des peines et le témoignage des avocats est concordant Les deux premiers éléments sont les seuls considérés comme prédo- minants par les parquetiers (64). d) Dans lu recherche d'autres motivations, on s'est enquis de savoir si le désir de faire 8xpérzmenter le regzme pénitentiaire pour donner plus d’efficacité au sursis, pouvait ou non jouer un rôle. Seuls trois magistrats du siège le contestent expressément, mais plus du tiers estirient que cela exerce sur cux un influence, et pensent volontiers qu il en est de même pour leurs collègues (65). Il est à noter à ce su]eL que les juges de l’applncatwn des peines doutent qu’un prévenu qui n’a 1amms été incarcéré puisse mesurer pleinement le régime auquel il serait soumis si le sursis dont il vient de bénéficier était ultérieurement révoqué. Les juges d’instruction partagent en majorité ce sentiment 4) Les effets réels ou supposés du sursis partiel ont fait l’objet de plusieurs questions À la première, qui portait sur le degré d’efficacité du sursis partiel sur la prévention de la récidive, selon qu’il était appliqué au sursis simple ou au sursis probatoire, la grande majorité des magistrats répond que l'application au sursis simple est plus efficace. Les parquetiers esti- ment unanimement que l’efficacité iest certaine dans les deux cas ; les juges d’instruction, ainsi que la majorité des juges de l’application des peines, partagent cette opinion avec plus de réserve. Il est à noter que les avocats sont très divisés sur ce point, et même sur la place que les juges accordent à une telle considération Une autre question portait sur l’éventualité d’un choc salutaire que la peine mixte pouvait produire. Les magistrats du siège admettent presque unanimement cet heureux effet, et leur avis est assez large- ment partagé par les juges d’instruction et le ministère public, À deux exceptions près, il en est de même chez les avocats (bien placés, en principe, pour en juger), qui estiment toutefois qu’il s’agit là surtout de l’opinion des magistrats (61) Citce 7 fois en première ligne, 8 en seconde, et 2 en troisième (62) Citée 7 fois en première ligne, 6 en seconde, et 2 en troisième (65) Citée 2 fois en p1em1erc ligne, 1 en seconde, et 9 en troisième. (64) La victime pitoyable n’est pas citée, mais un parquetier mentionne la vic- time agent de l'autorité. (65) En général, ils fixent entre 8 à 15 jours et l à 2 mois la durée d’incarcé ration nécessaire ; aucun n’estime que la durée de la partie ferme doive normalement excéder celle du sursis accordé — 140 — Le sursis partiel a-t-il pour effet de rendre une probation plus facile ? Aucun des juges de l’application des peines interrogés (ils s’é- taient tous occupés de condamn& à des peines mixtes) (66) ne voit d’a- vantage très marqué cet égard, dans l’institution. Ils ne sont pas d’avis, à tout le moins, d’en faire le préalable normal de la probation. Parmi les autres effets du sursis partiel, sont signalés positivement une meilleure intimidation que par le sursis total (d’où meilleure pré- vention de la récidive) et une plus. grande possibilité de réinsertion sociale ; négativement les risques inhérents à la promiscuité de la prison, en dépit des limites apportées à la durée de l’incarcération c) Comparaison des résultats de l’enquête sur dossiers et de ceux de l’enquête sur entreüens. Le plan adopté pour la mise en œuvre de la recherche supposait un examen comparatif des résultats respectivement obtenus par l’une et l’autre branche de l’enquête. D’ores et dé]à cette oomparmson peut être œqmssée 1) Principaux points de concordance. Les points de concordance concernent tout d’abord les délmquants susceptibles d’être bénéficiaires du sursis partiel. Le fait qu’ils soient généralement jeunes (67) ne résulte pas seulement d’une meilleure aptitude à remplir les conditions du sursis (68), mais correspond à une intention déibérée des magistrats qui rangent même cet élément parmi les facteurs déterminants (69). Cependant le passe md1czazre n’est pas sans importance. Si les 3/4 des bénéficiaires n’ont aucune condamnation à leur casier judiciaire (70), ce n’est pas l’effet du hasard, car les magistrats voient dans un passé sans tache un élément non seulement favorable mais déter- minant (71) et au contraire dans un casier judiciaire chargé un élé- ment éminemment défavorable (72) - (66) Il faut noter en outre que, à une exception près, les juges de l’application des peines assuraient également un service de juge du siège, ce qui renforce le poids des opinions qu'ils expriment (67) Voir supra p 123 (6S) À l'époque où les affaires ont été jugées, de telles conditions existaient encore pour le sursis avec mise à l’épreuve. {69) Foutes les réponses sont affirmatives quant à la prise en considération (bien qu'en note 2 « très peu »), et le jeune âge vient au second rang (12 réponses dont 6 en première ligne) parmi les éléments considérés comme déterminants (70) Supra p. 125 (71) Supra (72 Supra — 14t — L'importance de la durée d’une détention provisoire éventuelle est nettement confirmée La proportion des personnes poursuivies par la voie de l’information est certes considérable parmi les bénéficiaires (73), mais c’est la voie qu’il est nécessaire de suivre en matière correc- tionnelle (à l’exclusion de la citation directe) si l’on souhaite placer l’intéressé en détention. En présence de cette situation, les magistrats affirment en tenir un très grand compte pour le prononcé de la peine (74) et pour l’octroi du sursis partiel (75) (les analyses en composantes principales confirment d’ailleurs ces deux points) (76). Le souci de « couvrir la prévention » apparaît, dans une mesure importante (plus faible cependant qu’on ne le pensait au départ), à chacune des étapes de la recherche. Dans un nombre appréciable de cas, le caractère favorable ou défavorable attribué lors du dépouillement des dossiers aux circons- tances de l’infraction (77) se trouve confirmé par les entretiens avec des praticiens, plus encore que par les analyses en composantes prin- cipales. Ainsi en est-il, parmi les circonstances favorables, de la réparation du dommage causé ou de l'attitude provocatrice de la victime, et parmi les circonstances défavorables, de la violence (notam- ment l’emploi d'une arme), des facilités dont l’auteur disposait pour agir, du caractère pitoyable de la victime. Au contraire, certaines des circonstances supposées influer n’ont été confirmées ni par les analyses en composantes principales, ni par les entretiens. 2) Principaux points de divergence. a) C’est notamment le cas pour certaines circonstances de l’infrac- tion. Ainsi l’importance du dommage causé, qui est une variable explicative à la fois de la peine prononcée, de la durée du sursis partiel et de la peine ferme à exécuter, n'apparaît pas de façon aussi nette dags les entretiens avec les praticiens (78). (73) Supra p. 121 ; 2/3 alors que ce mode de saisine ne représente d’habitude que 15 % des poursuites correctionnelles. (74) Supra (75) Supra {76) Supra (7) Voir supra pp. 126-130. (78) Les conséquences de l’infraction ne sont jamais signalées comme détermi- nantes pour l’octroi du sursis partiel par les magistrats du siège ; d’autre part, le caractère minime du dommage subi n’est pris en considération, pour l’octroi du sursis simple, que par une faible proportion de ceux-ci (10 « oui », 9 « très peu », 5 « non »). — 142 — Dans des conditions moins frappantes, on constate que la simple tentative (79) ou la correctionnalisation de l’affaire (80) n’ont pas sur l’esprit des juges l’influence que l’on aurait supposée b) Il y a divergence aussi sur certaines caracténsthuœ du délmquant. Alors que le dépoufllement des dossiers paraissait faire ressortir un pourcentage élevé d’étrangers, et que la nationalité jouait un certain rôle dans l’analyse en composantes principales de la variable Sp (durée du sursis az cordé), la grande majorité des juges du siège estime que la qualité d’étranger ne doit jouer aucun rôle dans l’octroi du sursis (81) De même, il est curieux de remarquer que les charges de famille; qui n'étaient pas révélatrices au plan statistique, sont au contraire prises en considération par l’unanimité (à une exception près) des magistrats du siège, et figurait en tête des éléments que ceux-ci reconnaissent pré- pondérants pour l’octroi du sursis partiel (82), notamment à Lille et dans le Sud-Ouest On est enfin frappé de l’importance considérable que les magis- trats donnent aux « chances de neclassement » (83), alors que le tri à plat a permis de constater (84) qu’une importante fraction (25 à 30 %) des individus connus pour leur mauvaise conduite, leur fâcheuse moralité, leur triste réputation et leur alcoolisme, bénéficiait cependant d’un sursis partiel ; de même la récidive, la réitération ou tout au moins le concours réel apparaissent dans près des deux tiers des dossiers (85) Mais peut-être n’y a-t-il pas là véritablement contradiction, car les « chances de reclassement » sont une notion fort complexe qu’il était difficile de saisir intégralement dans une fiche de dépoufllemmt de dossiers. 3) Points qui restent obscurs. Ni l’enquefie sur dossiers, ni son exploitation, ni l’cnquête sur interviews n’ont permis de savoir si le sexe du prévenu exerce une —— (79) Supra p. 135 et note 52. (80) Chez les magistrats du siège, 13 réponses négatives, 2 « très peu» et 8 affirmatives ; le caractère déterminant n’existe que chez 2 magistrats, les membres u parquet ‘sont d’un avis contraire mais aucun ne range cet élément défavorable Parmi ceux qui sont déterminants. (81) 18 réponses négatives. Certains magistrats lyonnms et parxsxens parament Plutôt d’avis contraire (2 2 affirmatives, 3 « très peu » (82) 13 réponses, dont 3 en première ligne et 6 en secondc (83) Prises en considération par l’unanimité des juges du siège, ct 1ndxquccç Mmême comme déterminantes (11 réponses, quoique aucune en première ligne, voir Supra p. 133 et note 39. (84) Supra p. 125 (85) Supra p. 127 — 143 — influence sensible sur l’octroi du sursis partiel. Le rôle de la caté- gorie socio-professionnelle n’est guère moins confus (86). Il en est de même, on l’a vu, pour la nature de l’infraction (87). Enfin, les magistrats ont affirmé que le souci de ne pas pro- noncer une peine qui ferait perdre au prévenu l’emploi que celui-ci occupe, joue un rôle important dans leur décision (88) et font même figurer cet élément en 5ème place (89) parmi les éléments déterminants. L'enquête sur dossier n’a guère pu corroborer ce point : il était bien recommandé de noter si le prévenu occupait, au moment de sa comparution, un emploi que son incarcération risquait de lui faire perdre, mais, pour diverses raisons, il a été difficile de découvrir ce renseignement. CONCLUSION Il n’est guère possible, en l’état actuel d’une recherche qui n’est pas encore achevée, de dégager des conclusions définitives ; néan- moins, certaines tendances apparaissent déjà assez nettes pour mériter d’être notées. L’institution du sursis partiel a reçu, de la part du corps judiciaire, un accueil extrêmement favorable. Il était permis de s’en étonner, du fait qu’elle aboutit à multiplier les courtes peines d’emprisonnement, qui sont à l’heure actuelle très décriées comme plus néfastes qu’utiles. Une hypothèse fréquemment avancée était que la peine mixte échappait à cette critique parce qu’elle correspondait à un emprison- nement d’ores et déjà effectué, c’est-à-dire à la détention provisoire ayant précédé le jugement. Cette hypothèse n’est confirmée que par- tiellement par la recherche. La correspondance approximative (à quinze jours près) n’existe que dans 40 % des cas d’application, et cette pro- portion paraît avoir tendance à diminuer. (86) Sur la répartition dans les catégories socio-professionnelles, voir supra p. 124. Les analyses en composantes principales (supra pp. 130-131) fournissent des résultats complexes et parfois étonnants. Les juges du siège n’estiment prendre en considération le niveau socio-culturel qu’à une faible majorité (13 oui, 5 « très peu », 6 non) notamment à Lyon et Paris. et un seul d’entre eux lui attribue un rôle déterminant. (87) Supra pp. 121, 132 ct la note 35. (88) Unanimité (à une exception près) pour estimer qu'il y a lieu de tent compte de cet élément dans l’octroi du sursis ; cf. supra p. 133 et note 40. (89) 10 réponses en ce sens, dont 4 en première ligne. Le sursis partiel intervient dans de nombreux cas où il n’y a eu aucune détention provisoire sensible. Il en est particulièrement ainsi dans les procédures de flagrant délit ; il s’apparente alors à la pratique de certains tribunaux qui pensent qu’une très courte période d’incarcé- ration peut avoir un effet de choc salutaire (le « sharp short shock » britannique) (90). Aussi s’explique-t-on qu’il soit utilisé davantage en fonction de la personnalité du délinquant que de la nature de l’infrac- tion. Parfois la peine mixte remplit un autre rôle, celui de faire suffi- samment connaître le régime qui serait appliqué au prévenu en cas de révocation de son sursis ; c’est le cas lorsque la partie ferme est assez longue ; la durée de l’incarcération estimée nécessaire est en effet alors un peu plus forte. D’'autre part, il apparaît nettement qu’un partie du succès acuel du sursis partiel est due au fait que le sursis simple total a perdu une bonne partie de l’effet intimidant attribué à son caractère comminatoire (91). Enfin, il n’est pas rare de voir infliger des peines fortes accom- pagnées d’un sursis partiel important. Il semble que, de façon préto- rienne, le juge souhaite alors observer, et observe en fait une certaine proportion entre la durée de la peine totale et celle du sursis partiel, en sorte que celle-là est généralement plus longue que celle de l’incarcé- ration subie ou à subir dès lors que le tribunal ne s’est pas contenté de « couvrir la prévention ». Il faut enfin observer que l’enquête apporte de très utiles éléments à l’étude du sentencing. Grâce à elle, sont mis en lumière de très nombreux éléments ou circonstances (ainsi que le poids approximatif des uns et des autres) qui entrent en ligne de compte dans l’octroi du sursis, simple ou probatoire, total ou partiel et dans le dosage de la peine. La présente enquête, une fois menée à son terme, devra pourtant être complétée par une étude symétrique, déjà esquissée, de l’appli- cation du sursis partiel par les cours d’assises ; d’ores et déjà, cette application paraît obéir à des considérations différentes, du fait qu’il s’agit d’infractions particulièrement importantes. Un autre complément devra porter sur l’attitude des cours d’appel afin de rechercher si elles se montrent plus libérales ou plus exigeantes que les tribunaux correc- tionnels. (90) Une large majorité des juges du siège et des juges d’instruction se prononce en ce sens, et fixe la durée maximale à un mois environ (la moitié ne dépasserait pas 15 jours). Les membres du ministère public se montrent divisés sur ce point, mais l’admettent lorsqu’il prend la forme du sursis partiel. (91) La majorité des juges du siège a opiné en ce sens. La quasi-totalité des parquetiers pense que la partie ferme de la peine constitue, aux yeux du tribunal, UN minimum de répression assurée. — 145 — - À titre de pierres d’attente pour des recherches futures, des élé- ments ont été recueillis sur les réformes apportées au sursis avec mise à l’épreuve (92), ainsi qu’au sursis simple et à la restriction de sa révo- cation (93) ; sur l’institution des substituts aux courtes peines d’empri- sonnement (94) sur la révocation partielle ou totale du sursis pro- batoire (95), etc. Le fonctionnement de la justice pénale offre encore aux chercheurs un domaine presque illimité. (92) Approbaiion unanime de la part des magistrats du siège, des parquetiers, de- juges d’instruction et des juges de l’application des peincs. (93) La quasi-totalité des magistrats du siège ct l'unanimité des parquetiers approuvent ces réformes ; timent. : v_les avocats partagent tout à fait. bien entendu, ce sen- (94) La réforme Aest généralement approuvée par les juges du siège mais peu utilisée ; les juges d’instruction et les juges de l’application des peines sont beaucoun plus réservés ; les avocats se montrent divisés. (95) Les juges de l’application des peines demandent rarement la révocation par- tielle, varfois la révocation totale ; dans un cas comme dans l’autre. le tribunal suit leurs pronositions. — 146 — RESTlTUTION TO VICTIMS OF CRIMINAL OFFENCES IN SLOVENIA KATJA VODOPIVEC Institute of Criminology Ljubljana - S.F.R. Yugoslavia. In ancient times people resolved or sharpened the conflicts between the offender and the victim at the beginning among them- selves (individual revenge, composition) and later on with the help of the state authority or its institutions (composition, punishment). Nowadays, it is difficult to evaluate which was the main reason for the authority’s intervention in resolving private conflicts. Let us quote two examples. Vasiljevic is of the opinion that in the old accusatory procedure the pursuit could have been abandoned even if necessary, due to the lack of knowledge, incapacity or intimidation of the victim on one hand and on the other, it was possible to charge the person for the reason of revenge or other negative motives. He refers to the comedies of Aristophanes illustrating the deficiencies of private accu- satory procedure (Vasiljevic, p. 18). In the opinion of Mueller and Cooper « the rights of selfhelp were regulated and formalized by the state so as to limit their exercise and theoretically, to prevent divisive feuds and the exploitation of the weak by sirong… By the Middle Ages, the individual harmed in any way must always have recourse to law rather than taking the matter into his own hands. In terms of victim compensation, the price of state assistance in bringing the wrongdoer to justice soon becomes disproportionately high» (p. 86). After this time the interests of the victim have been replaced by the interests of the authority. From the sociological point of view the interests of the victim have been the most neglected in the inquisitory procedure. Nevertheless the theoreticians and judicial practice returned to the problems of the victims all over again, especially since the 16th century as e.g. Thomas More, Jeremy Bentham, A.B. Marsagny, Herbert Spencer, Raffaele Garofalo, Enrico Ferri and others. In some German countries the judicial practice conceded the so-called adhesive procedure as early as the l6th and 17th centuries and the French « action civile » has been enacted by the end: of 17th century (Kobe — 147 — 1977, IT). Also the international penological congresses in the last deca- des of the 19th century dealt with the problem of restitution as a mean for the resocialisation of offenders. Between Worlid War I and II the criminological and penolo- gical literature seems to neglect the victim of offences, but the adhesive procedure in the countries where it was enacted remained and developed. After the second world war the victim has won one of the first place in criminological, and finally in victimological literature. The reasons for such an approach are mainly threefold. 1. — The aspect of the victim. The attention given to the relationship of the personality of the offender to his surroundings included in the framework of crimino- genesis also the victim, at the beginning as a person who might have contributed to the criminal offence (Hans von Hentig, Benjamin Mendelsohn). Although Margery Fry recommended the restitution to the victim in 1951 with the correctional view, this belief has later on been replaced by the philanthropic attitude toward the victim who remained damaged especially in cases in which the offender was not identified or lived in bad material conditions. In 1957 she finally began to plead in favor of a special state fund for compensation of victims of criminal offences in Great Britain (Jacob, p. 152). After 1960, in many countries, public funds have been established for this purpose (Grebing, p. 476-482). 2. — The aspect of the offender — the correctional view. At the beginning at least this aspect has been in the forefront of Margery Fry’s interest. The English White Paper : The penal practice in the changing society of 1959 emphasized especially the corrective role of the restitution for the offender (Linden, p. 39). In 1958, Albert Eglash was of the opinion that « restitution as a cons- tructive activity may contribute to an offender’s self-esteem » (Jacob, p. 156). This aspect has been of special interest for the well-known victimologist Stephan Schafer from 1960 onwards. Finally, it is possible to include under this heading also those writers who draw attention to the detrimental effects of imprisonment and consider the restitu- tion as an effective alternative to it. 3. — Sociological view. The critique of contemporary sociologists is oriented toward the power of state, institution and the bureaucracy — in our case especially toward the state control — over the relationship among people. Nils Christie as adherent of the conflict theory (Dahrendorf) considers — 148 — conflicts as acceleration of the society>s development. The conflicts urge the participation of people, their activity and are eluciding the normative and value problems. « Segmentation according to space and according to caste attri- butes. leads into a depersonalisation of social life… The proceeding is converted from something between the concrete parties into a conflict between one of the parties and the state… Conflicts become the property of lawyers…. We have reduced the victim to a nomentity and the offender to a thing… The offender has lost the opportunity to explain himself to a person whose evaluation of him might have mattered » (pp. 6, 3-5, 9). Marvin E. Wolfgang reproaches to the criminal justice system that it is not « accountable to the public at large, to the victim and to the offender ». It is a closed system. As a compromise of all these and other views many recommend- ations regarding the indemnification process have been adopted on the international and national bases. The most important recommend- ations were formulated especially at the 11th International Congress of Penal Law at Budapest in 1974 (section IIT) and at the International Study Institute on Victimology at Bellagio (Italy) in 1975. On the national basis the most advanced proposals have been made by the Law Reform Commission of Canada (working papers 5 and 6 — Restitution and Compensation as well as working paper 7 — Diver- sion formulated in the years 1974 and 1975). Besides, as already men- tioned, public funds for the compensation of victims (especially of offences against the life and body) have been founded in different countries since 1960. There exist, however, well founded reasons against an exagge- rated use of restitution (e.g. Galaway p. 66. Goldstein p. 203, Linden p. 41, Mueller-Cooper pp. 94, 98, Williams p. 158). The legal provisions in Yugoslavia. The new constitution of F.P.R. of Yugoslavia from 1974 has decentralized legislative competence in the field of penal law among the Federation and 8 republics and autonomous regions. Consequently, on the Ist of July 1977 the new nine penal codes were enacted besides the new federal Code of criminal procedure. Regarding the indemni- fication possibilities there are no differences in principle between the previous and the new legislation. In order to give an accurate presen- tation oË legal provisions we shall describe in this section the present possibilities although the research has been carried out for the year 1973, Where necessary the differences will be outlined. Regarding the indemnification possibilities there are no differences among the penal codes of different republics and autonomous regions. Therefore we — 149 — shall take into account only the federal penal code (F.P.C.), the penal code of the Republic of Slovenia (P.C.S.) and the federal Code qf Penal Procedure (P.P.). For this purpose we shall apply the terminology proposed by S. Schafer and others and shall designate as restitution the indemni- fication given to the victim by the offender and as compensation the ipdemnification awarded to the victim by the public funds. 'We shall structure the presentation of our legal provisions by problems and recommendations adopted especially at international meetings and by the Canadian Law Reform Comrnission. For the more detailled information we refer to the national report prepared for the 11th International Congress of Penal Law at Budapest by Prof. Selih and Kobe. 1) The restitution as a mean for the diversion of the judicial procedure. The most advanced program in this respect has been developed by the Canadian Law Reform Commission (E.p.7.). In the U.S.A, an experimental Columbus Night Prosecutor Program in 1971 has been developed (Palmer). Such programs correspond to the principle of opportunity of public prosecutions as adopted in these countries in distinction to those legislations of the European conti- nental system which adopted the principle of legality for the public prosecution. It means that the public prosecutor is bound to prosecute all offenders who committed offences which are subject of public prosecution. The basic idea behind such provisions is equality before the law. In accordance with the penal legislation adopted in many states of the European continent the Yugoslav P.P. adopted the principle of legality. In the Yugoslav literature this principle until now has not been disputed and it is very unlikely that it will be changed ip the near future for very good reasons. —In spite of that one of the most natural ways for the diversio‘n of judicial process also in our country is the spontaneous reconcili- ation between the parties opposed with the consequence of non- reporting the criminal offence to the state control agencies or of non- charging the offender in cases in wich criminal offences are prosecuted by private prosecutor. For a part «of minor criminal offenses are prosecuted under the Yugoslav system by the injured party who initiates proceedings, files the charging papers and represents the pro- secution in court. The public prosecutor may not intervene in the case. Typical offenses prosecuted in this way are minor assaults and batteries, malicious mischief, defamation, etc. » (Damaska, p. 33). — 150 — In 1959 in Yugoslavia the lay reconciliation boards have been established at the local level. However, these boards did not enjoy enough credit among the people and were successful in reconciliation only exceptionally, especially at rural sides of the country (Pecar 1968) In this respect the new P.P. of 1977 brought the supplementary pro- vision that the judge should be bound to advise the subject matter (in above quoted offences) to the reconciliation board to see whether the reconciliation was possible (art. 445 P.P.). The law of the republic will determine when this provision shall be applied in the Republic Besides, it is possible that the private charge shall be dismissed by the court judge for certain reasons (art. 438 P.P.). The purpose of this provision is to eliminate the damage and inconveniences which might affect the offender due to an unfounded charge. In practice the private charges are regularly not controlled in this sense (Vasiljevic- Grubac and others, pp. 179, 180) Finally, the decriminalisation can be considered as a mean of diversion of the penal process and such considerations have been taken into account at preparatory work for the penal code of Slovenia as well as in some other Republics. 2) The restitution as the subject of penal procedure. Besides the « action civile» in French Law the so-called adhesive procedure be- came the overall accepted institute on the European continent since 1850, as enacted by the Austrian Code of Penal Procedure (Kobe 1977 ID). At this time, it was set up on the territory of Slovenia. The system of adhesive procedure is a system for restitution of damage civil in character but awarded in criminal proceedings. Its main features have been described quite often (see e.g. Schafer pp. 112, 113 ; Selih-Kobe pp. 305-313). Basically the Yugoslav solution does not d1fier much from the others. After World War II the Yougoslav practice enlarged the claims for indemnification in two directions. «The first enlargement is connected with persons entitled to make claim for indemnification Still more important is perhaps the enlargement for immaterial damage decided by the court, especially in the civil procedure » (Selih-Kobe, p. 311). One of the characteristics of the adhesive procedure is that it is simple and cheap for the victim, except for expenses for the legal adviser. In this connection Selih and Kobe are advocating the counsel on request of the victim, if by reason of his financial position he is unable to bear the costs for the counsel There are different possibilities to insure the claim for indem- nification by the confiscation of property benefits acquired by the — 151 — commission of a criminal offence (art. 84-87 F.P.C.), by the provi- sional remedies securing the claim (art. 112-114 P.P.) as well as by the provision for returning the illegally required things (art. 109 P.P.). Connected with these provisions the court is bound to collect evi- dence and inquire into circumstances necessary for the adjudication of the claim even before the claim has been made (art. 107 P.P.). Finally, the victim may himself assume the prosecution if public prosecutor finds that no grounds exist for initiating the proceedings for an offence subject to public prosecution (art. 60 P.P.). With the amendment of 1977 a new provision has been added according to which the public prosecutor and the court are bound to inform the injured party what he can do to realize this right (art. 60/5 P.P.). In the cases of more serious offences he has the right on the counsel of the poor (art. 65/2 P.P.), if he assumes the prosecution instead of the public prosecutor. Claims for indemnification can be decided by the court only in the case the offender is found guilty, but the court is not bound to take the decision provided such a decision would considerably delay the criminal proceedings and may direct the injured party to assert his claim by way of civil action (art. 103, 108 P.P.). Therefore, the injured party has not the right of appeal against the decision of the court in this matter (art. 360/4 P.P.). Besides, the victim may choose from the beginning to assert the claim for indem- nification before the criminal court or by way of civil action before the civil court. The provisions of art. 103, 108 and 360/4 P.P. are exposed to critique of some Yugoslav writers (e.g. Gvozdic), there is still the problem of the victim’s time spent at the court and in connection with that of the expenses (Acimovic) and finally, the suggestion can be found that instead of the injured party the public prosecutor should have the right to claim the indemnification or the court should have the right to take the decision in this respect on his own discretion (Caric). Against such a standpoint are the opinions of Vasiljevic and Kobe (1977 I). The latter with the argument that the injured party is and should remain an autonomous person and take the decision in this respect by himself. Instead of such paternalism Selih and Kobe propose the right on the counsel in the case the victim is not in the position to bear the expenses by himself. Finally, it is ques- tionable whether other physical (e.g. of parents, employers, committers, insurance organisations) could be made responsible in the adhesive procedure provided that the claim for indemnification may be decided by the court only if the offender is found guilty (Kobe 1977 II). — 152 — 3) The suspended sentence may be given by the court under the condition also that the offender will restitute the damage to the victim (art. 52, 56 F.P.C.). Such a provision was enacted in Yugoslavia as early as in 1929. 4) The restitution as a mitigating circumstance for the strictness of punishment can be taken into account under the general princrples in fixing punishment (art. 41 F.P.C.) and by the reduction of punish- ment (art. 42 F.P.C.). The two researches finished until now in Slovenia have demonstrated that this circumstance is taken into account by the judicial practice (Bayer, Bavcon and others). The new legislature of 1977 has enlarged the possibility to apply the judicial admonition (art. 59 F.P.C.) and it is possible that the restitution will have an influence on a larger application of this institute. 5S) The system of insurance and the compensation of damage by public funds. In Yugoslavia there exists a large system of insurance : health insurance, the insurance for disablement, the pension insurance and the obligatory insurance in traffic. Therefore, Selih and Kobe are of the opinion that it would be suitable to create only a limited public fund as a supplement for the compensation of damage not covered by the insurance systems. As far as I understand only the damage caused by the offences against life and limb would be additionally compensated with such fund, but the victims who contributed to the damage would be excluded (pp. 302-303). Caric is in favour of a public fund in case the offender remains unknown or lives in bad material conditions (p. 87). For the time beeing there is no such fund in Yougoslavia. 6) The possibilities for the indemnification our system does not apply. In our legal system the provision of probation is not foreseen for adult offenders. Therefore it is not possible to connect the possi- bility of this provision with the restitution of damage. Besides, our penal system does not know provisions as restitution of damage in lieu of the sanction (abolished in 1951) ; cession of a part of the fine to injured party for the restitution ; the offender cannot restitute the damage by his work for the victim but only with his earnings ; the wages of prisoners are too low for the indemnification of victims and the legal provisions do not foresee the restitution as a circums- tance for the approval of parole. Às far as we were able to confront the Yugoslav literature we have found no standpoints of Yugoslav authors regarding the proposals quoted. But it does not mean that some of them would not be worth being examined in the future. — 153 — Findines of the research (1l) The aim of the research. Public opinion often asserts that the legis- lative provisions are not applied in judicial practice. Therefore, the main hypothesis was that our judicial practice does not apply the available legal provisions for the restitution of damage to the victim So far such a hypothesis has not been verified This starting point means also the limitation of the research The findings will not give answers to the questions whether the restitution is a mean for the diversion of penal procedure (since we are observing only finally binding judgements) ; and whether the victims have got the real satisfaction since at this occasion, we were notable to follow neither the execution proœdure nor the compen- sation procedure by the insurance funds. The sample. The research has been carried out by nine students of postdiploma course and two undergraduate students under my guidance and with the help of Peter Kobe, professor of penal law. The data were collected from the court files for finally binding jud- gements for 1973 in the Republic of Slovenia (nationally Homogenous populaüon of 1 800 000 inhabitants). The research was carried out in 1976 in order to allow enough time between the finally binding judgements of criminal courts and the civil action to be resolved by the civil court. In 1973 11387 offenders were condemned. In the case of concurrence of criminal offences each offender is taken into account only once for the offence for which the most severe punishment has been adjudicated. The offenders examined for this research have been chosen by the method of elimination. We have excluded 1) AIl the offenders who were condemned on the basis of private charges. These are specific cases since the private prosecutor decides to charge the offender mainly if the damage was not restituted. Besides, the role of the private prosecutor in the penal procedure is different from the role of the injured party in the adhesive procedure. Finally, private charges mainly refer to minor assaults and batteries, (1) The research under the title : Restitution to Victims of Criminal Offences in Slovenia has been financed by the research fund Boris Kidric and carried ‘out at the Institute of Criminology at the Faculty of Law, University of Ljubljana. — 154 — malicious mischief. defamation etc. ie. offences for which the damage can hardiy be established 84 % 2) All the offenders who were condemned for traffic offences. In these cases a great amount of damage is cove- red by insurances. Nearly all of the offences were com- mitted with negligence and the legal provisions for these offences have been restricted with the new legislation. The research of these offences would demand a special exa- mination 308 % 3) All the offenders condemned for the offences for which the damage could hardly be established, e.g. offences against People and the State, against Civil Liberties and Rights, against the Dignity of the Person and Morality, against the Marriage and Family, against the Administra- tion of Justice, against Public Order and Legal Transac- tions, the offences of Unauthorised Use of a Motor-Vehicle (Joy riding) and similar e »% e .%.."................... 17,3 % 4) AIl the offenders condemned for the offences for which the total sum of frequencies did not reach the frequency of 50 with the exception of two cases which seemed to be of speçial interest for the research 4,2 % 5) Some offences for various reasons, e.g. murder (due to the fact that this offence is an exception and the damage caused exceeds the material possibilities of the offender) ; participation in a brawl (it is difficult to distinguish between the participation of the offender and the victim) ; illicite commerce and tax evasion (the dark number is very high) ; petty theft, embezzlement or fraud (due to low damage) and receiving stolen property (the damage is mainly established with the perpetrator of the main e e à ® % ® © © © 6 % à © © % © » ® » 6 6 © ® © » ® % e % % à © % * » © © à » e « @ offence) 90 % 6) Offences taken info consideration e 6 e 6 ® #% 6 € % %* % 6 © © » 9 30,3 % Total e …. ... 100,0 % We want to draw the attention that at least 26 % (items 1 and 3) of offences have been excluded due to the fact that it is almost im- possible to replace the damage with the monetary indemnification in these cases. — 155 — The outline of the offences considered for the research is the following one : Offences against life and limb : — grievous bodily injury — art. 141 P.C, (art. 52/53 P.C.S.) . e . %* 6 %* + e ® %* © % © « e 6 6 © % e à © © e % © % < à © » © © & © ® ® # @ 19 % — light bodily injury inflicted by weapons, dangerous tools or other instruments — art. 142/2 P.C. (art. 54/2 P.C.S.) . * e « © % © « 6 s * © + © 6 6 © % à % e 6 @ © à %* % % © 6 ® e # # 22,2 % —— 0 2 Total 30,1 % Offences against property : — theft — art. 249 P.C. (art. 165 P.C.S.) 32,6 % ———emeamer aggravated theft — art. 250 P.C. (art. 166 P.C.S.) .. 178 % — e e © 6 * © — e © e fraud — art. 258 P.Ç. (art, 171 P.C.S.) 49 % — embezzlement in office — art. 322 P.C. (art. 141 . ». © 6 » . 66 806.804 t + © € 8 6 % à ® © à © e % © à 6 % 6 6 6 6 % © 6 % e © # P.C.S.) 5,2 % Total e e e »% 00e 605 % Other (miscellaneous) offences : — business mismanagement — art. 213 P.C. (art. 127 . e « . % à % « 6 © 6 © % ® % © % © 6 — % 5 e © D ® & @ © à %* © 6 % % à à » ® — P.C.S.) 13 % — pillage (2) — art. 255 P.C. (art. 142 P.CS)) 08 % — malicious mischief — art. 257 P.C. (art. 177 P.C.S.). . 5.4 % 1,9 % ,— forest theft — art. 246 a P.C. (art. 158 P.C.S.) Sum-total 100,0 % 3451 if the value gained exceeds a certain amount. (2) Pillages are offences under the articles 250 P.C., 322 P.C. and similar Comment the articles following the description of the offence are the articles of the Penal Code valid in F.P.R. of Yugoslavia until 1977 July the tion these articles will be used. In the brackets are the articles of the new lst. Since the research has been carried out for the year 1973 in continua- Penal Code for the Republic of Slovenia. Between the two kinds of incrimi- nations there is only partial concordance. — 156 — The 25 % random sample has been taken out of these offenders except for the business mismanagement, pillage and forest theft for which a 100 % sample has been taken due to small frequencies. In computation of totals this fact has been taken into account. Thus the results are representative for the selected offenders. Besides, we have to call the attention to the fact that the numbers of mwred persons are greater than the numbers of offenders. Results. 1) Offenders and victims. Nearly one half of victims of all the property offences are legal persons (44 % — e.g. economic organi- sations, public institutions and similar). With criminal offences of business mismanagement, embezzlement in office, pillage and also malicious mischief the legal persons are the damaged parties in almost all the cases. In these conflict situations one cannot expect a great correctional advantage of direct interpersonal contacts bet ween the offenders and the legal representatives of the organisations It might be that in some cases the proceedings before the self- management bodies could replace the penal procedure what is already the practice in some situations. The perpetrators of criminal offences against life and limb malicious mischief, thefts and partially also of frauds are (in com- parison to the vocational structure of population) mainly workers and unemployed persons in distinction to the perpetrators of em- bezzlement in office, pillage and business mismanagement who mainly belong to the middie class of population. With personalized offences (against life and limb) the perpetrators and the victims belong mainly to the same social class. In these cases the victims are quite often also supported persons. The individuals as victims of criminal offences against property mainly belong to the middle social class and to supported persons (older people, disabled and similar) Bodily injuries are in 1/5 of the cases inflicted to close rela- tives or close acquaintances 2) The restitution of damage until the main trial. The damage involved was :mainly not high. With one half of offences against property the damage caused did not exceed 1 000 dinars (according to prices in 1973) and only in 15 % the damage exceeded 5 000 dinars The offenders of different offences restituted the damage before the main trial in quite different proportions, e.g. for thefts in 66 % to 2.5 % for grievous bodily injuries. But it would be a mistake to presume that with thefts we have to deal with spontaneous resti- tutions. Às already mentioned with thefts the damage has been restituted in 2/3 of cases, but out of these in 85 % in things with aggravated thefts the damage has been restituted in one half of cases — 157 — but out of these in 86 % in things. It means the judicial procedure involves chiefly those perpetrators of thefts and aggravated thefts (mainly burglaries) with which the things stolen have been found for this is a good evidence. Thus, we are faced with the selection of these offenders already at the stage of discovering and investigating. In connection with this statement we have to mention that according to a special provision of our penal code the court may, in theft cases, remit the punishment if the perpetrator had restored the stolen goods before he learned of the commencement of criminal proceedings (art. 165/3 P.C.S.). It seems that such a situation did not occur since there wWas no remigsion of punishment in our sample. It might be that the restitution is more spontaneous with offenders who committed offences against economy or in official duty (art. 213 — 49 Ÿ ; art. 322 — 44 Ÿ ; art. 258 — 30 % ; art. 255 — 24 %) : lation. that is, with offenders belonging mainly to the middle class of popu- Considering the proposal that the restitution should replace other sanctions if posible, we are of the opinion that the return of illegally required things can hardly be treated in such a manner. The consequence would be that the offenders of the middle class of popu- lation would be privileged. 3) Adhesive procedure. The victims could find the satisfaction during the criminal procedure by the restitution of damage, by recon- ciliation or with renouncement to the claim. There were 33 % of such victims. From the remaining victims the claims for indemni- fication have been asserted in 2/3 of cases (64 %). The most claims have been asserted by the victims of malicious mischief (80 %) and the least by the victims of bodily injuries (45 % and 36 %). We tried to compare the ranks of asserted with adjudicated claims with the presumption that the victims expect the adjudications in advance due to experiences of other previous victims. Between both ranks we could not find any correlation. Therefore, we presume that the proportions of victims for bodily injuries who assert the claim are low due to personalized relationships between the offenders and the victims. This seems to confirm our view-point that it would not be advisable to force the victims to assert the claims for indemnification or to replace their claims with the claim of public prosecutor or with official adjudiœtion of the damage by the court. From all the claims for the indemnification the courts satis- fied the claims in whole or partially in 2/3 of cases (65 %). Among the claims the courts did not adjudicate we found 1/3 of claims upon which the courts have forgotten to decide. — 158 — TABLE I By the victims asserted and by the courts decided claims in % -— — % of | %ofun-! Hofnon ; Art. PC Victims decided | decided asserted Total saturated claims claims claims — r r 141 N 12,5 19,4 56,9 99,9 11,1 142/2 197 13,8 15,0 13,8 47,4 100,0 249 297 55,9 23,9 10,4 98 100,0 250 236 30,1 32,2 23,2 14,4 100,0 258 83 33,7 27,7 10,8 27,7 99,9 322 50 42,0 38,0 8,0 100,0 12,0 213 41 36,6 31,7 (4,9) 26,8 100,0 255 25 (12,0) 60,0 (8,0) (20,0) 100,0 257 18,6 419 23,2 16,3 100,0 246 a 66 16,7 42,4 9,1 31,8 100,0 Total 1111 32,3 14,4 24,3 99,9 N 1 111 358 321 160 270 Ncor (3) 1 012 33,3 27,6 24,0 lÔ0,0 The total amount of claims is 481, out of that 321 claims were decided by — — the_ _co_mjs | _66,7 % (with the cgqqçti_on 64,6). — —— —— —— — — —— Regarding the legal advisers we have found that only 13 % of injured parties have had a legal adviser in penal procedure (the legal persons more often than the physical persons). As far as the injured persons have had the legal advisor they asserted the claims more often and the courts have taken more decisions in these cases. This result supports our suggestion for the legal advisers free of charge on the request of injured parties if necessary (Selih-Kobe). The time between the commitment of damage to the finally- binding sentence was not very long. In less than l year 69 % of cases of bodily injuries have been finished (91 % in less than two years) and 72 % of cases of property offences in the narrow sense of the word (art. 249, 250, 258 - 89 % in less than two years). An exception of this is especially embezzlement in office with the longer period of adjudication, but in this case the injured party is a legal person. — — (3) Ncor correction of the 100 % observed offences to 25 %. — 159 — 4) The confiscation of property benefits acquired by the commission of a criminal offence was a measure the courts have applied only with four criminal offences out of ten observed and that in very low percentages (art. 250 — 3 % ; art. 258 — 3 % ; art. 322 — 7 % and art. 255 — 5 %). These percentages are of course too low to be explained with reasons quoted in item 2 and the responsibility for this is with prosecution and courts. But partially it is possible to explain such judicial practice with the mentioned high percentages of return of things with thefts and burglaries as well as with high percentages of restitution done by the middle class of offenders. S) The civil action has been asserted in very few cases. First we had to exclude from the observed cases all those in which the damage has been restituted, the reconciliation obtained, the victim renounced to the claim, the court has decided the whole claim and by grievous bodily injuries those perpetrated between the married couples or between the children and their parents or vice versa, since we have found that among these relatives there are no claims for the indem- nification. So we have foreseen the possibility of civil action in 360 cases of injured parties. Out of these the civil action has been asserted in only 31 cases or in 8,6 %. After three years 7 cases have not been decided yet, in ! case there was a withdrawal of the action, 8 cases were resolved by reconciliation and in 15 cases the claim has been adjudicated (in 7 cases in whole and in 8 cases partially). The reasons for so few civil actions are expenses (out of 31 cases 22 victims have had the legal adviser in civil procedure and only 7 in the adhesive procedure), the uncertainty about the decision of the court and long duration of the procedure. It means that civil action is not the way the criminal courts could rely upon by directing the injured person to assert his claim by way of civil proceedings. The final evaluation of the situation is : TABLE II The restitution of damage for offences observed in % 'T.he total number of victims (4). 1012 — the damage restituted, adjudicated in the adhesive procedure, reconciliation obtained of the damage claim renounced e © …. « « — 6 0 n < € € © ® © à € ® « % … »s %* % % 6 © % * * 61,0 % (4) The total number of victims is given with the correction for the offences observed on 100 % basis to 25 %. The total number of offenders with the correction is 792. The total numbers of observations (without corrections) were : 886 offenders and 1 111 victims. — 160 — — the damage adjudicated in civil procedure or recon- ciliated during the civil procedure 23 % — the civil record in three years after the finally binding penal judgement not finished yet 0,6 % — the damage remained unresolved __361 % Total 1000 % The worst situation in this respect is with bodily injuries, but in these cases 1/5 of them have been committed among close relatives. 6) The restitution of damage and the strictness of penalty. We have established that the increase of damage influences significantly the increase in the strictness of penalty. Unfortunately the data collected do not allow the comparison between the restitution of the damage and the strictness of penalty due to small absolute numbers and the complexity of comparisons needed. Therefore we want to draw the attention to the research of Mrs. Bayer (1967) in which the findings showed that the courts condemned the offenders with no criminal records almost in all cases to conditional sentences if the damage has been restituted. The frequencies of unconditionally condemned offenders were so small that they were not suitable for the analysis (p. 92). The recidivists, on the other side have been treated differently. The fact of recidivism outweighed the fact of the restitution of damage (p. 93). Both aspects raise some doubts. The inflicted damage is espe- cially with thefts and burglaries more the consequence of an objective situation (how much did the offender find) than of his primary intent. The restitution of damage is in these cases mainly connected with the return of stolen things. Therefore, every automatism in instruc- tions to courts how to decide in these cases would be a hindrance for a wise criminal policy. Nevertheless, it does not mean that it would not be appropriate to foresee in the penal code a broader possibility of the remission of punishment for some other offenders who restituted the damage except the one we already have in the art. 165/3 P.C.S. (such provision we find eg. in the penal code of German Democratic Republic art. 24, par. 2). 7) The suspended sentence under the condition the offender will restitute the damage. The courts very exceptionally pronounce the suspended sentence under this condition. If we do not take into account the bodily injuries (due to immaterial damage hard to define) — 161 — and the pillages (in which the suspended sentence is not possible) in all other cases in which the suspended sentence was applied it had been pronounced under the condition of the restitution of damage in only 44 %. But with that we have to recall our previous statement that the courts have sentenced the offenders with the suspended sen- tence first of all in cases in which the damage has been restituted. 8) Conclusions, Our starting hypothesis was that the penal courts do not apply provisions for the restitution of damage provided by our penal legislation which are mainly of facultative nature. the find- ings of our research do not confirm this hypothesis since we have found an unexpected great proportion of adjudications of claims for indemnification and quite reasonable motives for the small appli- cation of other provisions. This does not mean, of course, that the judicial practice should not strive to improve its work in this respect. Such a conclusion is not justified merely by the results of the research but also by the reasons for which we have excluded from research some other offences. So, e.g. we have excluded 26 % of offences in which the damage caused cannot be estabished in mone- tary value. Besides we have excluded also the offences committed in traffic (31 % of all offences) which we consider as specific offences, perpetrated more or less in all cases out of neglect, and sentenced for the consequences. The great amount of damage is covered by insurances and the offenders should be bound to restitute the damage to insurance funds as well as to injured persons. But whether this is a penal or civil matter is a question of principle which has to be resolved in the future. We believe however that the public opinion according to which the courts do not adjudicate the claims for the indemnification in the penal procedure is mainly due to these offences which covered a great proportion of all offences dealt with by the penal courts. With the new legislation the proportion of these offences will become smaller due to the change in the incrimination. In the introduction it has been mentioned that the reasons for the greater interest among the writers for the indemnification of the victims should be searched in three aspects : — in the aspect of the victim ; — in the corrective aspect for the offender : — in the sociological aspect. Ad 1!) With the findings of our research we cannot answer the question how many victims have remained unsaturated after the penal procedure has been closed for we did not interview the victims of unknown offenders and we did not follow the execution procedure. AIl we can do at the moment is to hypothesize that the number of — 162 — such offenders is smaller than we expected for the finally binding sentence gives the right for the execution and besides, the damage has been restored before the main trial in 1/3 of cases. In spite of that, our findings do not allow the conclusion that public funds would not be appropriate to compensate the damage in a speedy way (espe- cially for the damage inflicted by the offences against life and limb and ev. by the traffic offences). Due to the extension of the insurances in our country it seems, however, that such fund would not need to be extensive and could easily be filled with fines and with small contributions of all offenders condemned by the penal courts (Selih- Kobe). Ad 2) From the correctional point of view it would be advisable that the offender restitutes the damage inflicted. But in addition to that the research has drawn our attention also to some obstacles which hinder such a generalization. We have seen e.g. that some pro- perty offenders are the subjects of penal procedure mainly in cases in which stolen things were found and returned. In these cases it is difficult to see the restitution as having a correctional effect. With criminal offences perpetrated in our country against the economy the victims are mainly legal persons. It is a special problem to sensitize the workers that with criminal offences perpetrated in eco- nomic or public organizations they are damaging themselves and other workers. This is a process stimulated in our country by the selfmanagement system, which the oflenders evidently did not accept. It is a question whether it is possible to force such a perception with the restitution of damage by individuals to institutions which dispose of great amounts of financial means. However, this does not mean that we should not use also this way of solution. In these cases, the psychological relations only differ from the relationships among individuals, and the suggestion one should strengthen the direct individual contacts between offenders and victims obtains another content in these cases. Ad 3) From the sociological point of view the direct reconci- liation between the offenders and their victims should restrict the power of the state institutions and bureaucracy over individuals. This is an aim one should accept. But the actual proposals found in the literature in connection with the restitution of damage (which were contained in our headings) contribute to such an aim only in a limited way. The compensation by public funds, the penal and adhesive procedure are also institutional interventions into the rela- tions among people. It is true, however, that a good adhesive pro- cedure eliminates the need for the additional civil procedure. Besides that it would be necessary to search for a way toward the limitation of the institutional power by disciminations, lay reconciliation as well as by the enlargement of offences prosecuted by private persons — 163 — instead of by public prosecutor. But with all that we should not return to old days of private accusatory procedure and to times which Aristophanes exposed to mockery. The macrosociological analyses are a challenge for searching new ways and solutions. But contemporary proposals generalizing one way of solution do not always contribute to other and better relationships among people. The way out of the pitfall is still to be searched. The most radical proposals are those made by the Canadian Law Commision (on Diversion) and by Nils Christie. The idea is in both cases the same, that is to say the penal prosecution should be optional and introduced only if necessary in the individual case with a special explanation why. With this the principle of legality, on the basis of which the public prosecutor is bound to prosecute if the law demands so, known in most of the countries of the continental Europe, should be abandoned. I am not sure that such solution is acceptable in my country today. It is true that equality in front of the law does not exist anywhere. But does this mean that aspirations in this direction should be abandoned also ? One can reasonably presume all kinds of pressures on the public prosecutor. Would it be wise to expose him in such a manner ? I am not convinced that the social reality should be forced by legal provisions in the first place. Finally, the results of the research and the interpretations should be understood from the point of view of my social reality and the time which I am living in. REFERENCES ACIMOVIC (M.) : « Polozaj ostecenog medju gradjanima i pred krivicnim sudom », Gledista, Beograd, 1970, n° 4, pp. 627-636. BAVCON (L.), VODOPIVEC (K.), UDERMAN (B.) : « Individualizacija kazni v praksi nasih sodisc », Institut za kriminologijo pri pravni fakul- teti v Ljubljani, publ. 12, 1968. BAYER (M. « Kriteriji za odmjeravanje kazne od strane sudova u Sloveniji », Institut za Kkriminologijo pri pravni fakulteti v Ljubljani, publ. 10, 1967. CARIC (A.) « Odskodnina zrtvi kaznivega dejanja v sodobnem kazenskem pravu », Revija za kriminalistiko in kriminologijo, Ljubljana, 1970, n° 2, pp. 82-88. CHRISTIE (N.) : « Conflicts as property », The British Journal of Criminology, London, 1977, n° 1, pp. 1-15. — 1648 — GALAWAY (B. « The Use of Restitution », Crime and Delinquency, NJ., Hachensack, 1977, n° 1, pp. 57-67. GOLDSTEIN (N.) : « Reparation by the offender to the victim as a method of rehabilitation for both » ; in Victimology : A new Focus, ed. by Israel Drapkin and Emilio Viano, vol. II, Lexington, Lexington Books, 1974, pp. 193-205. GREBING (G) « Die Verhandlungen der III. Sektion über das Thema Die Entschädigung des durch eine Straftat Verletzten », Zeitschrift für die gesamte Strafrechtswissenschaft, Berlin, 1975, n° 2, pp. 472-485. GVOZDIC (P.) : « Priviligirani sviedoci iz cl. 211 st. 1 toc. 2 ZKP », Nasa zakonitost, Zagreb, 1971, n° 3, pp. 210-220. INTERNATIONAL STUDY INSTITUTE ON VICTIMOLOGY : Conclusions and recommendations, Bellagio, Italy, July 1- 12, 1975 ; Victimology, Washington, 1976, n° 1, pp. 130-153. JACOB (R.B.) : « Reparation or Restitution by the Criminal Offender to his Victim : Applicability of an Ancient Concept in the Modern Correc- tional Process », The Journal of Criminal Law, Criminology and Police Science, Baltimore, 1970, n° 2, pp. 152-167. KOBE (P.) : « Polozaj zrtve v sodnem postopku s posebnim ozirom na adhezijski postopek », Revija za kriminalistiko in kriminologijo, Ljubljana, 1977, n° 1, pp. 3-11. KOBE (P.) : Primerjalnopravni prikaz in analiza, manuscript, 1977. LAMBORN (LL) « Crime Victim Compensations : Theory and Practice in the Second Decade », Victimology, Washington, n° 4, pp. 503-516. LAW REFORM COMMISSION OF CANADA : « Restitution and compen- sation », Working papers 5 and 6, October 1974, Canadian Journal of Criminology and Corrections, Ottawa, 1975, n° 2, pp. 1-25. LAW REFORM COMMISSION OF CANADA « Diversion », Working paper 7, Canadian Journal of Criminology and Corrections, Ottawa, 1975, n° 3, pp. 1-25. LINDEN (M.A) « Restitution, Compensation for Victims of Crime and Canadian Criminal Law », Canadian Journal of Criminology and Correc- tions, Ottawa, 1977, n° 1, pp. 1-49. MARKOVIC (M.) : « Neka pitanja u vezi sa pravnim pojmom stete, njenih vidova i odredjivanja naknade », Pravni zbornik, Titograd, 1975, n° 1, pp. 35-49. McLEAN (I.) : « Compensation and Restitution Orders », The Criminal Law Review, January 1973, pp. 3-6. MENDELSOHN (B.) « Victimology and Contemporary Society’s Trends », Victimology, Washington, 1976, n° 1, pp. 8-28. MIKLAVCIC (F) « Odskodnina za nepremozenjsko skodo pri telesnih poskodbah », Pravnik, Ljubljana, 1973, n°° 10-12, pp. 419-429. MISIC (V.) : « Uticaj krivicnog postupka na tok gradjanske parnice i uticaj krivicne presude na gradjansku presudu i obratno », Glasnik advokatske komore u SAP Vojvodini, Novi Sad, 1972, n° 6, pp. 12-21 ; n° 7, pp. 1-7. — 165 — MUELLER (O.W.G.), COOPER (H.H.A. « Society and the Victim Alternative Responses », in Victimology : A New Focus, ed. by Israel Drapkin and Emilio Viano, Vol. II, Lexington, Lexington Books, 1974, pp. 85-101. OSTROUMOV (S.S.), FRANK (L.V.) « O viktimologii i viktimnosti », Sovjetskoje gosudarstvo i pravo, Moskva, 1976, n° 4, pp. 74-79. PALMER (W.J) « Pre-Arrest Diversion : Victim Confrontation », Federal Probation, Washington, 1974, n° 3, pp. 12-18. PECAR (J.), VODOPIVEC (K.), UDERMAN (B.), KROFLIC (M.) « Porav- nalni sveti », Institut za kriminologijo pri pravni fakulteti v Ljubljani, 1968, pub. 11. RESOLUTIONS DU XIe CONGRES INTERNATIONAL DE DROIT PENAL, (Budapest, 9-14 september 1974) Revue internationale de droit pénal, Pau (France), 1974, n° 3-4, pp. 685-689. SCHAFER (S.) « Compensation of Victims of Criminal Offenses », Revue internationale de droit pénal, Pau (France), 1973, n° 1-2, pp. 105-135. SELIH (A.), KOBE (P.) : « L’indemnisation des victimes d’infraction : Yougos- lavie », Revue internationale de droit pénal, Pau, (France), 1973, n° 1-2, pp. 294-313. VASILJEVIC (T.) « Sistem krivicnog procesnog prava », Beograd, Zavod za (zdavanie udjbenika socijalisticne republike Srbije, 2nd ed…, 1971. VASILJEVIC (T.), GRUBAC (M), KUHAIJDA (V.), BAJIC (L.), MATIC (V.) : Primena ustanova krivicnog procesnog prava u sudovima Vojvodine, Zbornik radova VIII, Novi Sad, Pravni fakultet u Novom Sadu, 1974, pp. 129-196. VIANO (C.) : « The Study of the Victim », Victimology, Washington, 1976, n° 1, pp. 1-7. WILLIAMS (L.V.), FISCH (M.) « A proposed Model for Individualized Offender Restitution through State Victim Compensation », in Victimology : A New Focus, ed. by Israel Drapkin and Emilio Viano, Vol. II, Lexington, Lexington Books, 1974, pp. 155-165. WOLFGANG (E.M.) : « Making the Criminal Justice System Accountable », Crime and Delinquency, Paramus N.J., National Council on Crime and Delinquency, 1972, n° 1, pp. 15-22. LAWS Criminal Code — Collection of Yugoslav Laws, Institute of comparative Law, Vol. XI, Beograd, 1964 — Translation and Comments by Mirjan Damaska. Code of Criminal Procedure — Collection of Yugoslav Laws, Institute of Comparative Law, Vol. XIX, Beograd, 1969 — Foreword, Translation and Comments by Mirjan Damaska. Kazenski zakon Socialisticne Federativne republike Jugoslaviie — Uradni list S.R.S., Ljubljana, 1976. Kazenski zakon socialisticne republike Slovenije — Uradni list S.R.S., Ljubl- jana, 1977. Zakon o kazenskem postopku — Uradni list S.R.S., Ljubljana, 1977. — 166 — POLICE AND MINORITIES IN ENGLAND GARY T. MARX MARY MORTON Department of Urban Studies Department of Sociology and Planning, Queens University M.I1 T. Cambridge, Mass. Kingston, Ontario - Canada Department of Sociology University of California Lajolla, California - U.S.A. Police anywhere are problematic as a public institution for a varicty of reasons, such as their authoritarian and adversarial role, the fact that they can legally use force, the opportunities for corruption they are offered, lack of agreement over the morality they enforce, and the fact that they must act quickly in situations where only limited information is available. The English police are of interest in this regard because they occupy a rather unique position in appearing to be more loved than hated, something that can hardly be said for police in most other countries. The degree of consensus and voluntary compliance with law, though easily exaggerated by the anglophilic American, ssem more pronounced than in most other countries. As with all stereotypes, the image of the English police 1s nb doubt overdone. Americans operating with contrast conceptions often exaggerate the virtues of other societies (particularly English and Scandinavian) and the problems and failings of their own. With only a little poking beneath the surface, tarnish can be discovered on the badges of at least some Bobbies. While we were in England, the head of the Scotland Yard narcotics unit was arrested for dealing in drugs. Custom’s inspectors traced a suitcase filled with heroin to him. In another case which was the subject of a parliamentary inquiry, the special criminal police (C.ID.) were hanging suspects from a wall and whipping them to gain information. The Challenor case, also the subject of a government inquiry, saw a policeman who in his zeal to fight crime went so far as to plant evidence on people. Some claim that English police carry guns under the seats of patrol cars. The image of comfortable powerlessness conveyed by police can be deceptive. The English police appear to have extensive powers — 167 — which do not make the subtrafuge and defensive sub-culture of Ame- rican police as necessary. There is no written Bill of Rights or Supreme Court Reviews and doctrines such as that in Miranda. In the area of stop and search they face far fewer formal legal restraints. À New York detective reported, perhaps with a tone of envy in his voice, « it’s as if Hitler wrote the laws under which the English police operate ». The South African head of state has expressed envy of the British Special Powers Act. Foreign students and radicals sometimes claim there is an extensive political intelligence police who know everything that is going on. The visible police are seen as a subtrafuge. The somewhat bumbling image they represent is seen as a front. Other more secretive branches such as the C.I.D. or special branch are said to do the really dirty work. Any discussion of the merits of the English police must also be tempered by their historically oppressive role in Ireland. The image of the helpful, smiling Bobby is not held by many of Irish and lower class background in England, not to mention those subject to the British colonial experience. One should also not go too far in trying to « explain » the good English police-community relations without more carefully ascer- taining just how good they are and what the dimensions of « good- ness » are. If we wish to say the English police are somehow « better », how are they « better» and what is to be explained ? Is it positive attitudes toward police that citizens hold as revealed by public opinion polls ? Is it a higher level of citizen cooperation with police ? Is it in more compliant behavior on the part of those involved in adversarial relations with police ? Is it greater police efficiency in deterring or apprehending those engaged in crime ? Is it better service during emergencies or conflicts ? Is it lower per capita costs ? Is it lesser police dishonesty ? Is it resort to lesser degrees of physical force ? Is it more respect for the liberty of the individual ? Is it more rigorous sanctioning of police abuses and taking citizen complaints more seriously ? Is it greater tolerance for ethnic diversity and alternative life styles ? Were precise comparisons available on all of the above dimensions, it is doubtful that the English police would uniformly appear better. Yet holding apart such nuances, it is still our impression that English police-community relations are relatively unique, even if they are far from the idealized image held by some. How is this to be explained ? Beyond historical factors and things such as police organization. ethos recruitment and training, the relatively good quality of police- community relations has rested in the social characteristics of the — 168 — English people. Changes in this area now make it necessary to question just how good police-community relations are in England Important to understanding any social control relationship is degree of consensus on the rules being enforced and the homogeneity of the population. One English policeman remarked « we have a good police because we have a good public ». Relatively speaking, English society has been homogeneous, consensual and integrated, and hence high on self-policing. Social and cultural homogeneity are important factors in the quality of service citizens receive from public service bureaucracies — whether police, medicine, schools or welfare is considered. What then happens when the English police encounter the youth culture and alien non-white ethnic groups such as West Indians and Pakistanis ? The system tends to break down. There is more conflict, and police-community relations in England, at least to the extent that they involve immigrants and those with deviant life styles, start looking much more like police-community relations with such groups in the U.S In this paper we explore some aspects of police-minority relations in England. This draws on 1) our observations and in-depth semi- structured interviews with police and minority group members 2) a survey we developed in conjunction with Conrad Jameson associates The survey was carried out in London in 1972-3 and involved 334 West Indians and 438 whites. We first consider some of the survey data and then some qualitative data Respect for the police among West Indians and whites For the purpose of assessing the general standing of the police in the eyes of citizens in a given social context, the extent to which members of the public say that they respect the police has often been used as an indicator (e.g. Reiss, 1967). In the present study, a composite measure of general attitudes toward the police has been devised, and as well, specific components of this measure have been examined in the anticipation that they may shed light upon the image of the police in London Our respondents were asked specifically how much respect they had for the police, and as Table IA revea! he West Indians in the sample were much less likely than the whites to say that they had great respect. Youg persons of both races were less likely than were older persons of both races to say that they respected the police. Interestingly, while there appeared to be a linear relationship — 169 — between age and respect for the police for both racial groups, it was those West Indian persons who were between age 25 and age 44 who were most ambivalent on this question, while the youngest and the oldest blacks were more likely to express defimte views — strongly positive for the oldest. Citizens in the United States have been asked in several studies (e.g Bouma, 1970 ; Bayley and Mendelsohn, 1968) how well they liked the police in their neighbourhoods, and the results have suggested strong racial differences in this measure of general regard for the police. Data from the present study reveal a similar pattern in attitudes toward the police in London, England. Whereas 60 % of white males and 57 % of white females say that the police are liked in their neighbourhoods, less than a quarter of the blacks interviewed held this view and almost half (42 % of males and 48 % of females) say that the police are nor liked, in contrast to only 15 % of whites holding this view. When race was controlled, there were significant tendencies among both whites and West Indian for young people to be more critical of the police than their elders, as Table I B indicates. Indeed, it is impressive to note that over 90 % of the youngest blacks held negative or ambivalent views as to whether the police were liked in their neighbourhoods, while just over half of the white youths between 16 and 24 years of age felt this way. Age along with race is a significant determinant of whether the res- pondents claim that the police are liked The homily that the British police are the finest in the world is a reflection of the general belief that the police in Britain are highly regarded. Consequently, our respondents were asked to judge the truth of this statement as a measure of their general views of the police The racial differences which showed up on this item were striking, as l'able IC shows While almost four-fifths of white males agreed with this statement, only a quarter of the black males said that the statement was true ; the patterns were very similar for females except that less clear-cut opinions were expressed, more females than males of both racial groups being inclined to say that they did not know. The data also indicate that within racial groups age differences were also evident, with young persons of both races being less likely to agree that the British police are the world’s finest, although this trend was more striking for whites than for blacks An indirect measure of esteem for the police which has been employed in some American studies (Reiss, 1967 ; Bouma, 1970) is the question of whether one would approve of his son’s becoming a — 170 — police officer (l). Interestingly, in our survey differences between West Indians and white respondents do not appear to be as great regarding the judgment that for one’s son to be a policeman might be a gcod idea as are the racial differences in the expression of the view that this would be a bad idea. Whereas 21 % of the whites would not approve of their son’s becoming a policeman, over half of the blacks sampled decried the idea. One possible interpretation of these findings :5 that for West Indians the idea of one’>s son becoming a law en- forcement agent has a political significance that is not salient for whites who may view the question simply in terms of the practical merits of the occupation. Consistent with this supposition is the finding that for blacks the tendency to support the idea of one’s son becoming a policeman increases with age ; if a civil rights rhetoric underlies young persons’ opposition to a family member’s being a policeman, this politicized view of the question is not so frequently held by older West Indians for whom the practical possibility of one’s children attaining jobs may be more salient. À composite measure of general regard toward the police was com- prised of the component items mentioned above and, as Table II shows, there were differences by race and by age in the tendency of respondents to score high on this measure. For West Indians, the clearest age differences were found to be between those under age 25 and those over that age, while for whites, the tendency to express generally positive views of the police increased more gradually with age, fully 81 % of whites over age 45 scoring in the highest category of general regard for the police. These findings tend to lend support to Banton’s (1973) contention that the « generation gap » may be more severe among non-whites than among white Britons. Nevertheless, the eifects of race are greater than those of age, as over half of even the youngest whites scored high on the index while in contrast over half (54 %) of the blacks under age 25 fell into the lowest of three categories on this variable. Willingness to cooperate with the police Respondents’ stated willingness to provide the police with assist- ance in the performance of their various duties provides a beha- vioural measure of the more abstract concept of respect for the police. Of course, it is impossible to say from an attitude survey alone whether (1) In contrast to the U.S., the English police have not tried as hard to recruit minority group policemen, but there are several in London whose fathers were police in the West Indies or who were police there themselves. One of the black policemen is sometimes predominantly displayed directing traffic in front of South Africa House. — 171 — the expressed degree of cooperativeness is reflective of actual behaviour, and so such measures are probably more useful as indices of the general image of the police and the extent of positive feelings toward what the police do. These dimensions of attitudes toward the police are interesting particularly because they may involve two countervailing tendencies on the part of citizens in orienting to the policeman’s role [f, on the one hand, the questions tap a ‘general appreciation for the jobs that police may perform, one might expect whites to score higher on measures of cooperativeness than do West Indians. If, on the other hand, as several American studies (Cumming, Cumming and Edell, 1965 Bercal, 1970) have shown, lower socioeconomic status persons are more likely to call upon the police in time of need than are the more affluent members of a society, then one might expect that West Indians would express the willingness to call the police as often or more often than wh1te Bntons Table III shows responses to several questions which inquired of respondents how likely it was that they would call the police under certain specific circumstances. On the question of whether to call the police if a fight broke out at the respondent’s home, over 60 % of both racial groups said that they would contact the police. In fact, fully 47 %, of the black respondents said that they would be very likely to call the police under these circumstances, whereas only 32 % of the whites were that emphatic. At the same time, however, a greater proportion of blacks said that they would definitely not call the police under such circumstances, although the differences between the racial groups were not grœt in this respect. Willingness to call the police if a fight broke out in one’s home was positively associated with age for whites, but age did not make a significant difference for West Indians. On the whole, these findings lend support to the idea that some West Indians would be likely to call the police, as would many American poor people, in time of need, while others of their race would actively resist doing so, partly as a reflection of a general negative attitude toward the police. For whites, on the other hand, the fact that willingness to call the police is associated with gc is consistent with the interpretation that this measure reflects general attitudes toward the police, as we have shown in the previous section that regard for the police also increases with age among whites. One of the most interesting findmgs in Table III is the great racial differences in responses to the question of whether respondents would call the police in the event that they observed a store being broken into. Here we see that while almost all (90 %) of the whites said that they would assist the police in this manner, just over half of the blacks were so disposed. The relationship between age and willingness to call the police when a shop was being broken into — 172 — appears to be curvilinear for West Indians, with the youngest and the oidest groups sampled being somewhat more reluctant to call the police than were blacks between 25 and 44 years of age. This finding may reflect a tendency on the part of the young and the older respondents not to want to « get involved » while those in their prime might have less to fear as far as their personal safety was concerned When the question concerned the respondent as hypothetical victim of a crime, the picture changed. In the instance of having had some- thing stolen from them, West Indians were almost as likely as were whites to say that they would contact the police, although the youngest West Indians (under 25) were less likely to say that they would call the police than were their elders. If the victimization involved a person whom the respondent knew, such as a neighbour who threat- ened, both white and West Indian respondents were considerably more reluctant to call in the police, and this was especially true for West Indians under age. 25 When the individual measure of willingness to call the police were combined into a composite measure of cooperativeness, the strong racial differences held up as Table IV shows. In addition, there is also evidence that while willingness to cooperate with the police increases in a linear fashion for whites, there is a sharp disjunction between blacks under age 25 and those over that age in their willingness to call the police in a variety of circumstances. The findings indicate that there are differences by race and differences by age, as well as differences attributable to the interaction of race and age, as black youths are the least likely to express a willingness to cooperate with the police. Fear of the police Research in the United States has shown that minority group zersons are more likely than are whites to express a general fear of the police, and that such fears are negatively relatèd to willingness to support. the police (Block, 1971 ; Bouma, 1970 ; Bayley and Mendel- sohn, 1968). Therefore, we dewsed a scale of att1tude items designed to measure the concept of fear of the police using items derived from Bouma’s research on adolescent attitudes toward the police in the United States. “The data dcmonstrate that the relat10nsh1p between race and fear of the police is very strong even ‘when age is ‘controlled. In the youngest are group. over half of the blacks expressed a high fear of the police while almost half of the white youths gave low fear responses. Interestingly, the differences by race are even stronger in the older age groups, with 67 % of whites over 35 expressing little — 173 — fear as compared with only 6 % of blacks of this age, and fully 55 % of blacks over the age 45 expressing hieh fear as compared with only 3 % of the whites in the oldest age bracket The effects of age show up more clearly for whites than for blacks as Table VI indicates, with young white persons under age 25 being somewhat more inclined to express fear of the police than was the case for those over that age. Nevertheless, these age differences were far outshadowed by those of race, providing a very clear picture of the way in which West Indian residents in London orient toward the police in contrast to white Londoners ; this also suggests the attitudinal foundation for potentially hostile encounters between West Indians and the police if, as Lambert (1970), Humphrey (1972) and others have indicated, West Indians may mask their fear of the police by taking the offensive with them These findings regarding levels of fear of the police may have significant policy implications. Some police in both United States and Great Britain hold the belief that respect for the police can be instilled in citizens, and especially in minority group persons, through the use of tactics which indicate that the police are worthy of being feared. In fact, our data showed that there were very strong negative correlations between fear of the police and other measures of general attitudes toward the police, such as general regard for the reputation of the police (—.51) and willingness to cooperate with the police (—-.51). While police may feel that fear inspires respect, the evidence is that citizens who fear the police will withdraw legitimacy from them and will refuse to aid them even in their law enforcement functions unless they personally are the victims of crime In an effort to measure the amount of positive or friendly contact with the police that white and West Indian respondents had had with members of the London police force, we asked them to respond to a number of specific questions, as Table V shows. The findi reveal that West Indian respondents were significantly less likely to report having had some positive contact with the police than were whites, and moreover that this was especially true for West Indians under the age of 25, only a third of whom reported any positive contact with the police at all. West Indians were less likely to have had a friendly talk with a policeman, and less likely to know a policeman by name, than was the case for withe respondents. On the other hand, blacks over the age of 25 were about as likely as whites of those ages to say that they had at some time gone to the police for help over something important, a finding which is consistent with those American studies — 174 — which have shown that minority group persons may rely on the police in times of emergencies (Cumming, Cumming and Edell, 1965 ; Bayley and Mendelsohn, 1968). Table VF shows that the relative amounts of negative contact with the police which West Indian and white respondents report that they have had with the police are surprisingly consistent with the observational data of Black and Reiss (1967) in the United States. That is, whites were more likely than were West Indians to say that they had ever been stopped and questioned by the police. In fact, over half (58 %) of the white males interviewed said that they had been stopped by the police, as compared with only 36 % of the West Indian males. There was no statistically significant difference in the rate of arrest for black and white males, although the percentages indi- cate that blacks were more likely to say that this had happened to them than were whites. In the case of both stops and arrests, males of both races were significantly more likely to report such negative experiences with the police thaq were females. Black-white differences alsc appear in evaluations of the nature of the police role and the quality of police performance (Table VII), thus West Indians rated police performance much lower than whites did with respect to the giving of help, respecting the rights of citizens, and controlling political demonstration. In a related item 75 % of West Indian males and 55 %, of females agreed with the statement « tho police arc really protecting the interests of people with power », for whites these figures were 41 % and 40 %- It has been contended by some students of race relations in Britain that the problems which West Indians have encountered in British society, and correlatively their attitudes toward British social insti- tutions, are partly the product of the recency of their immigration and also partly attributable to the traditions of the islands from which they came — Jamaica being seen to be more anti-British than, say, the Barbados which is thought to be staunchly British. Consequently, it was informative to discover in the course of our research that neither the island from which respondents had migrated nor the length of time which they had been residing in Britain was of any impor- tance in predicting attitudes toward the police. With respect to social class effects, our data tended to bare out the conclusion which has often been arrived at in the U.S. (e.g. Bayley and Mendelsohn, 1968) that socio-economic factors are less significant determinants of attitudes toward the police than are race and age. Four different measures of socio-economic status were employed and none of these showed a clear relationship to general attitudes toward the police for either blacks or whites in our sample. — 175 — Contrary to the « optimistic » view that as West Indians become better acquainted with British society they will come to share the native Briton’s positive view of its most sacred institutions, we found that the length of time respondents from the West Indies had been in Britain had absolutely no effect upon their attitudes toward the police, and indeed there is some indication that for those who have been in the country longer, relations between the police and West Indians appear to be worse than for those who have more recently immigrated. Perhaps a decade or more of residence in the country with no tangible evidence of improvement in one’s life situation has the effect of embittering people and of passing that bitterness along the children who have been raised in Britain. RESPECT FOR POLICE Table ! West Indians Whites A) « Considering everything about the way British police do their job, would you say that you had a lot of respect for the British Police, not much respect for the British police or mixed feeling about them ? » (% saying « a lot of respect ») ». e8e0ess066sseenen006s (334) (438)* B) « What about this area, how wéll liked would you say the police are by the people who live in it ? » (% saying « very well liked » and * * + % .. 08500858e8e660ss 22 59 « liked well enough ») C) « People say the British police force is the finest in the world. Do you think this is true or not » ? (% saying «true») ..........… 25 T7 D) « Imagine you had a son aged 18 or 19, how do you think you would feel about the idea of him joining the Police ? » (% saying it would be a « good idea ») . à ... 53 (*) Numbers in our total sample, these vary slightlv depending on the item in question. — 176 — GENERAL REGARD FOR THE POLICE Table II —— — -— West Indians AGE . …. . ........% . 16-24 25-34 35-44 45 All ages % HIGH + » e « e © % € ® à 15 'Ï 38 39 | 34 Whites AGE 16-24 25-34 ! 35-44 45 + | All ages % HIGH » e …. .0%6%6e6e e 63 68 81 66 e =- 0s — srr 5 00e WILLINGNESS TO COOPERATE WITH THE POLICE Table III West Indians | Whites A) «If, for example, you had a party and a fight broke out that looked as though it might get out of hand, how likely is it that you might call the police ? » (% « likely ») 66 B) « And, if you saw someone breaking into a shop, how likely is it that you would call the police ? » (% « likely ») « « e « . à % 6 %%% 6 % © #% 0 ® © @ 57 C) « And, if you were having trouble with a neighbour who threatened you, how likely is it that you would call the police ? » (% « likely ») e …. .......................e.e 53 53 D) « And, if someone stole something valuable from your house, how likely is it that you 86 96 would call the police ? » (% « likely ») ... — 177 — COMBINED MEASURE OF WILLINGNESS TO COOPERATE WITH POLICE (based on items in table IIT) Table IV West Indians AGE oooooooooooooo 16-24 25-34 35-44 45 + AlÎl ages % HIGH ON WILLINGNESS TO COOPERATE 31 47 53 52 î 46 Whites AGE 16-24 25-34 35-44 45 + All ages _ % HIGH ON WILLINGNESS TO COOPERATE 74 1 Ë 78 rrs _ 87 Î 71 CONTACT WITH POLICE Table V _ p SE SE West ‘ Whites —- —— —— — Indians_ | A) « Have you ever had a friendly talk with a policeman who works in your e . . + ». » .. 26 35 neighborhood ?» (% yes) B) « Do you know any policeman by < … « © » 000460008980 name ? » (% vyes) 21 C) « Have you ever gone to the police for help over something important ? » (% yes) 33 39 D) « What about yourself, have you ever been treated unfairly by the police or Males 30 20 the law in general ? » (% yes) ... Females 13 10 E) « In fact, have you ever been stopped by the police and asked what you were Males 36 58 doing ? » (% vyes) Females 13 F) « In fact, have you ever been arrested ! Males ; 17 12 . ». e — e e % 6 — e © — — — —— _— ——e—sreatee yourself ? » (% yes)_ _Females — 178 — FEAR OF THE POLICE Table VI —— — — — West Indians AGE oooooooooooooo 16-24 25-34 35-44 Û 45 —- AlI ages % HIGH « « « . . ... 57 53 54 55 55 % LOW « …. © 0000600s â 9 6 l î Whites AGE . . .. .......... 16-24 25-34 35-44 } 45 AlÎl ages % HIGH .… | 15 7 6 ! % LOW . e » t — e — © = — 45 48 64 70 55 — —— p E —— — — EVALUATIONS OF POLICE PERFORMANCE Table VII —-— W est —0 22200 ce — Whites | Indians — ! A) % saying police are doing a « very ‘ Males ! 42 62 Females 51 64 good » or « good » job in giving help. B) % saying police are doing a « very Males 22 63 good » or « good » job in respecting the law and the rights of citizens Females 34 64 C) % saying police are doing a « very good » or « good » job in controlling Males 47 69 political demonstrations Females 57 75 D) % agreeing « the police are really protecting the interests of people with Females TS 41 power » . .. .0..................... Males — ——c >3 — 179 — Some qualitative data Let us move from the statistical data to some specific examples that illuminate increased police-minority conflict. An important part of the greater West Indian dissatisfaction with police would appear to lie in culture conflict and the breakdown of communication. Such explanations are often put forth to account for American police- community issues. Thus police may drive young men off street corners when hanging out is part of their life style or they may arrest people whose recreational interests involve gambling, homo- sexuality and drugs. In the United States context such an argument is often exaggerated because almost all groups share important elements of a common culture. Such an argument may also mask more basic causes of group struggle involving racism and class. But in the English case they clearly apply, given the newness of the immigration to England and the extensive differences between the immigrant and host cultures. À number of examples of this culture conflict can be seen. Pakistani Moslem immigrants have special rules about how the animals they eat should be slaughtered. Thus, Moslem chicken sales- man may go to immigrant areas with live chickens and kill them on the spot. This may upset English neighbors who summon a policeman, though there’s probably no law against killing chickens on the streets of London. Similar problems may emerge with the sacrifice of a goat. In another case a bearded and turbanned Seik came into a police station in a rural area late at night and said in excited and accented English « they’re fourteen men with sticks who are chasing me and want to kill me ». The old sargeant looked up at him in disbelief and said « sure there are fourteen men with sticks that want to kill you ». Fortunately another person was in the station who had some understanding of Seik customs and the local situation. The frightened man had taken an unmarried female alone in a car to a wedding, and this had embarrassed and humiliated her relatives. They then set out after the man. The police then intervened and protected him. The social structure of some Pakistani villages has been repro- duced in England. Along with this may come traditional social control mechanisms, such as village councils sanctioning people. In one case the police intervened on behalf of a man who had received a sentence of 50 lashes from his village council in England. We were even told in wWhispered terms oï one English policeman born in India, who could speak Hindi, being given the power to be a tribal judge. In West Indian culture, an emphasis on drama and a belief in conspiracy may make initial police efforts to get information more — 180 — difficult. Varying cultural definitions of the tactile may also be relevant. Touching people has a different meaning than it does in England. A West Indian may interpret a policeman grabbing his arm as a hostile act. For a native born English person this is less likely to be true. West Indian youth often carry nine inch wooden afro combs. This sometimes results in their being convicted for possessing « offen- sive weapons ». West Indians appear to place a greater emphasis on street life. West Indian youth often congregate outside their homes. Police are not accustomed to such a life-style and often try to move them indoors and off the streets. À tradition of open parties where all can join in, even people walking down the street who hear the music, can generate conflict with police. Liquor may be sold by the cup, violating licensing rules. Police receive many complaints of noise even though no laws are being broken. Some immigrants interviewed felt that to police, they all looked alike. Problems with mistaken identity appear more likely when immigrants are involved. À person may be accused of something he didn’t do by the inability of the accusers to differentiate facial types. There are also problems in locating or tracking down immigrants wanted for crimes, for help as witnesses, or simply to give infor- mation. À Pakistani involved in an incident with police will give his name as Singh and his address. When police show up at his house twenty people all named Singh will appear. In lower income white areas it is sometimes said that a man either becomes a cop or a thief. In such areas police and those they deal with know each other. Stable accomodative relationships may form. But for immigrants such relationships have not had time to form. Lack of knowledge of the English system may result in immigrants unknowingly breaking English law, or interpreting legal and depart- mental sanctioned police action (or non-action) as examples of discrimination. There are cases of West Indian parents beating a child as they would if they were still in the West Indies. In the West Indies, they are perhaps less likely to be sanctioned but in England such behavior can result in jail for the father and the child being taken away. Its not uncommon for Indians and Pakistanis to arrange marriages for their daughters. The daughters who have grown up in England, gone through the English school system and heard the Beatles and the Rolling Stones sometimes resist arranged marriages and run away from home. The family reports their missing daughter to the police, — 181 — who will locate the girl and tell the family that she is well, but they will not reveal her address. This infuriates the parents who believe the police are discriminating against them. An immigrant stopped by police on suspicion, may not know that police can legally do this. À stop and search by a policeman 1s more likely to be seen as an example of prejudice by a West Indian. It may be felt that police do not similarly stop whites. As one West Indian said, « Hell, every black in a white society is a suspicious person ». Police reciprocate and see West Indians as being on the defensive and as more likely to attack them than whites. Many police interviewed felt that black complaints often involved routine police behavior which the complainant mistakenly saw in racial terms. We presented to our respondents a hypothetical situation of a young man being stopped and questioned in the area where a number of robberies had rccently occurred and asked them to sav whether the policeman was allowed to do this by law. While over three-quarters of the whites actually knew that the police had the legal power to make such an investigation, only about half of the West Indians were aware that this was within the prerogatives of the police and over a third actually said that such a policeman would have been « going beyond the law ». These findings suggest that one reason why West Indians may feel that they are treated roughly or unfairly by the police, whereas whites are less likely to claim that this is the case, is that West Indians are not fully apprised of what the police have a legal right to do Vvis-a-vis any citizen in London. In the West Indies, police seem to play a more diverse role. They will punish as well as adjudicate. À policeman called to a house where the husband is drunk may hit him a few times and then take him to lail to cool off. À policeman may punish an unruly child. The police- man may help to adjudicate a landlord-tenant dispute. In England police actions are more formal and restricted. In landlord-tenant disputes, for example, police often claim that there is nothing they can do. They may refer the complainant to a rent board, however, the immigrant may see the policeman’s failure to take action as discrimi- natory. When called upon to settle a dispute between two or more parties, the police sometimes find that the expectations of citizens for their intervention actually exceed formal legal powers which the police are authorized to use. In part, such a situation may arise because citizens are not adequately aware of what the legal limitations on the police are, and in part the difference between expectations and reality may reflect dissatisfaction on the part of citizens about the extent to which they perceive the police as being willing to help them resolve their — 182 — difficulties. Consequently, we asked our respondents about a situation in which a policeman says there is nothing that he can do in regard to settling a disagreement between a landlord and a tenant and found that white respondents were overwhelmingly aware that the police were not legally authorized to intervene in such matters while almost a quarter of the West Indian respondents felt that the policeman in the situation described simply did not want to get involved. In a similar vein, we asked the respondents whether they felt that the police ought to be concerned with matters involving landlord discri- mination and found that while whites were sympathetic to such an idea (75 % felt that the police should be concerned), West Indians were somewhat more likely to feel that such matters should be part of the police mandate, well over half saying that the police should be very concerned with landlord discrimination. The tendency for West Indians to feel that the police ought to be very concerned with protecting the interests of individual citizens showed up even more clearly in regard to a question concerning whether the police should become involved in controlling the practice of overcharging on the part of shopkeepers. While just 60 % of the whites interviewed felt that such matters should be the concern of the police and about a quarter said that this should be none of the police’s business, fully 90 © of the West Indians said that this was a legitimate concern of the police. These findings may suggest that either the West Indians really are more likely to be victimized by unscrupulous busi- nessmen that are whites or that West Indians are more sensitized to this form of exploitation : in either case it seems clear that the West Indians sampled are overwhelmingly convinced that this 1s an area of life which the police ought to be regulating, even though it is not now part of their legal mandate to exercise control. While we did not collect systematic survey data on police attitudes toward immigrants, we did carry out numerous interviews. Relative to the U.S. there was much less realization of police-minority problems. One police commissioner said that 98 % of the immigrants were satisfied with the police. Another told us that there have been « no really serious immigration problems since Knotinghill ». For him a problem was seen as a riot and not police mistreatment or lack of service. We found a greater stress on a perceived need to « make immigrants become like us » than is the case in the United States. Pluralism seemed less appreciated, even at the level of lip service. This may be related to the fact that West Indians, unlike American blacks, are seen as non-citizen alien immigrants. There was considerable racial stereotyping among police. Just as American police offered George Wallace disproportionate support, so too do English police appear to offer disproportionate support to Enoch Powell. — 183 — West Indians, and especially Pakistanis, may expect police to be corrupt, based on experiences in their homelands. Pakistanis who may be accustomed to bribing police, may see an English policeman’s failure to accept a bribe as a type of prejudice against them and be resentful (2). The opposite, the failure to involve police or cooperate with them, may also be a source of conflict. As immigrants with a tenuous status in the country, minority group members are often hesitant to get involved with police in any fashion, even when they themselves are victimized or have done nothing wrong. Police report that whites are much more cooperative, even during the investigation of crimes where immigrants are victimized. Police drew distinctions between the Irish and West Indians. While the Irish were seen as rowdy, at least they were thought not to carry knives. The use of a knife in a fight was seen as an unmanly act. Irishmen were seen to fight ferociously with police in bar room brawls, but to be repentent the next day in court, apologetically telling the Judge, « the drink made me do it your honor ». Defiance rather than repentence was seen to be more characteristic of West Indian responses. Some Implications Over the course of the hundred and fifty years since modern metropolitan policing was begun in England and America, we have observed three phases of similarities and differences between policing in the two societies. At the beginning, large cities in the two nations faced similar problems and adopted similar solutions in the form of a bureaucracy of constables trained to serve and protect citizens. For more than a century, however, policing in England and America diverged from one another in important respects including the structure of the police organization, the powers of the police and their accoun- tability to civil powers and citizens, the degree of centralization of their operations and their isolation from the public ; and these struc- tural differences gave rise to a very different image of the police in the minds of the citizens of the two nations — the one, the benevolent and wise role-model, and the other, the crude and ruthless wielder of truncheons. Policemen, too, have shared these conceptions of the essential difference between British and American policing. When (2) Several people related the story of a rich Pakistani landlord who was involved in a dispute with a poor tenant on his land. While the landlord was traveling abroad, the tenant took the matter to court and a judgement was rendered. The landlord’s lawyer wired his client about the case « Justice previals ». The landlord wired back « Appeal at once ». — 184 — asked what might happen if the police forces of London and New York were exchanged, the Commissioner of the London Metropolitan Police said, «The New York streets would be filled with dead bobbies and the London jails would be filled with New York cops ». However, the second half of the twentieth century has brought about a renewed convergence of the patterns and problems of policing between England and the United States, a convergence which appears to be based most fundamentally in the structural characteristics of the societies being policed. Policing in England and the United States had similar sources in the perceived demand for order by the propertied classes of the nineteenth century who feared social changes brought about by rapid urbanization and industrialization and who associated their fears with the burgeoning urban masses which were thought of as the « dangerous classes ». However, whereas the solution adopted in Metropolitan Lon- don of creating a highly disciplined organization of « professional » constables rather quickly established a modicum of order and achieved acceptance by many citizens, in the U.S. there was not, as Allan Silver (1967:5) put it, «a comparable sense of relief or improvement ». Throughout the nineteenth and early part of the twentieth cen- tury, policing developed along diflerent lines in the United States and Great Britain. Whereas in the United States there was a centrifugal tendency for local, county and state jurisdictions each to set up their separate forces with little coordination amongst them, in England and Wales there was a spread out from the centre, London, of a standardized method of policing and a growth of the monitoring capacities of the central government. The result of this latter pattern was that the legal powers of the English police have continued to expand with little opposition from most citizens and little generalized sense that civil liberties have been endangered. In contrast, American police have found themselves in a position of having their legal powers systematically curtailed through the efforts of the Supreme Court in response to protest from segments of the population who fear the unbridled power of the police. The difficulty of developing a way of policing a society composed of disparate ethnic and racial groups in which interests are forcefully articulated has been seen as a major social problem in the United States for well over a century. David C. Perry (1975:47) takes the American case to exemplify the dilemmas of modern policing itself : « The history of the police function in the United States has been the history of a nation’s confusion over how to police itself…. The questions a metropolitan nation has yet to answer are : What is it we want the police to do ? How is it that we want our lives protected yet undisturbed in the name of such protection ? ». — 185 — Dismayed and sometimes appalled at the quality of policc performance in many areas of the United States, and disheartened by the apparently low level of regard for police expressed by many citizens, American students of policing have frequently turned to England for the example of how « good » police-community relations could be established. They have looked with reverence at the capacity of the British policeman to command respect from citizens even though he has no more coercive capability than do ordinary persons. Often they have interpreted the apparently great deference shown by British citizens to their police to the long-standing cultural traditions of respect for authority and civility which contrast with the frontier individualism and rebelliousness of Americans. Perhaps it required a British sociologist to identify for Americans the part that structural factors play in the extent to which legitimacy 1s accorded to the police by citizens. In 1964 Michael Banton drew attention to the relative homogeneity of the population of Britain as compared with the United States and predicted that as Britain became more racially heterogeneous, consensus about the police would decline and patterns of policing would necessarily change. Whereas the United States has experienced an almost constant flux in the composition and movement of its population over the course of its history, England had remained almost exclusively a « white and tight little island » until after World War IL. Not until the nineteen- fifties was there any appreciable in-migration of persons from the former colonies of the Third World. Thus, the chronicle of non- white immigration to Britain is the story of how in thirty years time a nation has been awakened to the fact it is no longer homo- geneous, that long-shared and often unspoken understandings might no longer be taken for granted, and that conflict over the rights of citizens and the obligations of authority would manifest themselves in connection with the role of the police. Policing in the United States since the Second World War has been analyzed as involving a growing attention to the law enforcement aspects of the social control function. Maureen Cain (1973) has suggested that such changes are already becoming apparent in British policing as well and has predicted that a new consensus surrounding law enforcement issues would emerge as conflicts about other aspects of the police role — such as peace-keeping and service — became manifest. At the same time, however, it is argued in both the United States and Britain that as police have relegated the tasks of order- maintenance and service to secondary positions in their priorities, they have tacitly acknowledged their allegiance to dominant segments of the society and have distanced themselves from accountability to most citizens through adoption of the stance of professional crime-fighters. — 186 — But the structural bases of conflicts which surface in demands for order and service do not evaporate because police disattend to them. Indeed, research in the United States in the past decade has highlighted the fact that the most vocal protests about the quality of police service and the conduct of police functions emanate from those segments of the society which perceive themselves as being alter- natively harassed and neglected by police, whom they view as ignoring theitr demands in favour of the interests of people with power. It appears that the same patterns of protest about the police are currently emerging in England ; namely, that minority group persons as well as young whites are less than satisfied with the quality of policing which they receive, and moreover that their discontents are centred upon the order-maintenance and service aspects of the police function. Granted, there is a fairly high degree of consensus that the police ought to be concerned about crime control, but those persons who are unsatisfied with the way the police serve their needs and answer their demands for order are likely to feel alienated from even the objective of crime control when it involves their participation in the process of preventing and apprehending offenders. If there is a « new consensus » about policing in Britain, or in the United States, it is a narrow one which leaves many persons feeling deprived of the respect and protection which, as citizens, they feel they are entitled to. There- fore, as police in both England and America emphasize their role as professionals in crime-fighting, they incure, on the one hand, a qualified appreciation for their specialized expertise and, on the other hand, a loss of the respect which had been associated with the more diffuse aspects of their function. As citizens become sensitized to the limited accountability of police to the public, the demand for cur- tailment of police power grows and the conflicts between the policers and the policed become salient issues for debate. REFERENCES : BANTON (M.) : The Policeman in the Community, Basic Books, New York. 1964. Police Community Relations, Collins, London, 1973. BAYLEY (D.H.) and MENDELSOHN (H) Minorities and the Police Confrontation in America, Free Press, New York, 1968. BERCAL (T.F. « Calls for Police Assistance Consumer Demands for Governmental Service », in Harlan Hahn (ed.) Police in Urban Society, Sage Publications, Beverlv Hills, California, pp. 267-278. BLOCK (R.L.) : « Fear of Crime and Fear of the Police », Social Problems, Summer 1971, pp. 91-101. — 187 — BOURMA (D)) : « Youth Attitudes Toward the Police and Law Enforcement », in J. Curran et al. (eds), Polzce and Law Enforcement A.M.S. Press, New “(>Th 197/2. Pp. 219-238. COIN (ME) : Society and the Pohceman S Role Routledge and Kegan Paul, London, 1973. CUNNING (E) et al. « Policeman as Philosopher, Guide and Friend », Soc1al Problems, n° 12, Winter 1965, pp. 276-286. HUMPHREY Police Power and Black People, Panther Books, London, 1972. LAMBERT (J) Crime, Police, Race Relations Oxford University Press, London PERRY (D.C.) Police in the Metropolis, C. Merrill, Colombus, Ohio, 1975 LiSS (A îx) « Public Perceptions and Recollections About Crime, Law Enforcement and Criminal Justice », Vol. 1, Field Surveys III, Studies in Crime and Lan Presidents Enforcement in Maior M opolitan Areas Conynission on Law Enforcement and Administration of Justice, Washington D.C 1967 SILVER (A.) ; « The Demand for Order in Civil Society », in D. Bordua (ed.) The Police, John Wiley and Sons, New York, pp. 1-24 -- 188 — V -Methodology DECISION THEORY AND COMPARATIVE CRIMINOLOGY LESLIF T. WILKINS School of Criminal Justice State University of New York Albany, New York - U.S.A. A Decision Network Model for Criminal Justice. There are many ways in which the criminal justice system may be seen. We may focus upon the information which is passed along various channels and attempt to relate the record to ideas about actual behaviour. On the other hand we may take the view that whatever decisions are made, they are made on the basis of available information, irrespective oË its quality. And there arc processes which lie some- where between these two approaches. Information which is believed to be more reliable may be more frequently selected as a basis for decisions, or items regarded as unreliable may be compared and combined with others and inferences made accordingly. For the accused who is caught up in the machinery of the system the whole system may be seen as a network of decisions made by persons in authority, treating him rather like a truck in the railroad marshalling yard. He may have some vague idea of what causes the « points » to be switched, but the location of any « point », whether correctly or incorrectly set, will determine his location at any time in this network. The idea of a system as a network of decisions is not new. In the criminal justice field the majority of « switchings » release the person, initially caught up in its network, from further consideration. The greatest number of persons go the shortest distance in the pro- cessing. The initial input to the system is, of course, the « suspect set ». From this set a selection is made of persons who may be inves- tigated, and from that smaller set, those who may be charged with a crime. And so on. Comparative criminology may be concerned to describe and compare the networks of decisions, or « switching systems » as they are organized in various parts of the world or at different times in the same location. The depth of any such analyses may vary quite validly from concern only with the officially determined machinery — 191 — of justice to concern with the ways in which the official forms are modified in practice. Another approach, and one which it is intended to develop in this article, concerns the nature of the « switching apparatus » itself, whether this official or unofficial. In other words. we are interested in how it comes about that a particular « switch » sà 1S moved in c particular manner. We will then begin to suggest ways in which this approach may be developed into a comparative form, giving some interim results. Information as « key ». The analogy of a « switch » applies to one of the main ways in which we use the word «decision». In order to generalise the model we will examine the idea of a « decision » a little more closely. This should also provide some basis for the further analyses we shall make. Although we often talk about decisions as though there was such a thing as a « decision process » it is clear that more often than not we use the word to indicate the ferrnination of some process, rather than the process itself. This fact is captured in the analogies and metaphores used in relation to the idea of decision-making. We talk, for example, of « having arrived at a decision » — clearly an allusion to a journey completed. We talk of decisions with allusions to « weighing » as in a balance. The etymological base for the word is related to the latin decidere, «to cut off » and also to our word deciduous « to fall off ». Sometimes we might say that we « concluded that… » as an alternative to saying that we had decided. If the « deci- sion » is an end, of what is it the end ? It is not very satisfactory to regard it as an end to itself. That is to say a « decision » is not well defined as that which terminates the « decision process». It is, of course, the information search process which is terminated by the decision. If a person says that he has « decided » but continues to search for more information related to the issue, we would assume that he was using the word « decision » incorrectly. He might have made an « interim decision ». a term which is almost a contradiction in itself. If the argument of the preceeding paragraph is accepted, we can now' identify a process of information search which takes place prior to the action of a « switch » (decision). It is possible to postulate that there will normally be « strategies » of information search, and that different decision-makers may prefer diflerent strategies. Ît is only a small step, then to argue that preferences for strategies of information search will be related to the utilities sought by the decision-makers concerned. We may be able to infer the utilities from a study of the information scarch, or conversely, we may make — 192 — projections as to the kind of information search strategy likely to be adopted, if we assume that we know the utilities sought. Certainly, if we postulate that different utilities are sought at different times or places, then we would expect the preferences for strategies of infor- mation search also to vary. Preferences for strategies of information search routines will be observed in relation to the sequence employed by decision-makers in examining data items. There may be other ways in which information- search strategies might be classified, but the sequence characteristics would seem to be a simple one to employ. Unfortunately it is by no means simple to identify and to classify search strategies, nor is it a simple matter merely to identify the sequence of search. The File Problem. Case files which are the repositories of much information about persons dealt with (decided about) in the criminal justice system present considerable difficulties for anybody who wishes to study deci- sion-makers and information use. The persons of the decision-makers themselves also present difficulties. Most social studies assume that persons engaged in an operation are reasonably self-aware — that they can report upon their activities with some validity. In the decision- making task this assumption is unsupported by any evidence. Perhaps because the decision-maker is thinking about the task in hand, he cannot at the same time think about the process which is itself of the same qualitv. In other words, it is difficult to think about some- thing, and at the same time to think - about - the - thinking about that thing. For whatever reason, decision-makers are poor witnesses as to the nature of their operations. This fact is important in several ways ; two of which concern us at this point in the analysis. The first problem was noted several years ago — probably first by Haÿes (1) who observed : « It is commonly assumed that the more relevant data one takes into account in making a decision, the better that decision will be. It is clear. however. that as one takes more relevant character- istics into account for comparing alternatives, the opportunities for confusion increase. If confusion were to increase rapidly enough as the number of characteristics increased it is conceivable that decision-makers would perform better if some of the relevant data were eliminated ». Note that Hayes refered to «relevant data» — not merely to more data. How fast does confusion increase as more relevant characteristics are taken into consideration ? The correct answer is — 193 — usually rejected by decision-makers as being totally unreasonable. It seems unreasonable because decision-makers are unaware of the ways in which they « work around » the problem of information overload, failing to recognise the devices they employ. Extensive testing by Hayes and others « has suggested that the maximum number of data items along different dimensions which can be processed profitably at the same time, without any formal decom- position and restructuring of the decision process (sic), is about eight » (2). Note the phrase, « decomposition and restructuring » ; it is by this means that a greater number of data items can be accommodated. But again, decision-makers are not self-aware of the restructuring processes they employ any more than they are aware of any information search strategy for which they seem to reveal à preference. The file — the usual means for the presentation of information to the decision-maker — is not organized to accommodate any individual strategy, perhaps because files are usually designed by committees, and committees will represent as many strategies as members. Even if a group of decision-makers were aware of the limitation of eight items of data and wished to design a short form, the number of items listed as essential by a committee of ten would be expected to exceed fourty, in order to ensure that every member had access to his specific eight items. For any decision-maker working with data from a file, the file contains an excess of information on some topics or areas in which he is interested, and a paucity of data on others, unless the file is extremely large. The decision-maker is then faced with the problem of extracting his particular « gold » from a very large qualtity of dross. Furthermore, in the course of this search for the specific items suited to his personal strategy, he will suren- dipidously hit upon other items of data which may register in his mind although not sought in relation to the issue of his decision. It is possible to characterise the case file or other record of infor- mation regarding offenders as an « information rich » source (*). A decision-makers seem to carry out two tasks in one operation and to be unaware of the transition from one to the other. First, when the task of deciding begins (such as consideration of parole), information will be sought which is seen to be directly related to the issue. However, quite quickly and imperceptibly, the task is re-structured to one of searching for information which would fest the quality of information that was among the items sought earlier in the process. The first task rich» field and is of significance in adaptations of all forms of econometric (*°) This is also true in the more technical use of the term « information- analysis to crime and criminal justice data. relates directly to the issue of concern (such as, say, parole), whereas the second is concerned with the possibility that evidence which had earlier tended to « swing » the decision towards one or other outcome, might be unreliable. The first question addressed may, for simplicity be stated as, « Can or should I parole this person ?», while the second question becomes, « Can I rely upon the impression I formed from xyz item(s) of data I saw m…ally ? ». These are, of course, two quite distinct decisions and require quite different kinds of mformaüon and different search strategies. When the task is a selection of an alternative, the items of information sought (in order to be of maximum power) should be strongly correlated with the discrimination (criterion), but should not be strongly correlated wih each other. To test an item for reliability, it is appropriate to seek an item which would be expected to be correlated with it. For example, if we suspect the data item «number of arrest», we might examine the data item, « number of convictions ». However, if we trusi the data item. « number of arrests » we Will gain no significant increment in information by looking also at the number of convictions From examinations of decision-makers operations under expe- rimental simulated conditions (3) it has been observed that after about six or eight items, decision-makers are more likely to seek items of data correlated with those they have seen earlier. This is a most reasonable strategy if we assume that the nature of the task performed has changed from being concerned about the initial problem to one of a search for verification of data. To put it another way, the problem addressed changes (imperceptably to the decision-maker, it seems) from a search for information relevant to the decision (the selection of alternatives) to a search for information-about-information In general we may state that information which assists in one decision may be contra-indicated for another decision, no matter how closely similar superficially they appear to be Decomposition and Restructuring of Decision Sets. The switch from one decision to one which appears to be related is one way in which decision-makers re-structure the decision task, probably unconsciously in most cases. Some re-structuring may be valuable in terms of efficiency of decision-making, while other forms may not be desirable in these terms. Some strategies may give the decision-maker some psychological satisfaction (a sort of warm feeling) which may be unrelated to performance, or positively or 195 negatively associated with efficiency. Some re-structuring may be sub- divisions of the task. Consider, for example, the disposition of a person found guilty by the court. The task is the selection of the appropriate sentence from the set of legally available dispositions. Thus Found guilty — (select) e e »* 0000 TFhe task ma be re-structured or sub-divided to form the foilowing model : Found — incarcerate — ) \ (select) How long (select) — not Guilty \ incarcerate — (select) What type ? Thus instead of the one task (marked «select») in the first approach, we have three tasks (marked « select ») which add up to the same outcome as the first. It is now possible for the decision-maker to consider more items of information in the total task, such as by using cight (different ?) items for each task (« select »). Ii will be necessary to return to the idea of decomposition and re-struciuring of decision tasks later in this article when we shall consider a form of operation which is at the same time more complex and of more practical utility. The utility relates to organizational pro- cedures as well as to analysis and theory in comparative criminology. The Quality of Decisions. AIl decisions in criminal justice are made in the presence of an audience. The audience may be present only in the mind of the decision-maker, but there is no doubt that there arz considerations of « an audience reaction » in all judicial decisions. In less picturesque terms, there is some measure of accountability demanded of all decision-makers in the criminal justice field — for some sreater and for some less — none can act alone and unobserved. The author oñce asked a very distinguished judge whether he had, in his opinion, ever made an incorrect judgement. It was agreed that this was a possibility and opened the way for the next question as to how he would know an incorrect decision. « If ninety percent of the persons in the public gallery thought it was a wrong decision, it probably was » the judge said. Another judge stated that he had never made — Îs6 — an incorrect decision — at least, not at the time he made it — later it might have been shown to be incorrect. In one case it Was the contemporary judgement of à democracy wich provided the test while in the other it was a test provided by future events or posterity. In any event it is always interesting to hear from decision- makers their views on their assessment of their decisions. It is not easy to say what qualities characterise an incorrect decision, but it is doubtless easier than stating all the necessary qualities of a correct one. Most decision-makers are concerned to make correct decisions, but they seldom or never obtain feed-back which distinguishes deci- sions which are « acceptable » from those which are of exceptionally fine quality. Decisions of parole boards, police, prosecutors and most officials in the criminal justice area tend to go unnoticed by the public while everything is going well. The determination of « well» is not always an unbiassed assessment. Nonetheless it may be that the failures of a process of administration are more informative as to its nature than are its successes. In some way or another the majority of officials and adminis- trators in criminal justice (and perhaps also most other governmental agencies) are concerned with the possibility of making a decision which, for some reason, is regarded as « wrong». Accusations of error may bc as much a problem to be avoided as actions believed to be genuine errors. JL is important to note that accusations of error made by others and self-recognized errors can be exacly similar or quite different kinds of error. Parole decisions serve as an example for decisions made at many other points in theo criminal justice system and have been the more widelv stutlizd than many others. Farole boards receive feed-back regarding their decisions wholly or mainly when these decisions are seen to be wrong. But the direction oË incorrectness is biassed — there is seidom much clamour that a board is detaining an inmate too long. There is little comment upon the decision of parole board except in terms of dramatic incidents (e.g. a paroled murderer commits another crime, or a notorious case is released). There is also a tendency for the critics to generalize from single incidents and to use these to attack the board’s policy. (Clearly, by definition, a rare event or dramatic incident 1s, by that very token, no usefu:! suide to policy.) The possibility of public criticism, inappropriately founded as it may be, is always in the minds of board members, even although suppressed. « If this case fouls up, what can I say 2» is one form that the reasoning may take. Another form, following upon an unusual incident, « How can we stop another one like that ? ». The words « like that » — 197 — refer to the ways in which the board was embarrassed by the presen- tation of information regarding the case — correctly or incorrectly coloured ! Clearly if parole is granted at all, there is always a chance of cases arising which present facts similar to those which gave rise to che embarrassment in the previous case. The majority of decisions in criminal justice have this kind of quality. À suspect may be permitted to go, and much later it is discovered that he was involved in the crime. The prosecution decides « not to paper » (proceed with) à case, only later to discover that had the investigation been pressed further serious offences could have been avoided. In view of the fact that information which may be used to discomfort the parole board, police chief or prosecutor is selected from one sector with a far greater probability than another, it is not unreasonnable for decision-makers to concentrate their defences in this area. Namely, they are concerned with the ways in which the facts may be represented against them. Anything which may be in their favour will require little or no attention. Thus it is good policy (as vulgarly, but preciselv put) « to cover your arse ». That is the nature of the system ; not the personal prejudices of the decision-makers within it. Decision-makers in different countries or in different regions within a country may expect different forms of attack. Which strategy of attack is chosen will depend upon assessments of the likelihood of !ts effeztiveness. The effectiveness of a strategy of attack will, in turn, depend upon the environment (political climate) in which it is used. For example, the New York Times may use one form while the San Francisco Chronicle may use another style. The style will also vary with the interests of any social action group involved. Criminal justice decision-makers, as we have noted, are sensitive to their environment. This is, perhaps, as it should be in a democracy. The survival of parole as a system depends upon its remaining acceptable to legis- latures and to the public. The press, together with other mass media, have an influence on both. Information Search and Decision Defences. It will be remembered that it was postulated that « preferences for strategies of information search» may be expected to be related to « the utilities sought by the decision-maker ». If the purpose changes, then the nature of the information search is expected to change. It is not suggested that criminal justice decision-makers select their method of information search solely because they are preparing for different forms of possible attack. Decision-makers in general and criminal justice is no exception, reflect in themselves something of the — 198 — qualities of the area in which they operate. States differ from each other. People who live in one environment are not the same as those living in another environment. Whether born in a particular state or not, people settled in an area are breathing that air, living in that information setting, reacting to that particular political climate. It may be that criminal justice decision-makers should reflect the essential elements of the political scenes in which they function. Perhaps even the same crime is not the same when it takes place in different environments ? As a judge once stated, « Cutting a wire fence in Vermont is different from cutting a wire fence in Colorado ». But it may be represented, people are the same — at least we are all expected to be equal under the law. How then can we say that there can be acceptable (ethically justified) local variations in the dispensation of justice ? It is possible to accept the postulate that all persons are « equal », but that the environment in which they function is not, and that justified differences may arise from that fact. It is possible to modify that statement a little and to formulate another possible principle, namely, there should be a consistent policy, but injustice may be done if all persons are pressed into the same policy mould. Is it possibe to reconcile the idea of a «consistent policy » with the requirements of justice in the decisions made about offenders ? Can the unique person of the offender be matched with a unique decision and yet that decision still fit a policy ? If we take the view that a basic element in justice is « equity», then this must be so. If like situations call for like decisions, this statement is itself a declaration of a policy. Policy and Case Elements in Criminal Justice Decisions. We have already discussed the principle of decomposition of a decision into sub-decisions. It is now claimed that a class of decisions found in criminal justice and certain other fields can be sub-divided by a standard procedure into a « policy» element and a «case» element. It is further claimed that it is possible to discover policy where it is present and to distinguish it from other factors. There are several benefits which accrue from this approach, both in terms of scientific inquiry and in practical operations of the system. It has commonly been found that decisions which individual decision-makers tend to claim are unique and unrelated to other decisions in fact show a definite but complex pattern. There is a tendency for decision-makers to do the « appropriate » thing, even where there are no rules laid down as to what is appropriate. This situation characterizes decisions where a group of persons (such as a parole board, members of the same district court bench, prosecutors — 199 — in the same office) make individual decisions and yet do so in an awareness of decisions made by their colleagues. Such decisions tend to be predictable to a greater or lesser degree. The element which is predictable is clearly not the unique case element due to the personal prejudices of the individual decision-maker. Whether recognized as such, or not, the predictable element must be implicit or explicit « poficy ». Ît is, then, possible to consider a decision in these situations as being composed of two or morce elements, one of which is the predic- table or policy element. As observed, the decisions have some similari- ties, but cach decision in nonetheless, unique. Decisions in this class can be considercd to be the summation of three elements, namely (a) the « policy » element, (b) the element due to the characteristics of the individual case which is subject to the decision, and (c) an element due to the idiosyncrasies of the particular decision-maker. The « policy » element (a) is the part of the decision which is predictable. It must be noted that this predictability is derived from the collective of decision-makers in the group (e.g. board, court, office). No reference is made to individual characteristics of the several decision-makers. This is because only the « policy » element is required to be specified. The objective of the analysis at this stage is not to explain the total complex of the decision outcome, but only the common elements or, to use a word we used earlier, the « similarities ». If it can be claimed that the prediction equation (or perhaps sequential network. or other model). fairlv represents the policy of the « collect- ive» of decision-makers in the specific unit (Le. board. court, office) then the task of decision-makers within this group can be re-struc- tured. Instead of approaching the information search task with the underlying question, « What should be done with this offender ? » the task is first to ascertain how the case would be dealt with if policy were to be followed « blindiy ». Since such decisions are the « predictable » element, this is a straightforward task. It requires the meticulous collection of the necessary information and the calculations to solve the equation. Once this has been done, the decision task is to assess whether there are any recognizable grounds for setting aside policy in the particular case under consideration. À search of the information available is carried out to this end as a separate task from that of the initial search to determine the policy element. An Application of the Decomposition Theory. The decomposition of decisions into policy and other elements linked with an information search strategy appropriate to this revised decision-making method has been developed into an operational system which is currently in use by the United States Parole Commis- sioners. The ways in which this was done and the background research are reported in a set of thirteen reports (4) and a short summary is provided in Carter R. and Wilkins L.T. : Probation, Parole anl Com:inuniy Corrections (5). 1u may serve here m LS w-j :.1rp‘w I to quote from a Memorandum of a decision by the Federal District Court in Connecticut. The court had heard a case which invoked the « guide- lines » system of decision-making adopted by the Parole Board (later, The Parole Commissioners). The court noted (6) : « These cases present interesting issues concerning the procedures used by the United States Board of Parole in reaching and explaining its decisions concerning parole. The issues arise because the Board, though not constitutionally required to give any reasons for its decisions (Menechino v. Oswald, 430 F 2d Cir 1970) has commendably adopted a new procedure designed to promote rationality in the decision-making process and to enhance under- standing of the process by all concerned… Key ingredients of the new procedure are (a) the use of a table of guidelines as an aid to deciding the appropriate length of time a prisoner should serve before being paroled…». The memorandum then goes on to describe the method whereby weights are giventotheseriousnessof the offence and to certain offender characteristics, and notes that these indicate ihe « expected » period of detention in prison. In the terms used in the description of the theory, the « guidelines » are a specification of the « policy element » in the parole decision. Although the guidelines in operation are prescriptive, (they indicate what should be done unless there are special reasons why they should be set aside in the individual case), they originated as descriptive equations. These descriptive equations, based on the seriousness of the offense and twelve offender characteristics (mainly prior criminal record) were found to predict a large percentage of decisions. Thus a policy element which had previously been implicit was made explicit by research and encapsulated into some simple equations. The equations are no more complex to read than the appropriate premium required for a life insurance policy and are of a somewhat similar form. An Expression of Cultural Values ? The « policy element » (guidelines) represent the previously unex- plained « vague similarities » between sets of unique individual deci- sions. If à « policy » is not made explicit, it 1s, perhaps, not correct to — 201 — refer to the discovered form as a « policy ». It might be more correct to describe the equations as an expression of the developed phenomenon of an approximate working concensus of colleagues faced with similar problems in the same general organization — a « culture of the office ». It is, of course, from this perspective that the analysis and equations may commend themselves to the interests of comparative criminologists. Social institutions develop a «style» by which they express their identity. Individuals within these social institutions both contribute towards this identity and blur its outlines by the exercise of their personal equations. There are two ways, at least, in which « cultural values» may be expressed in a latent policy and made visible by the development of descriptive equations which predict the decisions made by members of the organizations concerned. It is possible to attempt to fit a whole variety of equation forms as well as to derive sets of quite different weightings to data items within a particular form. The particular form and individual set of weights which were found to explain the greatest proportion of the variance in decision would, of course, be the preferred form. The failure of research studies to find an equation which predicts the decision outcomes seems to be the extremely unlikely. However,if the equations forms tried do not give a very good predictor, it may be that attempts have been made to fit an inappropriate form or alter- natively it could mean that individual variation was extremely large and the influence of « policy» was small. Similarly, different weights to individual items of data which might provide better or worse predictors do not necessarily reflect the weights given to these particular items by the decision-makers themselves. The prediction equation is fitted to the «collective» and not to any individual, and for this reason alone interpretation of the weights requires extreme caution. There are reasons for avoiding interpretation of the equations in other than operational terms. The same problems attend this kind of analysis as those involved in the fitting of econometric analyses to crime trends, criminal justice decisions or similar data. The data are drawn from an « information rich field ». Any matrix of correlations will present opportunities for a very large number of different sets of equations, each characterised by similar measures of efficiency. It is not possible to say which equation is the « best» (most powerful), because many could be found to satisfy this criterion. The use of any other criterion for the selection of one equation rather than another is attended by serious problems for which no satisfactory resolution seems to exist. Some persons prefer one, others another. Thus the selection becomes an exercice in subjective selection among mathematical forms. « 202 — If equations can be selected from a large number of apparently equally powerful expressions, how, it may be asked, can these be used to assess « policy» ? For many purposes the selection is not particularly important, any equation which appears to be as powerful as any other may be selected on other considerations (*). But this is not the issue in this case. If the prediction equations were derived to provide estimation oË the likelihood of recidivism (base expectancy), or were in the form of econometric models fitted to statistical data, the selection would cause greater difficulty. There is clearly no point in asking an offender whether he « feels » that his likelihood of recidivism is better predicted by equation (a) or equation (b), or whether it would be better to select the equation which gives a weight to (K) or to (L)! However, when the equations are used to attempt to discover « latent policy », we may recognize the fact that our subjects (the decision-makers) are somewhat self-aware in this process. Indeed they are, presumably, weighting data items themselves and the equations may be of the same form. We may capitalize upon this self-awareness. It was by this method that it was possible to test the decom- position of the sentencing decision noted earlier. It was discovered in the course of attempts to fit equations to sentences handed down by a court that the quality of the « fit» obtained if (a), the decision were assumed to be of the sub-divided format, (first decide whether to incarcerate or not, then, if incarcerate, decide for how long, and if not incarcerate, decide what form of community based disposition to select), was superior to (b), that obtained if we assumed that the decision was a straightforward selection from available dispositions. Could we rely upon the mathematical evidence for the use of the sub-divided format ? We could ask judges which form seemed to them best to represent their mental activities in the act of sentencing. The evidence of the judges asked was unanimous in favour of the sub-divided model as the more closely representing the process. Furthermore, we are not confined to evidence of this kind Once equations have been established, it is possible to carry out many tests by methods of stimulation. It is not possible to go into these methods at this time. The note of these methods is made only to establish the fact that the fitting of equations to decision outcomes from a collective of decision-makers is a procedure having several characteristics which distinguish it from the more usual derivation of prediction equations or curve fittings applied to other areas and other problems. (*) This is the basis of selection by the United States Parole Commissions only mathematically defensible, but philosophically «cleaner» than others. of a number of equations of equal power, one was selected which was not — 203 — An Example of Comparative Analysis. The long theoretical preamble to the presentation of a specific case of comparative criminology has by no means dealt with all the problems in this complex field. However, it is hoped that suificient points have been made to make the following example of an application of the method seem plausible. Soon after the United States Parole Commissioners adopted the « guideline » method discussed earlier, a number of states expressed an interest in investigating the possibility of their adopting a similar approach. The first step in such an approach is the development of equation which the decisions of the particular boards. he form of equation found to fit the decisions of the United States Board were found to fit also the decisions of some other state boards. But not all. Some states did not appear to give much weight to the seriousness of the offense, leaving this to be determined by the sentence of the court, while they gave most consideration to the behaviour of the offender while serving his term in prison. In one state, for example, an offender who had not been promoted to « honor grade» could not be further considered for parole. It was the task oË the research team to work out (in cooperation with the state boards concerned) an acceptable definition of their « policy » and to incorporate this into guidelines for future decisions. To do this the prior decisions of the boards provided samples of experience. The most useful and general approach to fitting an equation to such data either that of Multiple regression or Discrimi- nant fonction (which are sometimes identical). However, such methods presume a basic additivity in the items with the related idea of a « dimension ». Hence the idea of « seriousness » of the offense in the federal model could be regarded as a dimension increasing from trivial to the most serious with all crimes capable of being fitted in along this scale. Similarly, the oflender characteristics (« salient factor score ») was such that an offender might cancel out bad points by good points since it was the total score which determined the weight. In the Southern states this model did not hold as well as one which was more analogous to a steeplechase — there.are a series of « hurdles » to be overcome, and no matter at which one a fall occurs, the individual is out of the race. He cannot claim that he fell at the first, but had he been given a chance, he could have made all the remainder. He cannot say, for example, « If only I had been promoted to honor grade, I would have met all the other requirements of the board, so, since there 18 only one ticrm 24 which ! failod. F should te granted parole ». With the federal model, any one item in the salient 204 — factor score can be exchanged for another. so that all items are considered without regard to any sequence. There are other differences between the models which were found to fit the Southern state’s parole boards decisions and those of the federal system than the sequence element, although this is, perhaps, the most important difference. In the Southern state’s model there is a pattern of progression that ends with the specification of the decision as a dichotomy —- parole/not parole. In the federal model there 1s an inicrcept oï ‘wo dimensions which provides an entry in a cell which has a limited range of discretion for the individual decision- maker as to how long the offender should be detained. Within this range, the decision-maker is « within » the guidelines and, of course, he may decide outside this range, supporting his decision with reasons and subject to the other procedures for such cases. In the Southern model, the decision-maker can go against the indicated decision, but there is no range of discretion which is left unspecified. Not only do the models fitted to the prior decisions of Southern boards differ from those fitted to other states, but the consequences of fitting such different models also differ. The fitting makes clear, for example, tho ways in Which information items play their part in the outcome for the individual decided about. With the federal (matrix style) model, each item of information in the salient factor score contributes a small amount to the total and hence to the decision ; no single item will condemn the individual to any specific decision. With the « sequence » (network) model this does not apply. It is interesting to ask why a different style of model fits more closely the Southern states and why such a style of model commends itself to the decision- makers in that region. Às the research teams, point out in their report (7), « The transitton from descriptive to prescriptive is critical to our research and is a transition which we do not, ourselves, make. This must be the board’s decision in each state ». [n tho discussion of tho quality of decisions earlier in this article t Wais poted fhat “coisson-maters make their determinations before an audience, visible or invisible. The audience for decisions of parole boards in the Southern states is a different audience from that of the United States Board and the audience in other states. It is not surprising that the « received wisdom » of boards, courts and other collectives of decision-makers will reflect the culture of the local society. This we stated was to be expected. The empirical studies have shown that attempts to discover « latent » policy reveal such differences both in deta!! and in cencral structure of form. Do these differences make sense ” - 205 — A Tentative Explanation oË Differences. It was suggested that the ways which decision-makers selected to make their defense against accusations of incorrect decisions might reveal itself in the method of information search and decision strategies How might the United States Board seek to protect itself and from whom ? What is the constitution of this hidden critical audience for the federal Board ? Are they a different set of persons from the audience for parole decisions in Southern states ? It would seem so. The United States Board, located in Washington, has had plenty of experience of criticism and the audience is not much of a mystery to Board members ! (The New York Times and the Washington Post ; certain well-known law schools ; social action groups and political interests.) These critics, whether attacking from the left or right, use a form of argument which may be designated as « somewhat abstract ». The criticism is not likely, as a general rule, to deal with details of specific cases, but rather with principles. Are the decisions discrimina- tory ? Is there evidence of disparity ? How are the public protected from the dangerous offender ? And so on. These approaches are quite consistent with the « dimensional »> model which specifically relates to « seriousness » of offenses. The model and its adaptation into guidelines, addresses most directly the issue of disparity. If a dramatic incident focusses criticism, that criticism is still dealt with in general terms. The size of the audience inhibits a concern for detail — detail can be subsumed in the appropriate action in terms of principles. This seems to characterise the federal (and northern) model. Rather than detail being subsumed in the action to accommodate principles, the model identified with the Southern states, sees the prin- ciples in relation to the detail. This particular focus and model may well apply to other areas and probably to other countries where the social and political climate are similar. In the « sequence » model a more ready defense is available for the kind of criticism to be expected from the invisible audience found in that environment. Here offenders are somewhat less anonymous, and there is an interest in the detail of their activities. Decision-makers too are less anonymous and there is an interest in the detail of their decisions. It is not surprising that the models which fit are a reflection of the strategy likely to be more effective as a defense of the decision in the particular areas. The « sequence» model enables the defense to concentrate upon the individual items of information in respect of any «foul up». The « matrix » model enables concentration upon the correctness of a general policy at the expense of individual detail. The decision-making strategy has developed naturally over the years as a « home grown » product suited to the climate. There would seem to — 206 — be no reason why this should not be so — perhaps justice should have a local flavour ? The explanation of the different models (sequential and matrix) may be clarified by imagining the form of accusation in case a member of the invisible audience wished to suggest that a decision was in error. « Did this person, or did he not….» possess charac- teristic (x) ? « And did he not… ? » The specific questions relate, in this example, to discrete items of data. The imaginary defendant (the Parole Board or other agency) would be inferred to have erred if it had to answer questions formed in this way with a response showing that its guidelines were not followed. « Had this person made honor grade in the institution ; yes or no ? » If the board had to say «no » then it would be assumed to be in error in releasing the offender and hence responsible for the fact that he had committed another crime. The lack of logic (the non-sequitur) would not trouble many of the accusers ! The boards with the sequential model are ready to meet this kind of attack. Each decision is made in terms of the branching network of specific items ; not by adding together items to form some representation of a concept. There is some possibility, suggested by other studies of decision-making, that « sequential» models tend to represent a more risk-aversive decision process than do « matrix» models (8). It seems that risk aversiveness is increased due to the fact that at each node (branch) a subsidiary decision is involved. À node becomes an all-or-none determination, based upon one item of inform- ation rather than being decided conditionally upon several items. ÀA «safe » decision tends to be made at each branch, with an overall result that fewer persons get through the network than would be the case if alternative routes were available to them. This observation fits the general explanation of the differences observed in studies to date. Conclusions. The possibility of developing the methods discussed into a general tool of comparative analysis might be explored. In the present case, it was fortuitous that the data were to arise in a suitable form. The use of models developed with a different purpose presents different problems from those which are specifically constructed. Models which were used to test theories in comparative criminology would require more than a simple replication of the methods used to develop decision guidelines. If, for example, instead of attempting to discover the model implicitly used by parole boards, the research had established some a priori model or models, it is likely that some degree of fit would have been found to all states and to the federal system. This fit might have been nothing more than a meaningless centroid. — 207 — Despite the difficulties, it seems clear that the kinds of models which are necessary to give the best prediction of the decisions made by agencies reveal the dominant policy issues. Furthermore, while the weights to items in a multivariate analysis cannot be understood as having a direct relationship to the interpretation of policy, differences between weights should suggest questions. Such questions may be developed into hypotheses to be tested by other means, such as simulation or further 1ndependent 1nvest1gatmn ercte ‘It.is clear that decisions made 1ndependenfly by individual decision-makers in positions of authonty in the criminal justice system with regard to individual offenders in unique situations, represent a complex pattern which may be « modelled » in a variety of ways. In most cases.this pattern .may be equated:tosthe idea of an implicit policy or Lra sét of etiturat valués shared by members of the organ- IZation concerned. Although the decision rules implied in a policy may be varied by the decision-maker in respect of individual cases dealtwith by him, the modifications çan be, regarded.as added elements Aitiodifiersÿ, -Suctr-recomposition of tkie'decision ‘prôcess into policy and case elements (the central focus of this article) would seem to be valuable in comparative research as well as realistic and useful for purposes of policy control and in facilitating public accountability REFERENCES Ï _ HAYVES ( M0) Häman Da:a Processing Limits in Decision- malung Opera- é tiohal“Applications" Lab "HH‘ Forté 'Bléctronics,. Systems’ Divigon, Bedford; Mass*“ 1962. 2 l BURNHAM (R.W.) in Decision-making in the Criminal Justice System {\< I;1 i\‘30\> National Instituté of Mental Health. Monograph D H E.W. Publication - (A.D.M.) 1975, pp. 75-238. WILKINS (LT. « Information Overload » Journal of Criminal Law and Criminology, 1973, 64:2, pp. 190-197 VARIOUS arole Dcecision-making Project Reports Davis california N.C.C.. RKoscarch Center. 1973. (Also Law Enforcement Assistance Ann.n 15 1ml.on 3l L)°p\rtmcnt of _Justice.). - S Pr8baîmv Parole and Community CARTER (R.) dfld WILKIN9 (L.T.) : Corrections, (Qnd Edition). New York. Wilev. 1976. Lupo vs. Norton and Zagarino vs. Attorney General of the United States Federal District (ed.) Connecticut, 1970, (430) GOTTFREDSON î)) Ce (ll « Diffcrent Faces of Truth ? » in Report of Research Project Classification for Parole Decision-making in press To be published by L.E.A.A. shortly 8 ROBINSON (J) and TAKAGI (P.T.) : «The Parole Violator as an Re; ‘est »> -in CARTFR (R‘) and WILKINS (L.T.) see (S) ** “Organtzationa! above — 208 — THE COMPARATIVE STUDY OF PENAL SEVERITY e SOME METHODOLOGICAL CONSIDERATIONS* PETER N. GRABOSKY Yale Law School New Haven, Connecticut - US.A I. Introduction One explanation for the dearth of comparative studies of penal severity is the unhappy reality that such endeavors inevitably encounter methodological problems of a forbidding nature. While some problems of conceptualization and measurement have proven to be insurmount- able, thus precluding the testing of theory in a definitive, universal ense (accross all jurisdictions and for all time), such impediments should not deter prospective researchers altogether. While a few problems will remain intractable, others can be mitigated or circum- vented through a judicious choice of measurement strategy and research design. What does exist for the comparative researcher is the inviting opportunity to formulate general propositions, then to test them in a succession of different, albeit limited settings (Zetterberg, 1966 pp. 126-8). For example, while it may lie beyond the capability of social scientists to describe, much less explain variations in penal severity across the provinces of China in 1906, it may be possible to test the applicability of Sorokin’s (1937) theory in the context of contemporary European industrial societies. While no single study can establish the universal truth of a given theory of penal severity, the extent to which corroborative findings recur in subsequent and different settings is indicative of the theory’s scientific worth Fhe pages which follow are intended to alert prospective research ers of penal severity to some of the hazards which they might expect to face, and to acquaint them with a few of the basic remedies at their disposal. The discussion will focus first on some of the more fundamental considerations of conceptualization and measurement emphasizing the issues of comparability and equivalence. The essay will conclude with a discussion of alternative research designs which can serve to mitigate the hazards in question *) ’TThe author wishes to acknowledge the support of the Russel Sage Foundation, and the he]pful advice of colleagues in the preparation of this essay — 209 I[. The Concept of Penal Severity. The task of delineating the conceptual boundaries of « penal severity » is a difficult one, for the degree of suffering inflicted on an individual by authorities in response to the commission of an officially proscribed act is essentially subjective. Hardly a modern insight, this recognition of the subjectivity of penal severity appears, for example, in Grotius’ contention that a fine weighs more heavily on the poor man than upon his more affluent counterpart (1853, II[: 20 : 33). Indeed, the notion of differential propensities to suffering seems implicit in the elaborately graded systems of punishments according to social status that were implemented in many ancient societies (Sellin, 1976, Chs. 1-3). The most sophisticated discussions of the inherent subjectivity of penal severity, those of Nils Christie (1968, 1975), assert that penal severity takes on meaning only when experienced against life at large. Thus, the severity of a day’s deprivation of liberty has increased in most societies over the past century, and the experience of incarceration «is more grievous in a society in which life at large is easy and enjoyable than in one in which it is difficult and dangerous ». (1975, 293). This notion of social or experimential time as opposed to physical time in conjunction with deprivations of liberty has also been noted by Sliwowski (1974). To insist that penal severity can only be studied from the pers- pective of the recipient of punishment, however, is too confining a posture. It would preclude the possibility of any historical studies and demand that any future comparative research be based on equivalent measures of individual attitudes. Contemporary researchers have thus adopted the more flexible alternative of focusing upon the objective properties of penal sanctions, presuming them to be comparable under certain conditions. What Gibbs (1975) terms presumptive severity thus appears to be the only viable alternative for those who wish to undertake comparative research on any significant scale. Indeed, this appears to be the only practical alternative for those who wish to regard penal severity as a property or characteristic of political systems rather than a subjective state of mind. Another conceptual problem encountered in the comparative study of penal severity concerns the relative salience of penal sanctions within a society’s overall system of social control (Newman, 1976, 27). Penal sanctions constitute but one instrument of social control, and it has been suggested by Black (1976) that they will be less important to the extent that control functions are shared with alter- native institutions. In the People’s Republic of China, for example, a considerable part of the sanctioning function is performed informall by family, neighborhood, and workplace groups (Cohen, 1968). — 210 — In western industrial societies, the sanctioning function may also be shared by penal and non-penal administrative institutions (Jasinski, 1976). Feeley (1976) suggests that taxes, subsidies, licenses, and grants are often used as calculated instruments of social control. In Sweden, there has long been a visible emphasis on administrative regulation in lieu of the criminal sanction, with regard to juvenile misbehavior and alcohol abuse (Kinberg, 1933 ; Bengtsson, 1938 ; Gôransson, 1939). Even medical treatment exists as an alternative to punishment in some societies (Kittrie, 1971 ; Goldstein, 1967, Ch. 11), and there exist some tentative research findings to suggest that rates of imprisonment vary inversely with the availability of mental hospital facilities (Biles and Mulligan, 1973). There are, quite simply, very different ways of responding to deviant behavior. By focusing exclusively on penal sanctions, it is possible to over- look the subtle interrelationship between these and non-criminal or informal sanctioning mechanisms within a society. Accordingly, inter- societal comparisons are doubly jeopardized. These problems of con- ceptualization raise commensurate difficulties in measurement. IIL. The Measurement of Penal Severity. Before venturing further, it is important to confront directly one of the fundamental problems of comparative research — the issue of equivalence in measurement. The earlier discussion of the inherent subjectivity of punishment, and the relationship between penal and alternative sanctioning systems reflect a problem common to all comparative researchers and which is particularly threatening to research on penal systems. The issue, briefly stated, is this : Is the phenomenon being investigated in society À essentially comparable to the phenomenon under scrutiny in society B ? Is the measurement language to be applied equivalent across the two systems ? If the rate of imprisonment is twice as high in the latter, this does not necessarily imply that society B is twice as punitive. Only 1f the overall structures of social control in the two societies are themselves comparable may such an inference be made (Smelser, 1976, pp. 185-93). It follows that explanations based on incomparable phenomena should be viewed with utmost caution. When different meanings are couched in facially equivalent measures, the researcher must determine whether it is possible to modify one or the other of the measures so as to render them at least crudely equivalent. Given the present state of social science, researchers will have to settle for what might best be termed « gross comparability ». One of the more important tasks for the future is the development of cross-systematically equivalent measures. — 211 — The task of measuring penal severity is a difficult one, as has been apparent to scholars since the late nineteenth century (U.S. Department of the Interior, 1896 ; Sorokin, 1937 ; Gibbs, 1975). The problem is essentially twofold — data are available for only a limited number of jurisdictions, and those data which do exist may not be comparable across jurisdictions (1). The problems posed by the scarcity and questionable validity of data are best confronted through a strategy of multiple measurement. Given the inevitable imperfection of any one measure of penal severity, it is wise to rely upon a number of alternative measures. The justifi- cation for such a multiple measurement approach rests in the assump- tion that each of the measures chosen will be subject to different sources of inaccuracy, while still sharing elements of conceptual relevance (Campbell and Fiske, 1959). From an evaluation of alter- native measures, the researcher can move to select those which appear most appropriate to the task at hand, or to construct a composite index (Sullivan, 1974 ; Gurr, 1972. pp. 49-59). Any analysis of penal policy must take into account the fact that formal sanctioning is imposed at a number of stages or dispositional points. Initially, the severity of response appropriate to a given category of deviant acts is specified by statute. Following the initial invocation of the criminal sanction, deprivations may be inflicted in the course of investigation or processing prior to adjudication. The most obvious example of such a form of sanction is pre-trial detention. Subsequent to adjudication, severity may be measured in terms of the sentence awarded, or ultimately, in terms of the punishment actually imposed. It is by no means certain that measures of severity taken at successive points in the dispositional process will show consistent covariation. Indeed, there may well be a significant degree of « slip- page » between statutory prescription and severity as applied, as well as between the various stages of application (Newman, 1976, p. 18). Ross (1976) for example, has shown how relatively harsh provisions are «neutralized » by lenient enforcement and administration. The magnitude of the gap between symbol and action is itself deserving of analysis in a comparative context. The measure of severity as reflected in a jurisdiction’s statutes (hereafter normative severity) is based upon the premise that the penalties prescribed in a society's penal statutes are generally indi- cative of its punitive predisposition. While scholars of the criminal (1) A number of polities, for domestic and/or international political consi- derations, conceal some or all of their sanctioning activity. — 212 — law suggest that individual provisions are generally initiated and championed by parochial interests (Heinz e/ al, 1969, Steyn, 1975) it seems reasonable to conclude that an aggregate or summèar”y measure of normative severity will constitute a more valid represent- ation of the concept under study. Indices of normative severity can be constructed from statutes by selecting a set of offenses in accordance with one fundamental criterion : the behavior defined by each offense should be comparable across units of analysis (Newman, 1976, p. 214). This is important, of course, since the terms used to label a given criminal act can vary across jurisdictions and over time. In seeking to compare the severity with which authorities purport to respond to what is conventionally termed armed robbery, for example, it will be necessary to ensure that the penalties in question are not attached to quite different pheno- mena — robbery with actual violence in one jurisdiction and robbery with the mere threat of violence in another. Ideally, the set of offenses chosen for analysis will be large enough to permit its division into subsets, each representative of a conventional index category. This division into subsets will permit the construction of a summary measure of normative severity for each jurisdiction under investigation, as well as permit comparisons within and between jurisdictions in the relative severity with which different offenses are ostensibly punished. The actual measurement of statutes can be accomplished in a number of ways. In dealing with similar legal cultures in which impri- sonments and fines constitute the most common sanctions, one may use the duration of imprisonment (in years or fractions there of) prescribed for each offense (Newman, 1976, pp. 228-246). To accom- modate those statutes which mandate indefinite sentences or a range of possible punishments for each offense, two measures might be generated for each offense, one each for the maximum and minimum possible penalties prescribed (2). For those statutes which prescribe a fine in addition to or as an alternative to imprisonment, the amount prescribed can be expressed as a percentage of mean annual income in the jurisdiction, and equated to an appropriate period of deprivation of Hiberty. Some penal codes actually provide a conversion scale, thus obviating the necessity for such judgmental translation. Capital punish- ment and life imprisonment may be assigned arbitrary values at the high end of the scale. (2) Where no upper or lower limits are stipulated by statute, the researcher may apply judgmental standards or solicit expert opinion. See Newman (1976, P ) — 213 — For the more ambitious researcher who seeks to undertake com- parative analyses across widely differing cultures and time periods, a standardized measurement scheme was devised by Sorokin (1937, I, p. 582). The scheme consists of a ten point scale along which a range of ostensibly commensurate punishments are arrayed. Aggravated capital punishment is accorded a weight of ten, a simple fine a weight of one. Gravely mutilating bodily punishment is equated with life imprison- ment, and both are coded with a score of eight. While such a measure- ment strategy may be appropriate for broad comparisons of otherwise incomparable societies, it is of less utility for the analysis of modern penal systems. In such cases, the development of more precise measures 1s preferable. The degree to which authorities rely on procedural punishment, deprivations attending the period before and during adjudication, poses a particular problem, for such reliance varies widely across juris- dictions, and systematic assessment is most difficult. These dispositional considerations may not be wished away, however, since in a number of jurisdictions, it is customary to consider predisposition periods in determining the sentence actually awarded. Thus, high dismissal rates and frequent commutation of sentence to time served do not necessarily imply that no punishment at all is being meted out. The measurement of pre-adjudicative severity should incorporate two dis- tinct properties the proportion of defendants detained prior to trial, and the average duration of pre-trial detention. The severity of sentences actually awarded may be expressed in a number of alternative ways. Where data are available, the researcher could calculate an appropriate summary measure such as the mean or median sentence imposed. Alternatively, it is often possible to represent sentencing severity in terms of the proportion of sentences surpassing certain thresholds of severity. Data reflecting the percentage of defendants sentenced to sanctions involving deprivation of liberty, or the percentage of sentences involving incarceration in excess of one year, have been employed to this end (Jasinski, 1976). In addition to those techniques, the researcher may employ a variation on the method employed by Kobrin and his colleagues (1972) in their comparative study of California counties. Their additive index of sentencing severity was constructed by assigning one of six weights to each sentence awarded, summing the weighted scores, and deviding by the total number of sentence : Severity = {(0 x fine only) --- (! x probation onty) (2 jail and probat'on’ (3 x jail only) -+- (4 x state prison) (5 x deoth e_ænnl?y \ (S x total sunarier cour!s convictions) — 214 — In undertaking to measure the severity of punishments ultimately imposed, it will be important to distinguish between two distinct properties — the intensity of sanctions applied, and the scope of their application. The most common measure of severity thus far employed has been the median number of months served by inmates of a juris- diction’s prisons at year’s end (Gibbs, 1975, p. 129). This measure is insensitive to the property of scope, as it reveals nothing about the proportion of a jurisdiction*s population which is subjected to penal sanctions. Alternative measures of severity, based on the rate of imprisonment, fail to reflect the intensity of punishments inflicted. To further illustrate the distinction, it is helpful to envision two societies, one which imposes short prison sentences on a relatively large proportion of its population, and another which imposes long prison terms on only a small percentage of its citizens. The extent to which the scope of punishment varies with intensity is an interesting question in its own right, but a measure of actual severity should be sensitive to both of these properties. The process of developing and refining measures of actual severity should test different additive and multiplicative combinations of scope and intensity to arrive at a general measure of the magnitude of punishment delivered by a society. The contention that penal severity can only be measured in light of the quality of life in society at large suggest one measurement strategy which can significantly enhance equivalence of indicators across jurisdictions as well as within a jurisdiction over time. The simple adjustment to be described may be applied to any of the measures of severity reviewed above. The use of residual scores from a simple two variable linear regression is a relatively uncomplicated technique described by T'ufte (1974, pp. 24-29). When a quality of life measure such as life expect- ancy is regressed on a measure of penal severity, the residual scores reflect the degree to which that measure of severity is more or less severe than would be predicted by social conditions at large. The residual score thus represents what might be described as an « adjus- ted» measure of severity, one which accounts for the relative well being or misery of a societys general popuation. This adjus*meut may be made for differences between societies in the general level of well being ; it may also be used to control for improvement, deterio- ration, or fluctuation over time in the quality of life within a given jurisdiction. Such an adjustment may be made to measures of severity at any of the above procedural stages ; it may also incorporate other variables thought to detract from measurement equivalence. — 215 — IV. Research Designs. The formulation of a research design, or rationale by which cases are selected for analysis, is one of the more crucial issues confronting comparative researchers, for it is this choice which determines two fundamental considerations explanatory potential and generaliza bility. To compare present day penal policies in North Dakota and South Dakota, for example, would be a less than fruitful undertaking for two reasons. First, there would be little variation between the states either in terms of the policies themselves or in terms of the social and historical characteristics which might influence these policies. Second any findings which might emerge from a comparative study of the Dakotas could not be generalized to other jurisdictions or other time periods The task faced by researchers of penal severity, then, given the inevitable constraints of resource and data limitations, is to select those jurisdictions for study which, by virtue of their distribution across time and space, provide for greatest explanatory power and generalizability Given the serious difficulties in conceptualization and measurement which confront the student of comparative penal severity, the choice of an appropriate research design becomes one of great significance It should be apparent by now that the rich diversity of penal systems over time and across space preclude any single test which might yield findings of universal applicability. Reality dictates what might be called a modular approach — the testing of propositions within com- parable groups of cases, systems whose operational indicators have equivalent meaning (Smelser, 1976, 169-70) It can be argued that a wide variety of distinct research settings provide the optimal context for theory building. Consistent findings emerging from a range of diverse settings each with its own measure- ment strategy and research design appropriate to the set of systems under analysis are of much greater knowledge value than ambiguous results emerging from a «loose fitting » macro-study of every juris- diction for which data might be available (Stinchcombe, 1968, pp. 18- 22). What is 1mportant is that the limited contexts which serve as the objects of inquiry be chosen with some care. Within these general guidelines, it is appropriate to review some of the more widely employed design stategies and alternative analytic frameworks, and to suggest the utility and limitation of their appli- cation to the comparative study of penal severity. Ideally, this discussion Will provide prospective scholars with guidelines for choosing the approach which seems most useful for the purpose at hand — 216 — 1. Basic Design Strategies. One of the more popular alternatives available to the comparative researcher is what Lijphart (1975) terms « the comparable cases stra- tegy ». By selecting fundamentally similar systems for analysis, albeit systems which vary significantly in terms of independent and dependent variables, the researcher is able to achieve a considerable amount of control for extraneous or irrelevant influences. This general approach has also been termed the « most similar systems design » by Przeworski and Teune (1970). In addition to the advantages of control, the similar systems approach provides the additional advantage of minimizing problems of equivalence in measurement. Sanctioning systems and accounting practices vary much less markedly across the political systems of Western Europe than they do across a random sample of the world’s polities. It is important to note, however, that the systems selected for analysis must have sufficient variation on theoretically relevant variables (penal severity and its hypothesized determinants). Without such variation, the design lacks explanatory potential, as did the above example of the Dakotas. In contrast to the advantages provided by the most similar systems design, there are some questions which can be approached fruitfully through comparison of relationships visible in drastically different penal systems. While problems of comparability would pre- clude an explicit comparison of the magnitude of punishment delivered in Sweden, Uganda, Equador, and Singapore, consistent patterns visible within each of these societies would be of considerable scientific interest. If, despite the myriad of political, legal, and cultural characteristics which differentiate these four societies, short term economic contrac- tions are regularly followed by a decrease in penal severity, a resear- cher can infer that the relationship is unlikely to have resulted from some systemic singularity (3). This « most different systems design » (Przeworski and Teune, 1970, p. 35) should be regarded as one of the more fruitful approaches to the development of general theorv. 2. Specific Analytic Frameworks. One of the more common designs in comparative research is the cross-sectional or synchronic study. Common procedure in this type of analysis would involve the selection of a set of penal systems and an attempt to explain the relationship at a given point in time between penal severity and contemporaneous social and political characteristics. (3) Note, for example, Newman’s (1976) study of individual attitudes in six cultures. — 217 — The limitations of cross-sectional analysis, however, are severe. Since the generalizability of one’s findings is determined by the repre- Isentativeness and size of the sample under observation, equivalence is easily jeopardized. A broadly representative sample of the world’s penal systems would be characterized by incomparable sanctioning processes and non-equivalent indicators. As far as research on penal severity is concerned, therefore, cross-sectional analysis would best be confined to the most similar systems context. One might employ such a design, for example, to explore the relationship between penal seve- rity and income concentration in the nations of Western Europe during 1975. Cross-sectional analyses admit of another, equally severe short- coming, however. While a considerable part of the mission of compa- rative research is to explain differences between systems at a given point in time, it is fundamental importance to note that many pheno- mena which researchers seek to explain are longitudinal in nature (Le. occurring over time). Much of the theory of penal severity (Durkheim, 1969, Sorokin, 1937) is developmental in nature, and the logic of causal explanation requires a longitudinal perspective. To infer developmental relationships from cross-sectional contemporaneous covariation necessitates drastic assumptions, and these may not always be tenable. Political systems may well arrive at their respective states of development by means of very different causal processes, and these alternative paths to the present can be concealed by a synchronic analysis. It is, of course, possible to modify a cross-sectional design to permit the incorporation of some dynamic variables. The use of the simple arithmetic mean, for example, would permit the introduction of such potential independent variables as the average annual rate of population increase for a preceeding span of years. Similarly, one may introduce a measure encompassing the average annual rate of impri- sonment during the same time span. Relationships between such summary measures or « gain scores» (Bohrntedt, 1969) no doubt constitute an improvement over those based on simple static measures. They nevertheless lack the manipulability and potential explanatory power of observations taken over time. One analytic strategy which permits a more direct observation of change over time is time series analysis (Blumstein et al, 1976; Hibbs, 1974a). While the taking of observations at equal intervals over time raises the problem of historical equivalence suggested by Christie (1968), the adjustment technique discussed above provides a partial solution. However, even if a measurement strategy were to compensate satisfactorily for an initial lack of equivalence, there are significant problems to be confronted. Time series analyses pose — 218 — inherent problems of generalizability ; it is uncommon to find a study based on more than a handful of cases. Assuming one were to discern a negative relationship between short term economic fluc- tuation and penal severity in France from 1945 to 1975 one may not infer that such a relationship obtains anywhere else, or indeed that it would obtain in France during another period. Moreover, there are a number of variables which may be expected to explain differences in severity which are more or less invariant over time within a society and which cannot be accommodated to time series analysis. Statutory severity itself may change only very infrequently, and may thus not be picked up in a span of years under observation. Racial homogeneity, which may well have signi- ficant relationship to penal severity in the United States, has been virtually constant in Sweden. Despite the disadvantages, time series data may be fruitfully employed to assess the impact of a sudden intervention. Short time series have been artfully analysed to assess the effect of war on crime, dissidence, and official reaction (Archer and Gartner, 1976, Stohl, 1976), while longer time series lend themselves to more powerful multivariate techniques (Hibbs, 1974b) (4). In some research contexts, when observations over time are available for more than one case, it is possible to pool cross-sectional and time series data (Nerlove, 1971). This approach has become increa- singly popular with social scientists (Hibbs, 1976). While strictly speaking a technique of data analysis, the pooling of observations can bear on design strategy by influencing the choice of jurisdictions and the period of their observation. By treating each nation year as a case, it is possible to mitigate those effects of data scarcity which would ordinarily inhibit gene- ralization. Assuming ten annual obervations for each of five jurisdic- tions, one can undertake analysis with a sample size of 50, and thereby apply more sophisticated statistical techniques. While the pooling of time series and cross sectional data permits a researcher to share the advantages of longitudinal and cross-sec- tional designs, it admits of one essential shortcoming. The structure of causal relationships themselves may differ from time to time and from place to place, and such differences may be masked by an analysis of pooled observations. This is not to suggest that penal severity may defy explanation altogether, but merely that the structure of that explanation itself may change. Penal policy might be sensitive to economic contraction during a certain period, and yet be profoundly influenced bv demographic changes a year later. (4) For a helpful introduction to the uses and limitations of quasi-experimental designs, see Caporaso and Ross (1973). For a more advanced perspective, see Glass, et al. (1975). — 219 — One design which accomodates this need for flexibility is the panel study (Pelz and Andrews, 1964, Shingles, 1976). By observing the systems under investigation at each of two or three time points, one is able to undertake temporally distinct cross-sectional analyses as well as to explore relationships over time. Such an approach would permit the researcher to discern whether a relationship between the unemployment rate and penal severity across western European polities in 1970 was also apparent during 1960 and 1950. While less flexible than the incorporation of lagged terms in the time series context, the researcher may produce lagged correlations from panel data, and thus determine the extent to which economic conditions in 1960 explain penal severity a decade hence. This capabilitv is quite consistent with the logic of causal inference. Although the panel design seems to provide the most advantageous posture for an initial approach to the question at hand, it should by now be apparent that no one measurement technique or research design strategy can hold the key to a general theory of penal severity. Indeed, it has been argued that the greatest knowledge gains will result from a richly diverse application of measurements and methods. What is important, however, is that the technique fit the setting, and that the researcher be cognizant of the inherent limitations of a given undertaking. With these considerations in mind, researchers should find the comparative study of penal systems an inviting one. Bibliography ARCHER (D.) and GARTNER (R.) : « Violent Acts and Violent Times : À (‘’oOmparative Approach to Post-war Homicide Rates », American Socio- logical Review, December 1976, Vol. 41, pp. 937-963. BENGTSSON (H.) : « The Temperance Movement and Temperance Legislation in Sweden ». The Annals of the American Academy of Political and Social Science, 1938. Vol. 197, pp. 138-153. BERK (R.A.) : The Measure of Justice, Academic Press, New York, 1977. BILES (D.) and MULLIGAN (G.) : « Mad or Bad ? The Enduring Dilemma », British Journal of Criminology, 1973, Vol. 13, n° 3, pp. 275-79. BLACK (D.J.) : l'he Behavior of Law, Academic Press, New York, 1976. BLUMSTEIN (A.), COHEN (J.) and NAGIN (D. «The Dynamics of a Homeostatic Punishment Process », Journal of Criminal Law and Crimi- nology, September 1976, Vol. 67, n° 3, pp. 317-334. BOHRNSTEDT (G.W.) « Observations on the Measurement of Change », in E.F. Borgatta and G.W. Bohrnstedt (eds.) Sociological Methodology 1969, Jossey-Bass, San Francisco, 1969, pp. 113-136. — 220 — CAMPBELL (D.T.) and FISKE (E.M.) : « Convergent and Discriminant Vali- dation by the Multitrait-Multimethod Matrix», Psychological Bulletin, 1959, Vol. 56, pp. 81-105. CAPORASO (J.A.) and ROOS (LJ. jr.) : Quasi Experimental Approaches Testing Theory and Evaluating Policy, Northwestern University Press, Evanston, 1973. CHRISTIE (N.) : « Changes in Penal Values », Scandinavian Studies in Crimino- logy, 1968, Vol. 2, pp. 161-172. « Utility and Social Value in Decisions on Punishment », in Roger Hood, (ed.) Crime, Criminology and Public Policy, Free Press, New York, 1975. COHEN (J.A.) : The Criminal Process in the People’s Repuhlic of China, 1949- 1963, Harvard University Press, Cambridge, 1968. DURKHEIM (E) « Two Laws of Penal Evolution », (William Jeffrey, Jr. tr.) University of Cincinnati Law Review, 1969, Vol. 28, p. 32. FEELEY (M.) : « The Concept of Laws in Social Science : À Critique and Notes on ain Expanded View». Lauw and Society Review, Summer 1976. Vol. 10, n° 4, pp. 497-524. GIBBS (J.P.) : Crime, Punishment, and Deterrence, Elsevier, New York, 1975. GLASS (G.V.) et al. : Design and Analysis of Time Scries Experiments, Colorado Associated Universities Press, Boulder, 1975. GOLDS'I:EIN (A.S.) : The Insanity Defense, Yale University Press, New Haven, GORANSSON (H.) « Treatment of Criminals and Other Asocial Individuals ». The Annals of the American Academy of Political and Social Science, 1939, Vol. 197. GROTIUS (H.) : De Jure Belli et Pacis, Cambridge Universitv Press, Cambridge, 1853. GURR (F.R.) : Politimetrics : An introduction to Quantitative Macropolitics, Prentice Hall, Englewood Cliffs, 1972. HEINZ (J.)) GETTLEMAN (R.) and SEESKIN (M.) : « Legislative Politics and the Criminal Law», Northwestern University Law Review, 1969, Voi. 64, pp. 277-358. HIBBS (D.A.) : « Problems of Statistical Estimation and Causal Inference in Time Series Regression Models », in H. Costner (ed.) Sociological Metho- dologv 7973-1974, Jossev-Bass, San Fransisco, 1974. pp. 252-289. On Analysing the Effects of Policy Interventions : Box-Jenkins and Box-Tiao vs. Structural Equation Models, Center for International Studies, MI.T Publication C/74-33, Cambridge, 1974. « Industrial Conflict in Advanced Industrial Societies», The American Political Science Review, December 1976, Vol. 70, n° 4, pp. 1033-1058. JASINSKI (J.) : « The Punitiveness of Criminal Justice Systems (AÀ Cross- National Perspective) ». The Polish Sociologica! Bulletin, 1976. n | pp. 43-51. KINBERG (O.) : « Criminal Policv in Sweden During the Past Fifty Years ». Journal of Criminal Law and Criminology, 1933, Vol. 24, pp. 314-335. KITTRIE (N.N)) : The Right to Be Different : Deviance and Enforced Therapy, The Johns Hopkins University Press, Baltimore, 1971. KOBRIN (S.) et al. : The Deterrent Effectiveness of Criminal Justice Sanction Strategies, U.S. Department of Justicee Law Enforcement Assistance Administration, Washington, 1972. LIJPHART (A.) : « The Comparable Cases Strategv in Comparative Research ». Comparative Political studies, July 197S, Vol. 8, n° 2, pp. 158-177. NERLOVE (M.) : « Further Evidence on the Estimation of Dynamic Economic Relations from a Time Series of Cross Sections ». Econometrica, March 1971, Vol. 39, pp. 359-382. NEWMAN (G.) : Comparative Deviance : Perception and Law in Six Cultures, Elsevier, New York, 1976. PELZ (D.) and ANDREWS (F.) : « Detecting Causal Priorities in Panel Data », American Sociological Review, 1964, Vol. 29, pp. 836-848. PRZEWORSKI (A.) and TEUNE (H The Logic of Comparative Social Inquiry, John Wiley, New York, 1970. ROSS (H.L.) « The Noetitralization of Severe Penalties : Some Trafïic Law studies », Law and Society Review, Spring 1976, Vol. 10, n° 3, pp. 403-413. SHINGLES (R.) : « Causal Inference in Cross Lagged Panel Analysis », Political Methodology, 1976, Vol. 3, pp. 95-133. SLIWOWSKI (G.) : « La Sociologie du temps et la peine privative de liberté », 7 Revue de science criminclle et de droit pénal comparé, 1974. n° pp. 295-312. SMELSER (N.J.) : Comparative Research in *he Social Sciences, Prentice Hall. Englewood Cliffs, 1976. SOROKIN (P.) : « Fluctuation of Ethicojuridical Mentality in Criminal Law », in Sociocultural Dynamics, American Book Co… 1937. Vol. 2. pp. 523-632. STEYN (J.H.) : « The Punishment Scene in South Africa », in Roger Hood (ed.), Crime, Criminology, and Public Policy, Free Press, New York, 1975. STINCHCOMBE (A.L.) : Constructing Social Theories, Harcourt, Brace and World, New York, 1968. STOHL (M.) : War and Domestic Political Violence : The American Capacity for Repression and Reaction, Sage Publications, Beverly Hills, 1976. SULLIVAN (J.L) « Multiple Indicators : Some Criteria of Selection », in Hubert Bialock (ed.). Measurement in the Social Sciences, Aldine Publishing Co., Chicago, 1974. TUFTE (E-R) : Data Analysis for Politics and Policy, Prentice Hall, Englewood Cliffs, 1974. U.S. DEPARTMENT OF THE INTERIOR, CENSUS OFFICE : Report on Crime, Pauperism, and Benevolence in the United States at the Eleventh Census, 1890, Government Printing Office, Washington, 1896. ZETTERBERG (H.L.) : On Theory and Verification in Sociology, Bedminster Press, Totowa, (3rd. ed.), 1966. ee — 222 — CONSENSUS, CONFLIT ET ATTITUDES ENVERS LE CRIME ET LA LOI PENALE PHILIPPE ROBERT, Service d’études pénales et criminologiques, (équipe de recherche associée au Centre national de la recherche scientifique). PARIS - FRANCE. Il est devenu courant, surtout depuis une dizaine d’années, de parler de crise du système pénal. L’une des premières idées qui soit alors venue à l’esprit a été de se demander si ce sentiment de crise ne provenait pas essentiellement du monument législatif — voire jurisprudentiel — par rapport à l’évo- lution de « l’opinion publique ». C’est au fond la problématique initiale des travaux « K.O.L. » (1) du comité de sociologie du droit de l’Association internationale de sociologie. On s’assignait pour tâche de détecter et de mesurer le décalage existant éventuellement entre « l’opinion publique», les prescriptions légales et même quelquefois leur application prétorienne dans la jurisprudence. Le problème paraissait alors simple, du moins dans son principe : il s’agissait de connaître l’évolution de « l’opinion publique », puis de moderniser la loi pour l’adapter à l’état nouveau de «l’opinion » prise comme indicateur de l’évolution des mœurs reçues dans la société. Parmi beaucoup de tâtonnements, le bilan de cette première vague de travaux a, au moins, permis de découvrir un point tout à fait essentiel : la faiblesse de la dimension cognitive dans les représentations de la loi et de la justice pénale. Mais que la connaissance soit limitée et de mauvaise qualité ne doit pas conduire à conclure à l’inexistence de représentations et de structures d’attitudes. Seulement, elles sont caractérisées par leurs dimensions normatives et affectives (2) plus que par celle de connaissance. (1) Knowledge and Opinion about Law (Réf. 1). (2) D’après ma propre expérience de recherche, j'aurais même tendance à préciser que la dominante est en général normative, tandis qu’elle est affective pour les gens qui ont eu maille à partir avec la justice pénale (Réf. 2). — 223 — C’est d’ailleurs la difficulté à laquelle se sont heurtées la plupart des études classiques de sociologie du droit qui ont axé sur le cognitif leurs instruments d’investigation.. outre qu’elles ont souvent sous- estimé les problèmes de visibilité (3). Quoiqu’il en soit, ce point doit être considéré comme définiti- vement acquis aux débats depuis le magistral rapport de Berl Kutchinsky au Conseil de l’Europe (Réf. 4), d’ailleurs corroboré — pour ce qui concerne cette question — par un ouvrage récent de G.R. Newman (Réf. 5). À vrai dire, on n’aurait même pas pris la peine de s’étendre ici sur ce problème de la dimension cognitive si l’on ne relevait encore des ambiguïtés et des méconnaissances dans des articles pourtant récents (Réf. 6). Malgré cet acquis — d’ailleurs souvent découvert seulement a contrario et a posteriori — il faut bien admettre que la problématique initiale d’un décalage entre loi et « opinion » n’a guère permis une production décisive et qu’elle ne le pouvait pas. En premier lieu, elle charriait une espèce de postulat optimiste et naïf d’évolution linéaire qui est directement importé des pseudo-évidences du sens commun, mais dont le statut scientifique demeure tout à fait flou — et d’autant plus flou qu’il reste implicite (4). En second lieu, il s’est peu à peu avéré que l’ainsi nommée « opinion publique » constituait un pôle de référence moins simple et plus complexe qu’on ne l’imaginait naïvement au début. A priori, tout le monde parle de « l’opinion publique » et, d’ailleurs, nul ne doute de la connaître aisément et de pouvoir se targuer de son autorité souveraine (5). On céderait les armes devant tant d’assu- rance si l’on ne s’avisait que chacun l’invoque en des sens différents, voire contradictoires. De sorte naît le soupçon que ce syntagme ne soit que la couverture idéologique (6) des idiosyncrasies des uns et des autres. On découvre alors que l’on baptise de cette expression — apparemment indiscutable — qui les avis de ses familiers, qui les propos de son journal favori, ou de sa chaîne de télévision accoutumée, (3) Cette question de visibilité est en fait très importante pour notre matière et le plus souvent négligée ; mais je ne peux en traiter ici et je dois donc renvoyer à ce qui en est dit dans une communication, d’allure plus méthodolo- gique, faite avcc un de mes collègues lors d’un colloque du C.N.R.S. (Réf. 3). (4) L’effort scientifique, c’est-à-dire de connaissance et de compréhension, commence seulement lorsqu’on se prend à questionner les assertions même les plus assurées du sens commun et les évidences les mieux admises (Réf. 7) et cette prescription doit rester tout particulièrement à l’esprit quand on prétend étudier des matières tout empreintes de normativité (Réf. 8). (5) Souveraine car évoquant la vox populi. (6) Sans qu’il soit pertinent ni même utile de supposer des intentions perverses ou même une conscience claire chez l’agent. — 224 — qui enfin ces sondages commerciaux si fort à la mode actuellement et que s'offrent régulièrement journaux, gouvernants, groupes de pression. Ce n’est pas le lieu ici de démontrer à nouveau les vices le plus souvent rédhibitoires de cette activité si lucrative et je prends la liberté de renvoyer à d’autres développements d’allure plus méthodologique (Réf. 9). Le fâcheux est de constater que maints travaux scientifiques sont tombés, faute de conceptualisation suffisante ou par défaut méthodologique, dans des pièges sensiblement analogues. Là encore, Je ne peux détailler ici ces derniers et je dois procéder par renvoi (Réf. 10). Quoiqu’il en soit, il a bien fallu se rendre à l’évidence : travailler au niveau des manifestations d’opinion — c’est-à-dire traiter comme variables les données immédiatement recueillies (7) — revenait tout simplement à prétendre s’intéresser au phénomène des marées tout en se contentant de recueillir l’écume des vagues, Il est clair que les manifestations d’opinion (8), tout comme les autres données comporte- mentales, ne peuvent avoir d’utilité que comme indicateurs traitées de manière agrégée et non séparément afin de découvrir les régula- rités sous-jacentes, elles permettent de faire émerger les structures d’attitudes, puis les types de représentations (9). A ce moment, on s’aperçoit combien on se trouvait embarqué dans des voies trompeuses par la notion même d’une « opinion publi- que ». Premièrement ce pseudo-concept conduit fatalement à rai- sonner en termes d’opinion moyenne sans pouvoir dire si cette moyenne a une signification quelconque. Deuxièmement, elle fait tomber presque inéluctablement dans le piège des homologies apparentes. Prenons un exemple très simple on est tenté de ranger dans la même catégorie tous cux qui se déclarent mécontents de la justice pénale. Or, une analyse plus attentive montre que, ce faisant, on classe comme sem- blables des gens que tout sépare en réalité car ce mécontentement vient, par exemple, pour les uns du sentiment d’une trop grande indulgence mettant en péril les valeurs éternelles, pour les autres d’une impression de rigidité s’opposant à toute évolution jugée souhai- table, ou, du moins, la contrariant. On est donc conduit à substituer à l’étude du décalage entre loi et « opinion publique » un examen des sortes d’attitudes et des types de représentation. Mais, alors, on soumet fatalement à interro- (7) A supposer que ce recueîl des données ait été opéré èdrrectement. (8) Qu’il s’agisse d’un discours et de fragments de discours selon la nature plus ou moins standardisée de la méthodologie de recueil des données. (9) Il est impossible dans ce bref article de préciser l’agencement conceptuel dans lequel j’use de ces différentes notions. On en trouvera un exposé dans d’autres publications (Réf. 11). — 225 — gation, comme le montrait déjà T. Sellin (Réf, 12), l’assertion qui fait de la loi l’expression de la « volonté générale ». Dans le même temps où l’évaluation des travaux sur loi et « opinion » faisait découvrir l’obligation de raisonner en termes d’atti- tudes et de sortes de représentations, d’autres raisons militaient encore en faveur du développement de telles recherches, Longtemps la sociologie du crime et de la déviance n’a accordé au système pénal qu’une attention marginale. Toute entière adonnée à l’examen du déviant et de la déviance, elle négligeait largement d’étudier la main qui lui tendait, pourtant, le plus clair de son infor- mation. On réputait — implicitement — que le système pénal était parfai- tement transparent ; on le considérait comme une sorte de moule qui recevait passivement l’impact de la criminalité — ce qui mettait le chercheur à même d’étudier ensuite cette dernière. Bref, la justice pénale n’était pour le sociologue du crime que la plaque sensible sur laquelle s’était imprimé son objet d’étude. On considérait donc le processus pénal comme opérant un simple transfert de données, une translation ayant pour effet de mettre l’objet d’étude à la disposition du chercheur (10). Pareille attitude n’a pas été sans inconvénient pour la sociologie du crime et de la déviance. En effet, elle se résignait — ce faisant — à ignorer les conditions de production de sa matière première. Comme un myope, elle s’enfermait dans un objet d’étude prétendument évident selon les évidences de sens commun. Mais ceci lui dissimulait un phénomène fondamental et rendait impossible un autre, tout aussi important. Cette boîte qu’on croyait de verre, qu’il n’importait pas d’ouvrir — la justice pénale — s’avérait être une boîte noire bien difficile à connaître, En outre, ne pas ouvrir la boîte noire de la justice pénale est se condamner à une sociologie provincialiste qui ne pourra jamais rattacher — sinon nonchalamment — son propos à une sociologie historiquement datée de la société entière. Ces deux faiblesses classiques ont été durement reprochées à la sociologie du crime et de la déviance. Sous l’effet de pareille critique, notre discipline est venue à s’intéresser plus activement au système pénal, et d’ailleurs d’autant plus que celui-ci émergeait lentement comme problème social, comme institution en crise : il devenait objet de préoccupations pour l’homme politique ou l’administrateur, d’engagement pour le militant… fina- lement d’étude pour le chercheur. (10) Toût au plu$ se désolait-on que des contraintes administratives, juridiques des renseignements sur le crime et le criminel. ou le manque de moyens viennent parfois restreindre l’étendue ou la complétion — 226 — Avec le relatif recul du temps, les coups d’essai ne sont toutefois DIs apparus ocomme ayant été pleinement concluants. D’un côté, nous avons observé depuis le début des années 60 le pullulement d’écoles anglo-saxonnes — et surtout américaines — qui prétendaient étudier les modalités de contrôle social de la criminalité. Elles y parvenaient mal cependant par une ambiguïté dans la problé- matique. Malgré leur appellation, ces écoles se bornaient le plus souvent à conïérer au contrôle social le statut de variable explicative de la déviance, oubliant que la différence de logique entre les deux phéno- mèncs oolige à y voir deux problématiques irréductibles l’une à l’autre (Réf. 13). D’un autre côté, nous avons vu se développer une sociologie technologique qui focalisait certes son attention sur le système pénal, mais dans un laxisme épistémologique lui faisant accepter comme des postulats tous les lieux communs, toutes les idées reçues sur le système pénal. Il aurait été pourtant préférable de les traiter comme de simples hypothèses à tester : en agissant comme l’a fait ce courant de pensée, on arrive à supprimer tout objet de recherche. Le résultat n’ajoute Tien aux prémisses excepté de l’ésotérisme ou de la « recettologie ». En fait, ce courant n’a guère été fructueux il a redit en termes compliqués les assertions — même erronées -— du sens commun ; ou alors, il s’est borné à du rapiéçage tandis qu’empirait imperturbablement une crise qui se comprend au niveau de la fonction sociale du système pénal et pas tellement (11) à celui de son fonction- nement. Enfin, on a souvent admis que le système pénal se ramène à un ensemble — plus ou moins complexe — d’institutions ou d’agences (12) entretenant entre elles des relations systémiques et l’on borne l’attention à ce complexe institutionnel. On risque ainsi de revenir ou de rester à un simple fonctionnalisme. Or, les institutions constituant le système pénal sont seulement comme la partie émergée d’un iceberg… Pour la comprendre finalement il faut s’avérer capable de sonder la masse bien plus importante qui se trouve immergée. En clair, non seulement le contrôle social de la criminalité n’est pas entièrement institution- nalisé (13), mais encore ce qui l’est ne fonctionne qu’adossé à cette (11) Ou seulement par ricochet. (12) Police, ministère public, juridictions, agences d’exécution pour majeurs ou mineurs, de milieu fermé ou ouvert. - (13) On (KRéf. 14) a ainsi montré que la double spirale de ségrégation qui carac- térise la situation de bandes de jeunes et qui organise ses rapports à l’environ- nement, se développe fort bien même sans étiquetage par le système pénal, donc déjà au seul niveau des relations informelles. Des travaux récents ont en outre souligné que les représentations de la déviance et de son contrôle confinent généralement le système pénal dans un rôle marginal et limité à certaines cibles bien délimitées. — 227 — énorme masse peu — ou pas — institutionnalisée dont l’étude est trop souvent négligée. Ce faisant, on se condamne à mal comprendre le fonctionnement de la justice pénale et pas du tout à sa place dans la structure sociale. La réalité du système pénal n’est donc pas épuisée par le seul examen de ses règles et modalités institutionnalisées et de ses produits. Elle se découvre encore dans les représentations qu’on en a et leur connaissance est indispensable si l’on veut ouvrir la boîte noire. Les produits et modes d’opérer du système pénal sont indisso- ciablement liés à ce qu’on en pense, à ses représentations. Ce n’est pas que ces deux aspects de la réalité pénale soient réductibles l’un à l’autre. Mais il y a interaction entre les deux (Réf., 15). Tout d’abord le fonctionnement concret de la justice pénale dépend de manière importante des représentations qui en ont cours. Pour qu’une affaire arrive aux institutions répressives, il faut que les faits soient suffisamment visibles ou qu’il se trouve quelqu’un pour avertir la police. La visibilité, quant à elle, dépend de conditions objectives et fort variables selon les cas. Il n’est pas besoin d’être grand clerc pour réaliser que certains passent le plus clair de leur vie à l’abri des regards indiscrets tandis que d’autres vivent en perma- nence à découvert : les illégalismes de ceux-ci sont plus « visibles » que ceux des autres. Et cette visibilité est fort liée au statut (notamment professionnel) des agents, donc finalement à leur place dans la structure sociale. Mais la visibilité n’est pas tout. Les faits que les institutions répressives découvrent vraiment de leur seule initiative ne sont pas les plus nombreux. En définitive, la majeure partie de sa matière première vient au système pénal parce que quelqu’un est venu de quelque manière lui signaler l’affaire. Ceci met naturellement en jeu les conditions objectives de reportabilité : certaines infractions — sans victime individuelle directe (14) — auront moins de chances d’être dénoncées puisque nul ne s’en reconnaît directement et individuellement victime c’est le cas d’une bonne partie du crime organisé et de la criminalité d’affaires. En sens inverse, les atteintes à la propriété — du vol de vo'ture au hold-up — auront le maximum de chances d'être reportées à la police pour cette simple raison que les compagnies d’assurances exigent souvent pareil préalable à la déclaration de sinistre. Mais la responsabilité met encore et surtout en jeu d’autres mécanismes moins clairs mais tout aussi influents. La décision de (14) En fait elles pèsent non sur un individu particulier mais sur des fractions entières de population : ainsi de la fraude fiscale qui porte tort à tous ceux qui ne peuvent guère y recourir car leurs revenus sont déclarés par autrui (donc les salariés). — 228 — reporter un fait à la police dépend finalement beaucoup des repré- sentations que l'on se fait du délit (Jle comportement dont nous sommes témoin ou victime est-il considéré comme un « vrai délit » ?), du délin- quant (rangeons-nous son auteur parmi les « vrais » délinquants relevant de la répression pénale ?), finalement du système pénal et de son adéquation (à quoi doit-il servir ?). Bref, la connaissance d’un fait par la justice dépend donc à la fois — de conditions objectives variables selon l’appartenance sociale comme la plus ou moins grande visibilité ou encore les circonstances poussant à la reportabilité ou la rendant improbable, — Inais aussi des représentations du délit, du délinquant et de l’adéquation de chaque appareil de contrôle social, dont le système pénal. L'influence des représentations sur le fonctionnement du méca- nisme pénal ne s’arrête pas à cette phase préliminaire — et d’ailleurs fondamentale. On a montré que le système de justice pénale opère — parmi la matière première qu’il découvre ou qu’on lui amène — un tri et une progressive construction de son produit fini. Celui-ci s’élabore à travers les phases successives de ce mécanisme et, à chaque moment, entrent en jeu les représentations que se font les opérateurs du système pénal, bref leur idéologie professionnelle (15). Ajoutons que ces représentations — qui conditionnent de manière importante le fonctionnement de la justice pénale — ne sont pas de génération spontanée. De quelque manière, elles dépendent en fin de compte de la structure de l’appareil pénal, de son fonctionnement, surtout de sa production. Dans les représentations du système pénal, on constate l’existence de fortes stéréotypies reposant sur la production d’une image type dc son produit fini. On observe également que les appareils de production idéologique — comme par exemple la presse — diffusent de telles stéréotypies et de semblables images types Or, l’existence de cette colonisation idéologique s’alimente au mécanisme par lequel le système pénal construit progressivement son objet. Parmi les données et individus constituant la matière première initiale, certains éléments sont sélectionnés et d’autres rejetés. Puis, ceux qui ont été conservés sont reconstruits selon la logique interne propre à œ système et à sa fonction sociale, de sorte que son intervention — permise en fait par les conditions objectives et les représentations des intervenants extérieurs, lesquels éléments ne sont pas aléatoires mais significativement différents selon la structure sociale — paraît avec (15) Idéologie qui, tout comme celle du public, est loin d’être transparente pour le chercheur. Il y a tout à la fois intériorisation des contraintes institu- tionnelles, négociations au niveau des pratiques, et remaniements afin d’éviter une trop grande dissonance entre normes et réalités des pratiques. — 229 — les atiributs inéluctables du destin. Par cette opération, s’édifient des images types de la clienièle de la justice pénale et de l’adéquation de cet appareil à telle ou telle intervention, Or, il se trouve que c produit fini émerge seul — et encore partiellement — de la boîte noire, qu’il est plus visiblee On comprend la place qu’il tiendra dans la production des représentations du délit, du délinquant et du système pénal. Toutefois ces représentations ne sont pas une photographie du fonctionnement. Tout n’est pas également visible dans l’appareil pénal, tout n’y est pas non plus également compréhensible. De surcroît, visibi- lité et ésotérisme varient selon la situation des différents groupes sociaux pertinents (16). Ajoutons encore à cela l’existence d’intermédiaires, d’entrepreneurs idéologiques — comme la presse — qui informent, forment et déforment. Tenons ainsi compte du fait que la justice pénale a une histoire fort ancienne, de sorte que le changement d’aujourd’hui peut mettre longtemps à effacer la trace d’hier. N’oublions pas que l’on ne se représente pas le système pénal seul en isolat, que son image doit donc s’insérer dans des conceptions du monde, des visions de la vie sociale. Enfin, dans la détermination des conduites effectives des différents groupes, leurs attitudes devront composer avec des contraintes de situation — variables d’un groupe à l’autre — et qui rétroagiront sur les représentations et les attitudes. Voici autant de considérations qui expliquent pourquoi les idées qu’on se fait sur le système pénal (qui traduisent certaines modalités de relations à ce même système), ont une consistance propre et une autonomie certaine — quoique relative — par rapport à son fonc- tionnement. Mais on n’oubliera pas qu’en sens inverse les représentations conditionnent en bonne partie le fonctionnement et la production de la justice pénale. Ce qui paraît du domaine du symbolique s’avère posséder des vertus très concrètes, Et voilà pourquoi une bonne connaissance du système pénal ne peut s’arrêter à l’analyse des produits et modes d’opérer mais doit encore s’élargir aux représentations. Au reste, la réalité de l’appareil pénal ne se borne pas à ce qu’il est aujourd’hui compte même tenu du poids de l’histoire. Elle est (16) Lorsque nous parlons ici de groupes sociaux pertinents, il ne s’agit pas de groupes définis a posteriori par la recherche — ce qui correspondrait à un « emnirisme » récusable — ni aux classiques catégories socioprofessionnetlles. qui elles-même recouvrent une théorie souvent peu explicitée. Il s’agit de grounes définis nar une démarche théorico-conceptuelle préalable et indisso- ciable de l’ensemble de la recherche. Sinon, on ne fait que se laisser piéger par l’usage d’instruments préexistants. — 230 — encore dans les contradictions, conflits, attentes qui peuvent s’observer à son propos. Et cette nécessité — pour bien connaître l’appareil pénal — de le situer historiquement dans un présent instable entre un passé et un devenir augmente l’urgence d’une investigation sur les représentations… La conjonction de ces diverses considérations explique le délais- sement progressif de la problématique initiale du décalage entre loi et « opinion » (17) au profit d’une interrogation sur la diversité des attitudes et des représentations concernant les matières pénales. Mais cet approfondissement, en principe satisfaisant, risque d’être réduit de manière telle qu’on passera une fois de plus à côté des questions pertinentes. C’est à propos que je voudrais consacrer ce papier à quelques observations sur l’utilisation de la dichotomie consensus vs conflit dans l’étude des attitudes envers le crime et la loi pénale. » Les recherches sur les attitudes envers le crime et la loi sont, en effet, en passe d’être dominées par une opposition simple — voire simpliste — entre consensus et conflit. Cette dichotomie a surtout fait fortune en Amérique du Nord ; elle prétend ranger les sociologues de la déviance en deux camps : ceux qui postulent l’existence au sein d’une société donnée — voire de l’humanité — d’un accord minimum au sujet des problèmes criminels ; et ceux qui se refusent à pareil postulat et ne voient à ce propos qu’attitudes contradictoires, voire opposées. Celui qui s’aventure dans l’étude des représentations et attitudes concernant le crime et la loi pénale est sommé de se ranger dans l’un de ces camps, faute de quoi on y pourvoira pour lui. Je voudrais montrer que ce simplisme est peu fécond — ou, du moins, insuffisant — qu’il risque même de faire passer à côté des interrogations pertinentes. Pareille discussion me sera occasion de traiter, à la lumière de diverses recherches récentes, de quelques questions essentielles à l’étude des attitudes envers le crime et la loi pénale. Û.Û Considérons iout d’abord que cette opposition — qui est appa- mmmt simple — s’avère en fait être à contenu variable. (17) Encore que certains ouvrages récents y reviennent, comme le fait G.R. et Caricatural de la dichotomis consensus vs conflit. Newman (Réf. 16), mais justement, me semble-t-il. par l’effet d’un usage abusif — 231 — On s’en aperçoit tout de suite en cherchant à préciser ce que l’on recouvre de l’étiquette « conflit ». Dans une acceptation étroite, l’expression « école du conflit » est appliquée à ditférents auteurs qui s’inspirent de R. Dahrendorf (Réf. 17), notamment naturellement A. Turk (Réf. 18), mais aussi, dans leur première époque, W.J. Chambliss (Réf. /9) et, dans une certaine mesure, R. Quinney (Réf. 20). Chambliss est d’ailleurs l’un des respon- sables de cette dichotomie puisqu’il la pose comme summa divisio en tête de son ouvrage de 1971 (Réf. 2/). Tl n’est toutefois que juste de rappeler que Chambliss subdivisait l’opposition consensus- conflit. Il distinguait dabord ceux pour qui toute société reposait sur une certaine communauté consensuelle de ceux qui y voient fonda- mentalement un champ de conflit ; mais parmi ces derniers, il subdis- tinguait ceux pour aui l’Etat est au-dessus de la mêlée de ceux aux yeux de qui 1l est l’enjeu du conflit. Quoiqu'il en soit, cette appellatiôn — école du conflit — avait de essentiellement pour fonction — dans la pratique courante distinguer la filiation de R. Dahrendorf de celle de G.H. Mead (Réf. 22), c’est-à-dire tout ce que l’on range lato sensu sous le vocable d’inter- actionnismes. Toutefois, malgré la différence de filiations, la distinction apparaît finalement assezpeu au niveau des produits. Malgré les efforts de Chambliss notamment (18), on y étudie surtout les aspects différentiels et inégalitaires de la répression pénale, mais sans savoir focaliser décidément sa problématique sur la norme pénale et son effectivité au sein d’une sociologie globale, Dans une acceptation plus large, on étend l’étiquette « conflit » à des post-interactionnistes plus déclarés comme les criminologues radi- caux ou critiques et même jusqu’à l’aile la plus radicale de l’ancien group> de Berkeley : la tendance manifestée depuis 1970, par T. Platt (Réf. 23) ct les Schwendinger (Réf. 24). En fait, là encore, la rupture est plus verbale que réelle. De tcls auteurs adhèrent souvent à une alternative humaniste modrne qui ne rompt pas substantiellement avec l’idéalisme antérieur. Lode Van Outrive (Réf. 25) a justement souligné que cett école ne parvient pas à dépasser le niveau atteint par les interactionnistes et les phénoméno- logues-ethnométhodologues. De proche en proche, le terme de conflit est parfois utilisé pour recouvrir tous les post-interactionnistes ; on l’applique pêle-mêle aux « radicaux », aux «critiques », aux néo-marxistes, etc… En fin de compte, ce syncrétisme n’est pas tellement étonnant puisque — à (18) Tout spécialement dans ses travaux de sociologie de droit. — 232 — quelques rares exceptions près (19) — on retrouve dans tout ce foison- nement de prises de position un vieux fonds commun de nominalisme idéaliste et d’humanisme dont on vient de dresser un bilan sans complaisance (Réf. 26); il ressort bien à la définition que donne N. Poulantzas de l’ultragauchisme petit bourgeois (Réf. 27). Au gré des auteurs, l’opposition consensus-conflit sert donc soit à identifier la filiation de Dahrendorf, soit à désigner tous les post- interactionnistes, qu’ils se baptisent anti-criminologues, criminologues critiques, criminologues radicaux ou néo-marxistes. Mais la géométrie variable de cette opposition peut même être utilisée pour couvrir une marchandise plus vaste et désigner tous les sociologues de la déviance postfonctionnalistes. C’est d’ailleurs la porte qu’ouvrait Chambliss en identifiant le consensus à T. Parsons (Réf. 28). p Il est d’ailleurs bien vrai qu’il y a — toujours à quelques exceptions près — moins de rupture fondamentale qu’on ne pourrait le penser dans la multiplication des écoles qui marque la littérature criminologique des quinze dernières années… ou, du moins, une véritable rupture épistémologique à été préparée plutôt qu’accomplie. C’est que, malgré toutes les différences de tendances, ces auteurs ont surtout songé à inverser le sens de l’équation, à atteler la locomotive en queue et non plus en tête de train. Mais un tel changement du sens de la marche laisse intact le problème fondamental de l’objet de recherche. Toutes ces criminologies demeurent finalement focalisées dans leur log'que ultime, sur le crime et le criminel, en ce sens qu’elles restent des criminologies du passage à l’acte (Réf. 29). Simplement : — les interactionnistes ont fait de la réaction sociale une — voire la — variable explicative de la criminalité ; — les théoriciens du conflit ——n striclo sensu —— ont mis l’accent sur le traitement discriminatoire par le système pénal (20) ; — la plupart des radicaux ont proposé de considérer d’autres crimes que ceux habituellement ressassés, et donc d’autres criminels que ceux recrutés dans les habituelles « classes laborieuses, classes dangereuses » (Réf. 32). (19) Qui se situent généralement dans différents pays européens et qu’il ne faut pas traiter par prétérition sous prétexte qu’ils sont souvent interna- tionalement peu connus. (20) Souvent sans aller beaucoup plus loin qu'un fonctionnalisme rénové ; en tout cas en ignorant ou €n traitant de manière infirme tout ce qui a trait aux idéologies. De ce point de vue, peu de progrès scientifique a été accompli depuis le bilan dressé dans Crime and the Legal Process (Réf. 30). En fait, la plupart de ces travaux n’ont pas su égaler D. Chapman qui demeure, pour ce courant, un précurseur solitaire et souvent ignoré (Réf. 31). — 233 — Mais tout cela n’empêche pas que ces différentes « écoles» se laissent généralement imposer un objet de recherche préconstruit par les errements du sens commun, par les représentations de l’idéologie dominante. Elles n’ont généralement pas su mener à bien une construc- tion réelle de leur objet, bref, elles n’ont pas pu réussir une véritable réflexion sur leur problématique. De la sorte, la « rupture épistémo- logique » n’est pas véritablement assurée avec les errements criminolo- giques classiques, notamment avec les culturalistes et fonctionnalistes. Même en adoptant une géométrie variable, l’opposition consensus- conflit hypertrophie à mauvais droit les changements théoriques inter- venus ces dernières décennies à propos de notre ob]ct d’étude. Elle a donc peu de justification. Malgré quoi, on assiste à un nouvel avatar qui vise à en faire une division permanente des théories en sociologie de la déviance Dans un ouvrage récent, Lynn Mc Donald entreprend de reconstruire toute l’histoire de cette discipline depuis le 18* siècle, et même au-delà, sous les espèces d’un combat entre tenants du consensus et tenants du conflit (Réf. 33) L’exemple de ce livre — par ailleurs sympathique et intéressant — est utile en ce qu’il permet de découvrir l’un des vices majeurs de l’opposition consensus-conflit : l’idéalisme. En extrapolant de proche en proche à partir de la considération de la filiation anglo-saxonne de Dahrendorf, on parvient à reconstruire un sorte d'histoire de la sociologie de la déviance dépeinte comme l’affrontement sans cesse renouvelé de « l’école » du consensus et de celle du conflit Mais, ce faisant, on ne présente plus qu’un théâtre d’ombres où les idées paraissent évoluer en soi, l’une poussant l’autre, sous l’effet d’on ne sait quel deus ex machina Qu'il y ait sans cesse, en sociologie de la déviance, une majorité de conformistes qui acceptent les soit-disant évidences du sens commun — Singulièrement celles d’un consensus social qu’exprimerait la loi c’est-à-dire l’Etat — et une minorité qui se réserve le droit d’interroger ces postulats, c’est-à-dire de ne pas se borner à ressortir en termes abscons les constructions d’objet du sens commun, voilà qui est clair En faire pour autant des sortes « d’écoles» qui traversent les siècles me paraît cependant exagéré et dénué de pouvoir explicatif. On a l’impression de relire chez Loyola la bataille des deux étendards (21) (Réf. 34) (21) Je voudrais répéter une fois de plus que l’utilisation de cette image à titre de comparaison ne veut pas dire que je mésestime les qualités certaines et l’intérêt de l’ouvrage de Lynn Mc Donald, mais j’ai eu l’occasion de dire plus longuement mon sentiment sur ce travail dans une autre revue (Réf. 35) —— 234 — Au reste. Lynn Mc Donald est obligée — pour s’en tirer — de supposer un échange d’argumentaires en cours de joute peu à peu, ceux qui adoptent une vue « conflictuelle » auraient délaissé les théories à étiologie économicisante pour mettre l’accent principalement sur le développement cancérigène des systèmes spécialisés de contrôle social, tandis que les avocats du consensus auraient récupéré les notions de problèmes sociaux, d’inégalités et de privation. Comment identifier sous le même courant ceux qui séparent des renversements aussi radicaux d'argumentaires ? Comment, sinon au prix d’un incompréhensible idéalisme, faire de cette dichotomie consensus vs conflit deux courants qui se partageraient le champ de la sociologie de la déviance et suffiraient à expliquer son développement ? De surcroît, cette opposition n’explique rien puisque les différences ainsi mises en exergue ne sont pas fondamentales : du point de vue de la problématique, de l’objet de recherche, bien des « conflictuels » ne se séparent pas radicalement, nous l’avons vu, des « consensuels ». se ° L'usage de l’opposition consensus vs conflit expose — outre l’idéalisme — à un second danger caricaturer une opposition et, ce faisant, passer à côté de certaines questions fondamentales pour l’étude des attitudes envers le crime et la loi pénale. Ainsi un travail récent présente-t-il les deux positions, l’une comme “ stipulant un assez large accord sur les conduites à réprouver et à réprimer, l’autre comme prétendant qu’il n’existe aucun accord sur ce point (Réf. 36). Et G.R. Newman prétend trancher le dialogue de sourds qu’il croit s'être instauré entre les deux « écoles » du consensus et du conflit celle-là n’appliquant son effort qu’à des infractions classiques et peu discutées, celle-ci bornant son intérêt aux cas litigieux des crimes sans victime. D’où, soutient-il. l’impossibilité de trancher Sérieusement entre elles. Lui, prétend, au contraire, résoudre définiti- vement cette querelle en appliquant à des populations tirées dans six pays (Inde. Indonésie, Iran, Italie, Etat-Unis et Yougoslavie) un questionnaire comprenant des cas variés d’infractions (mmala in se de type inceste, mmala antiqua de type vol qualifié, mala prohibita comme es crimes sans victime, mala nova comme la pollution, mala ambigua comme l’omission de porter secours ou les infractions liées aux mani- festations de protestation). Il souligne un très large accord transculturel sur la réprobation tandis que d’importantes variations apparaissent dans les définitions, la connaissance de la loi et la perception de l’action répressive des agences officielles. Qu'’il trouve un large accord sur un sentiment de gravité-répro- bation l’amène à se ranger aux côtés de P. Sorokin (Réf. 37) : les lois sont l’expression de l’état de l’opinion. On se souvient, en effet, que — 235 — A.K. Cohen (Réf. 38)-—employant avant la lettre l’opposition consensus- conflit à propos de la genèse des lois pénales — avait ramené la controverse sur les attitudes envers la loi énale à une opposition entre la position de Sorokin et celle de Becker (Réf. 39). Quant à Newman, ce résultat le conduit à conclure que les théories consensuelles se trouvent ainsi vérifiées ct celles du conflit rejetées et à revenir à la problématique en termes de décalage loi-opinion. Toutefois, cette conclusion me paraît invalide en raison du caractère caricatural de l’opposition sur laquelle repose ce travail. Pourtant, Newman prend un certain nombre de précautions conceptuelles louables — et qui constituent un apport solide de son ouvrage. Ainsi précise-t-il que le consensus ne correspond pas à un problème de majorité-minorité, s'agirait-il même d’une majorité absolue. Il distingue, de manière convaincante, quatre situations -— CONSENSUS à une position structurée ne s’en oppose aucune autre ; —- dissensus plusieurs positions structurées s’opposent, l’une majoritaire, les autres minoritaires : — conflit : plusieurs positions s’opposent sans qu'aucune prédomine clai- rement ; — ASEnNsus : il y a atomisation des positions. Seulement, il ne tire pas des conclusions très convaincantes de cette excellente classification (Réf. 40). Qu’il y ait un assez large accord pour désapprouver les conduites analogues à celles retenues dans son questionnaire, qu’il y ait même un certain accord pour considérer qu’elles relèvent adéquatement du système pénal — ce qui apparait déjà beaucoup moins nettement dans ses résultats — donc qu'’elles doivent être considérées comme des infractions. voilà qui est assez trivial et ne permet pas de trancher du consensus. * %% En fait 1l faut sérier plusieurs problèmes dans les représentations du crime et de la loi pénale. Les représentations sont structurées par plusieurs dimensions d’attitudes (22). Une première et essentielle distinction sépare le sentiment de gravité et celui d'adéquation d'une modalité de contrôle à un type de (22) Et les positions sur chaque dimension d’attitude peuvent varier selon les groupes sociaux pertinents, de sorte que l’on pourra observer diverses combinaisons de ces attitudes, c’est-à-dire différents types de représentation. — 236 — conduite (Réf. 4/), Beaucoup d’auteurs (Réf. 42) confondent, en effet. les deux ils partent de l’idée d’un continuum linéaire unidimensionnel sur lequel s'étaleraient les perceptions de la déviance, avec, au sommet, la plage de la criminalité, qui serait caractérisée comme la déviance la plus grave — ou du moins celle considérée comme telle. Ce faisant ils postulent une continuité perçue là où bien des travaux ont surtout mis l’accent sur l’existence de solutions de continuité perçues (23). Et dans le même temps, ils confondent deux dimensions de perception la gravité et l’adéquation. Or, une même conduite peut très bien, dans un type donné dc représentations, être considérée come relevant adéquatement du système pénal — donc comme infraction — et pourtant, en même temps, comme moins grave que telle autre estimée relever adéquatement d’un contrôle social non pénal, donc pas comme infraction. Sans vouloir s’étendre ici trop longuement sur un point qui a fait l’objet de travaux par ailleurs (Réf. 43), on dira seulement qu’il est impossible d'’étudier ensemble ces deux dimensions perceptuelles que sont la gravité de l'adéquation. Celle-ci domine les représentations générales des déviances. Elle correspond en matière de représentations au problème juridique du seuil de criminalisation-décriminalisation. La sociologie des représentations ne peut aborder ce dernier qu’en termes de perception d'adéquation et non en termes de perception de eravité. D'ailleurs, Newman aperçoit bien cettc distinction conceptuelle et ses résultats en soulignent clairement la pertinence. Il montre notam- ment qu’il existe d'assez larges variations dans ses six populations sur l’attitude d’adéquation de telle forme de contrôle à telle déviance (24). On lui reprochera seulement de ne guère respecter par la suite ce qu’il a lui-même contribué à clarifier, notamment en construisant deux indices et surtout le second que l’on peut assez justement soupçonner de mélanger de manière abusive ces différentes dimensions. _— (2.31).'Â'_'cthoutes les conséquences méthodologiques qu’entraîne l’emploi d’instruments supposant une continuité linéaire. (24) Encore ses résultats ne pouvaient-ils être très développés pour trois raisons a) comme son sous-titre l’indique plus que son titre, le questionnaire utilisé bornait l’investigation à quelques cas d’infractions sans permettre d’étudier les représentations de la déviance en général ; b) il a traité chaque population nationale — dont on ne connaît guère les modes de stratifications — comme un tout faute d’y appliquer une analyse selon les groupes sociaux pertinents c) enfin — comme nous le redirons en conclusion — sa méthode compa- rative postule une sorte d’universalisme idéaliste qui évacue — mal les apparences — toute compréhension de l’historicité des sociétés et de leurs instances juridiques - normatives, ainsi que des phénomènes idéolo- giques qui y sont produits. — 237 — Mais abandonnons maintenant le problème du seuil de crimi- nalisation-décriminalisation et donc celui de la perception d’adéquation d’une forme de contrôle social à une déviance donnée. Nous retrouvons alors la question de l’attitude envers la gravité. Ce que ne comprend pas Newman — et c'est le principal grei que l’on peut lui faire — c’est que les divergences ne peuvent pas — ou seulement marginalement —- porter sur la perception de gravité-réprobation. Aboutir à faire dire aux gens que des conduites incriminées par la loi pénale sont plutôt réprouvables que recomman- dables paraît finalement assez trivial, En fait la perception de gravité doit se traiter en termes d’ordre de priorités relatives. La loi pénale incrimine un nombre très large de conduites et ce n’est un mystère pour personne que son effectivité répressive varie de manière très importante de l’une à l’autre. Elle recèle d’ailleurs de superbes contradictions internes qui ne peuvent se résoudre en pratique qu’en introduisant des choix, des priorités de fait au niveau de l’effectivité. Et ce « mode d’emploi » résulte de la combinaison d’effets idéologiques et de choix dans l’agencement des systèmes répressifs et dans l’allocation des ressources (25). C’est là qu’est le nœud gordien. C’est là aussi qu’il faut faire porter l’investigation sur la perception de gravité. Mais la conceptualisation et l’agencement méthodologique de la recher- che de Newman ne permettaient pas d’y atteindre. Je vais tenter de montrer un peu plus bas comment il est possible d’y parvenir. Auparavant, il faut ajouter que la perception de gravité (relative, d’ordre de priorité) ne règle pas toute la représentation de la loi pénale. Elle correspond — si l’on veut une comparaison juridique — à la partie de droit pénal spécial qui règle les infractions et leurs sanctions. L’adéquation dont on parlait tout à l’heure est mesurée selon une représentation de la délinquance qui combine deux dimensions : la gravité reconnue à l’acte et l'intentionalité imputée à l’auteur ; une investigation complète sur les représentations et attitudes envers la loi pénale ne doit pas borner son investigation à la perception de gravité. Il faut la compléter par une investigation sur la perception de l’inten- tionalité — ce qui correspond aux problèmes de droit pénal général. $ =s* Faute de pouvbir exemplifier ici chacune de ces trois dimensions d’attitudes. je vais, pour finir, tenter de montrer spr un exemple (‘iS)‘lä rècherché——— par ailleurs intéressante de N. Herpin (Réf. 44)—présente à mon avis la faiblesse de miser touté l’explication de la production du système pénal sur des raisons tenant à son agencement (et avec des données parfois fragiles) mais en négligeant totalement les mécanismes idéologiques. — 238 — comment il est possible d’étudier la perception de gravité des infractions en termes de priorités relatives (26). L’instrument utilisé se compose de 65 descriptions en termes « quotidiens » de comportements incriminés par la loi. Selon la métho- dologie mise au point par T. Sellin et M. Wolfgang (Réf. 45) et souvent répliquée depuis, ont s’est assuré d’étudier seulement la dimension d’attitude en excluant des descriptions toute information sur l’intentiona- lité et en les bornant aux conditions et aux conséquences « objectives » des faits. Ces comportements ont été choisis de manière à représenter un éventail aussi diversifié que possible de toute les sortes d’incriminations légales en droit français, y compris celles qui sont consensuelles, conspiratoires ou à victimisation indirecte, etc. Les personnes interrogées étaient invitées à donner une note exprimant la gravité relative attribuée à chaque comportement. Ce n’est pas ici le lieu d’entrer dans une discussion sur les échelles de notation. Je me bornerai à dire que l’on a choisi l’utilisation d’une échelle entièrement libre à la seule exception de l’interdiction du O et des notes négatives dont la signification aurait pu être ambiguë et exprimer une appréciation sur l’adéquation en même temps que — ou à la place de — une appréciation sur la gravité relative. Cet instrument a été administré à un échantillon représentatif de la population française de 15 à 69 ans révolus, tiré selon la méthode des quotas. On a donc recueilli 1 800 séries de notes. Avant de pouvoir utiliser ces données, il était nécessaire — compte tenu du choix d’une méthode à échelle libre — de réduire les amplitudes utilisées à une amplitude unique. Pour cela, on a rapporté les notes données pour chaque item par un individu à l’amplitude de son échelle personnelle. Si l’on appelle : — Mij la note donnée affectivement à l’infraction ; par l’individu : : — mi la note la plus basse utilisée par cet individu ; — Mi sa note la plus haute. On analysera le tableau fi; où Mij_-——— m/ fij = M — mi (26) La recherche ici retenue comme exemple a fait l’objet de diverses publi- cations plus détaillées auxquelles je prends la liberté de renvoyer pour plus «mple information (Réf. 44). —w— Concrétement les notes à analyser seront un rapport à 100, valeur attribuée à la note maximum donnée par un individu. Pour un item donné la notc scra tel pourcentage de l’échelle de gravité. On a appliqué aux données ainsi transformées une méthode de traitement statistique où les relations entre éléments statistiques sont figurées dans des représentations graphiques planes, obtenues par condensation à partir d’une représentation dans un espace à n dimen- sions : les éléments analysés sont représentés par des points et leurs rela- tions par les distances qui les séparent ; les proximités entre ces points permettent d’opérer des groupements dont les oppositions conduisent à l’interprétation des facteurs déterminant la répartition des perceptions d’infractions. Dans l’analyse factorielle des correspondances (27), (Réf. 47), on travaille sur des « profils» et non sur des nombres bruts. Les séries de notes de gravité obtenues par chacun des délits ont été placées en variables actives et les données concernant les caractéristiques socio-culturelles des interviewés en variables supplémentaires. Il était nécessaire d’utiliser comme variables actives tant les antinotes que les notes (28). En effet, dans un tableau qui n’aurait comporté que des notes, les profils des infractions ayant recueilli en moyenne une note élevée se seraient trouvés aplatis par rapport aux profils des infractions ayvant obtenu en moyenne une note basse. Supposons. par exemple. que D+ soit une infraction ayant suscité une forte réprobation et D’+ une infraction jugée mineure entre deux notes données par les individus / et /’ n (( D+) = 10 et n ( D+) = 8 Il y a la même différence qu’entrè | n ( D'+) = et n (ÿ, D'+) = 3 Comme entre 8 et 10, la variation relative n’est que de 20 %, le profil des notes D+ sera presque plat : le point se placera au centre des graphiques. Au contraire, D+ aura un profil au contraste très accusé puisque la variation relative est de 200 % ; le point pourra être à la périphérie des graphiques. Afin de rétablir l’équilibre. il convient de traiter un tableau dédoublé. (27) Je nc m'étendrai pas ici sur la convenance de cette méthode statis- tique au problème ici traité la question est longuement exposée par J.C. Weinberger dans la thèse qu'’il a tirée de cette recherche (Réf. 46). Il est inté- ressant de noter que Newman utilise aussi une méthode d’analyse factorielle. mais sans en tirer peut-être tout ce qu’il était possible d’en obtenir au niveau de l’interprétation des résultats. (28) Les notes indiquent la pari de l’amplitude de l’échelle 100 donnée à une infraction, les antinotes représentent la été donnée à cette infraction. part de l’échelle 100 qui n’a pas — 240 — -Mais cette opération a comme conséquence de donner aux variables une structure particulière que fait apparaître le premier axe factoriel : il.a une valeur propre très élevée (40 % de l’inertie totale) mais tout simplement parce qu'il oppose notes et antinotes, Au contraire, la multiplicité et la grande variété des comportements présents dans le questionnaire aboutit nécessairement à un nuage dont on peut extraire de nombreux facteurs. Ceci explique la faible part d’inertie extraite par chacun des axes d’ordre égal ou supérieur à 2 (5.3 % ; 4,2 % : 3 % : 22 % ; 2 %). Mais il faut insister sur la stabilité des facteurs 2, 3 et 4. On constate, en effet, que chacun d’eux correspond à la contribution de variables nombreuses dont aucune ne contribue de manière massivement disproportionnée. De la sorte, une modication des réponses à tel ou tel ifemn y contribuant ne modifierait pas profondément la nature de cet axe. On peut tirer deux conséquences de cette constatation — en premier, on identifie là trois facteurs manifestant des clivages d'attitudes profonds, — en second, la sphère de non-clivage apparaiît très limitée au niveau des priorités relatives ; seulement 25 % des infractions incluses dans le questionnaire demeurent en dehors des clivages. Certaines ont été jugées pas graves, d'autres moyennement ou très graves ; il s’agit d’une liste assez hétérogène et l’on peut seulement remarquer que l’accord de gravité se fait plutôt sur les comportements pouvant entraîner l’atteinte physique à des personnes, surtout si elles paraissent innocentes ou étrangères à l’affaire. Ceci étant précisé, il est maintenant possible d’indiquer les clivages essentiels sur les priorités relatives dans la perception de la gravité des infractions. “ Un premier clivage s’opère dans la population selon qu’elle est plus ou moins tolérante vis-à-vis de mœurs minoritaires « déviantes », ou même tout simplement nouvelles. Il s’agit en fait d’une attitude de libéralisme Vis-à-vis de comportements qui mettent en jeu le droit de disposer de soi-même, en bien ou en mal. Bref, c’est un clivage autour de la permissivité. La partie « permissive » de la population tolère relativement bien, en particulier, les délits ayant trait à la sexualité, l'avortement (29), la drogue et des manifestations d’opinions politiques. Par contre, la fraction « non permissive » de la population tolère particulièrement mal ces mêmes comportements ; alors qu’elle attribue, en valeur relative, moins de gravité que la première au vol même avec meurtre, à l’escroquerie, au viol même collectif, au rapt d’enfant. Ce qui importe le plus. pour elle, semble être une hostilité aux changements de mœurs, (29) Encore pénalisé à l’époque de l’enquête. — 241 — Dans l'ensemble, les « permissifs » sont plutôt des « cols blancs », des étudiants, plus urbanisés et ayant fait plus d’études que les « non- permissifs ». Les premiers se placent plus à gauche sur l’échelle d’opinions politiques, Les « non-permissifs » sont plutôt des femmes au foyer, des commerçants, des professions libérales et des ouvriers. Un second clivage oppose ceux pour qui l’important est de défendre l'individu dans ses possessions, à ceux pour qui la priorité est surtout une défense collective contre les dangers moraux et physiques mena- çant individus et collectivité, On pourrait appeler les premiers les défenseurs de «l’avoir individuel », et les seconds les tenants de « l’être social ». Cet axe peut être décrit comme celui de la défense de la propriété privée individuelle. Les défenseurs de l’« avoir individuel » jugent très sévèremeni toutes les formes de vols. Par contre, ils acceptent relativement mieux la drogue et l’ivresse, la pollution industrielle, les infractions à la législation sociale et le trafic d’influence. À l’inverse, les défenseurs de l’« être social » tolèrent très mal les infractions à la législation sociale, les abus de pouvoir, l’usage et le trafic de drogue. Ils acceptent, mieux que les autres, les délits ayant trait au vol sous toutes ses formes. Les tenants de l’« avoir individuel » sont plutôt des non-salariés, des ruraux ou provinciaux, ayant un faible niveau d’étude. Ceux de l’«être social» sont principalement des salariés et des étudiants. Enfin, un troisième niveau de clivage oppose ceux qui tolèrent particulièrement mal la délinquance des classes dominantes ou de leurs agents, à ceux qui jugent plus grave la délinquance des marginaux ou des contestataires. Les premiers jugent sévèrement les abus de pouvoir de la police, la fraude fiscale des commerçants et des industriels, la corruption de fonctionnaires, les atteintes à la législation du travail ; ils acceptent par contre, mieux que les autres, l’usage et le trafic de drogue, le vol à main armée, le rapt d’enfant, Pour les seconds, les délits les plus mal tolérés sont l’usage et le trafic de drogue ; par contre, ils acceptent relativement mieux les abus de pouvoir de la police, les infractions au droit du travail, les diverses fraudes des industriels et des commerçants. On pourrait appeler ce troisième axe « intégration sociale » dans la mesure où il traduit l’acceptation ou non de normes en fait destinées à maintenir le système en place et la domination de certains groupes par d’autres. Les premiers sont plutôt des ouvriers, des chômeurs et des jeunes, alors que les seconds se recrutent chez les retraités, les femmes au foyer, les commerçants, artisans et professions libérales. Ces trois niveaux de clivage montrent essentiellement qu la notion de gravité n’est pas unidimensionnelle. Par conséquent, il serait vain de — 242 — chercher une échelle continue de gravité au sein de la société française. Cec axes désignent bien plutôt des niveaux de tensions, des courants de pensée qui traversent la société, des rapports conflictuels plutôt qu’une transformation continue et monotone des normes. Trois groupes ont été essentiellement étudiés la petite bour- geoisie non salariée, la petite bourgeoisie salariée et les ouvriers. La petite bourgeoisie non salariée (professions libérales et petits patrons) se montre globalement peu permissive, peu tolérante envers les mœurs nouvelles ; elle privilégie la défense de la propriété indi- viduelle ; elle est sévère à l’égard de la délinquance des marginaux, relativement plus tolérante aux manquements des puissants. Il n’em- pêche que, sur ce fond général de conformisme, les rmêmes clivages se retrouvent parmi cètte petite bourgeoisie que ceux observés parmi l’ensemble de la population. Sur l’axe de permissivité, les plus intolérants sont les exploitants agricoles (et accessoirement le clergé), les croyants-pratiquants, les gens se disant de droite. De même, ce sont ceux qui sont les plus représentatifs de la France rurale, cléricale et de droite qui sont les meilleurs défenseurs de la propriété individuelle. En ce qui concerne le dernier axe, l’« intégration au système de normes dominant », le clivage semble s’être cristallisé autour de la drogue. Mais ici le tableau se transforme légèrement. La population la plus tolérante à l’égard de cette délinquance est plutôt composée de seoteurs les plus en marge de cette catégorie (retraités, femmes au foyer), les plus en déclin (petits commerçants) mais aussi les plus à gauche ou d’un niveau d’éducation plutôt supérieur à la moyenne. La petite bourgeoisie salariée (cadres, employés) est globalement plus permissive, plus attachée à l’« être social» que la précédente et plus tolérante à l’égard de la délinquance des groupes minoritaires. Mais elle aussi est relativement clivée par les trois mêmes grandes dimensions. Toutefois, si l’on rencontre les mêmes différenciations entre droite-gauche, croyants. pratiquants-non croyants, petites villes-grandes villes, le tableau se complique en raison de la place dans la structure hiérarchique (par aileurs redondante avec le diplôme obtenu), ce qui rend cette catégorie plus ambiguë que la précédente. Les moins permissifs se rencontrent surtout parmi les provinciaux, croyants-pratiquants, se situant plus à droite, mais aussi parmi ceux qui excercent un pouvoir hiérarchique. Les défenseurs prioritaires de à « la fois aux couches Jles la propriété individuelle appartiennent plus aisées et les plus démunies de cette population, à la fois des gens ayant un niveau d’étude supérieur et d’autres qui ont un niveau — 243 — primaire. Parmi ceux qui tolèrent mieux la délinquance des marginaux et minoritaires que celle des puissants, il y a des jeunes, des chômeurs des inactifs, très urbanisés, des gens qui exercent un métier n’imptiquant pas l’exercice d’un pouvoir hiérarchique, mais aussi des gens ayant un niveau d’études élevé Pour cette catégorie se croisent la dimension d’évolution et celle de place dans la structure de pouvoi: En ce qui concerne la classe ouvrière, on retrouve, pour chacun des trois axes, une opposition entre une population plutôt masculine jeune, qualifiée, urbaine, non croyante et non pratiquante, et une population plutôt féminine, plus âgée, moins qualifiée, davantage rurale et provinciale, croyante et pratiquante. La seule note discor- dante est pour le troisième axe, où les personnes les plus âgées se montrent aussi plus tolérantes En résumé, d’importants clivages traversent la société française à propos des incriminations et de leur ordre de priorité. Ajoutons encore qu’il s’agit d’axes stables, durables car ils ne reposent pas sur une querelle ponctuelle, mais manifestent des oppositions intéressant chaque fois de nombreuses infractions. Leur ordonnancement d’'importance n'est d’ailleurs pas démuni d’intérêt. Enfin ils traversent chaque groupe social, même si fondamentalement 1ls les opposent entre eux — Le premier concerne la permissivité vis-à-vis de l’évolution des tous les autres. mœurs ; il oppose la petite bourgeoisie salariée à — Le deuxième regarde la défense de la pr0pnete individuelle et oppose les non-salariés aux salariés. — Enfin le troisième oppose la delmquanœ des marginaux à celle des puissants ; ouvriers (et jeunes) s’y distinguent cla1rement surtout des non-salariés. Si le désaccord n'apparaît guère Vis-à-vis de la répression des atteintes à la personne sur quoi tout le monde semble plus ou moins tacitement se retrouver, règnent ensuite de profondes divergences sur l’attitude de la loi pénale l’égard de l’évolution des cultures et des mœurs, de la défense de la propriété individuelle (30), enfin des deux délinquances antithétiques des puissants et des marginaux (31) De tels résultats peuvent difficilement apparaître dans une recher- che comparaüvc comme celle de Newman (Réf. 49) par exemple, qui se résume à l’administration d’un instrument très standardisé dans six pays différents. ——c se (30) Et pourtant combien de fois n’a-t-on pas dit que tout le monde s’accordait pour voir là une des priorités de la répression ? (31) Le simple fait que le clivage apparaisse de manière stable (encore qu’en troisième) montre quelles contradictions règnent actuellement dans le secteur pénal de l’idéologie acuelle. —w— Je crains toujours qu’une telle manière de procéder fasse sombrer des-travaux comparatifs dans un idéalisme anhistôrique — comme on l’a souvent reproché à des exercices comparatifs juridiques- (Réf. 50). - En tout cas, la nécessaire focalisation sur la perception de gravité en termes d’ordres de priorité y devient quasiment impossible. Et je pense même qu’une telle sorte de recherche comparative a du mal'à réussir les focalisations convenables, également, sur les dimensions d'attitudes concernant l’adéquation sorte de déviance-forme de contrôle social ainsi que sür l’intentionalité imputée à l’auteur. A temps qu'à procéder de manière comparative, il me paraît préférable — sur la base d’une problématique et d’une conceptuali- sation harmonisées dans les différents terrains — de se borner à apparier les démarches et non les instruments en commençant chaque fois, comme on l’a fait au Québec, par répliquer les explorations qua- litatives (Réf. 5/) sans se laisser obséder par la volonté d’une unité d’instruments ultra-standardisés appliqués de plano dans une multitude dç sociéÿés dont les histoires sont profondément différentes. REFERENCË.S ,— Sur les reèherches K.O.L., on peut voir par exemple a) POGDORECKI (A.) e: al. Knowledge and opinion about law, London, Martin Robertson, 1973. b) VERSELE (S.C.) ct al. « Actés ‘du colloque international de sociologie du droit et de la justice », Rev n° spécial. . vlnst, Socio. (U.L.B.). 1970. t> Voir sur ce point : ROBERT (Ph.), LAFFARGUE (B.) et coll. : L'image de la justice pénale dans la société 1977, ronéo. : le sy;rènw pén4 vu par ses clients, Paris. SEPC a) ROBERT (Ph.) et FAUGERÔN (C.) : Représentations du système pénal dans la société française, rapport au colloque ouest-européen du C.N.R.S. sur « Connaissance et fonctionnement de la justice pénale », Lyon, 1977. ronéo. Voir aussi b) op. cit. (2) et ref. cit. 4.— KUTCHINSKY (B.) « La perception de la déviance ; aperçu sur les recherches empiriques », in La perception de ta déviance et de la crimina- lité, Strasbourg, Conseil de l’Europe, 1972. 5.— NEWMAN (G.R.) : Comparative deviance perçeption and ÿlaw im six --cultures, New York, Elsevier, 1976. 6.-—- SAYAG (A.) et TERRE (F. « Connaissance et conscience du droit problèmes de recherche », Année sociologique, 1975, XXVII, pp. 465-496. 7.— Voir par exemple BOURDIEU (P.), CHAMBOREDON (J.C.), PASSERON : Le mélm’ de sociologue, Paris. Mouton, 1973. 8.— Voir MIAILLE (M.) : Une introduction critique au droit, Paris, Maspéro, 1976 9.— op. cit. (3a) 0.— op. cit. (3a) 11.— a) FAUGERON (C.) et ROBERT (Ph,) : « Les représentations sociales de la justice pénale », Cahiers internationaux de sociologie, 1976, LXL pp. 341-366. b) op. cit. (3a) 12.— SELLIN (T) « Groupe de pouvoir, législation et criminalité », Revue de science criminelle, 1960, XL, p. 8 13.— a) ROBERT (Ph) : « La sociologie entre une criminologie du passage à l’acte et une cnmmologne de la réaction sociale », Année sociologique, 1973, XXIV, pp. 441-504. b) ROBERT (Ph.) : « Recherches en criminologie de la réaction sociale », in Neue Perspecktiven in der Kriminologie, Zürich, Verlag der Fach- vereine an den Schweizerischen Hochschulen und Techniken, 1975 pp. 55-83 c) ROBERT (Ph.), LAMBERT (Th.) et FAUGERON (C.) Image du reconstruction Genève-Paris, Médecine et viol collectif et d'objet Hygiène-Masson 1976, chap. 1 Les handes d'adolescents, une 14.— ROBERT (Ph.) et LASCOUMES (P.) théorie de la ségrégation, Paris, Editions ouvrières, 1974, chap. 5 1S. — Sur ces points voir par exemple a) ROBERT (Ph.) : « Les statistiques criminelles et la recherche, réfilexions conceptuelles », Déviance et Société, 1977, I, 1, pp. 3-27, et ref. cit b) BOTTOMLEY (A.K.) et COLEMAN (C.A. « Criminal statistics : the police role in the discovery and detection of crime », Internationaf Journal of crim. and pen 1976, IV, pp. 33-58 16.— op. cit. (S$) 17. — DAHRENDORF (R. Classes et conflits de classes dans la société industrielle, Paris Mouton, 1972 18.— TURK (A.) Criminality and the legal order, Chicago, Rand Mc Nally, 1969 19.— a) CHAMBLISS (W.J.) (Ed) : Mc Graw Hill, 1969. Crime and the legal process, New York b) CHAMBLISS (W.J.) and SEIDMAN (R.B)) : Law, order and power, Reading, Mass., Addison Wesley, 1971. 25.— a) QUINNEY (R) The social reality of crime, Boston, Little Brown and Co 1970 b) QUINNEY (R) The problem of crime, Dodd, Mead and Co., 1970 c) QUINNEY (R. Critique of legal order, Boston, Little Brown and C 1973 214— op. cit. (19b) — 246 — 22.— MEAD (G.H) Mind, self and society, Chicago, University of Chicago Press 1934 23.— PLATT (A. « Prospects for a radical criminology in the United States », Crime and social justice, 1974, I — a) SCHWENDINGER (H.) and SCHWENDINGER (J.) « Defenders of order or guardian of human rights », /Issues in criminology, 1970 V, 2, pp 123-158 b) SCHWENDINGER (H.) and SCHWENDINGER (J.) « Continuing debate on the legalistic approach to the definition of crime ». Jssues in criminology, 1972, VII, 1, pp. 71-82 25,— VAN OUTRINE (L « Interactionnisme et néo-marxisme, une analyse critique », Déviance et société, 1977, I, p. 3 26 — op. cit. (25) 27 — POULANTZAS (N.) Les classes sociales dans le capitalisme d'aujour- d’'hui, Paris, Seuil. 1974, p. 314 28 op. cit. (19b) 29 op. cit. (13) op. cit. (19a) N — CHAPMAN (D : Socwlogv and the stereotype of the criminal, London. Tavistock, 1968. 32 — Pour reprendre le titre de l'ouvrage de CHEVALIER (L Classes laborieuses, classes dangereuses, Paris, Plon 1958 33 — Mc DONALD (L.) The sociology of Law and order, Montréal, Book Center 1976 — LOYOLA (I. de) Exercices spirituels, Paris, Gigord, 1948, p. 136 35. — ROBERT (Ph.) : Recension des op. cit. (S) et (33) in Rev. can. crimino,, 1977, XIX. 4, pp. 398-408. 36.— op. cit. (5) 37.— SOROKIN (P.A.) Social and cultural dynamics New York, Amer Book 1937 38.— COHEN (AI) La déviance, Gembloux, Duculot 1971 39.— a) BECKER (H.S.) Outsiders, New York, The Free Press of Glencoe, 1963. b) BECKER (HS) (Ed) : The otherside, New York. The Free Press of Glencoe. 1 40 Dans un sens analogue, voir les remarques de WEINBERGER (J.C.) : Idéologie et gravité attribuée aux comportements incriminés, Bordeaux. Université de Bordeaux I, 1975, ronéo (thèse) 41.— op. cit. (13a) 42.— a) RIBORDY (F.X.) Réaction sociale à la déviance, Varèse. Journées inter. socio. droit. 1971. ronéo _— 247 — b) NGUYEN THI HAU : La réaction sociale à la déviance au Canads, Montréal, 1972, ronéo. c) VERSELE (S.C) : « La perception de Ja démmce et de la cnmmahté ». in op. cit. (4). 43 FAUGERON (C.) et al. - De la déviance cl du confrole social, repré— sentatzons et attitudes, Pans D.GR. S.T. 1975, ronéo. 44 HERPIN (N) : Lappl:calmn de la loi Pans Seuil, 1976 45 a).op. cit. 40) b) WEINBERGER (J.C.), JAKUBOWICZ (P.) et ROBERT (Ph.) « T1 declino del diritto… como strumento di controllo sociale », Questione criminale, 1976, IL, 1, pp. 73-96. c) WFINBERGER (JC) JAKUBOWICZ (P.) et ROBERT (Ph) « Société et gravité des infractions », RS.C., 1976, 4, pp. 915-930. 46 SELLIN (T.) and WOLFGANG (M.ÆE.) The measurement of: delin- quency, New York Wilev 1964 47 op. cit. (40) 48 BENZECRI (J.P.) L'analyse des données, Paris, Dunod. 1973 49 op. cit. (5) 50 op. cit. (8) 51 LEVY-(R.) e ZAUBERMAN (R.) « Les enseignements d’une. recherche exploratoire l’image de la justice pénale au Québec»,. Déviance et Société, 1977 | 4. pp. 411-426 Imprimerie administrative de Melun 2281 - 1978