Nous contacter|Plan du site|Mentions légales|Accès réservé

Assessing Deviance, Crime and Prevention in Europe
Version française PDF

Brochure n°2bis - Self-Reported Delinquency Surveys in Europe

Marcelo F. Aebi

Full text

I - Introduction

In the framework of the 6th FPRTD, the European Commission financed the coordinating action Assessing Deviance, Crime and Prevention in Europe (CRIMPREV). This coordinating action includes different thematic workpackages. Workpackage 7 is devoted to Methodology and Good Practices, and among the activities included in that workpackage there was a workshop on Self-Reported Delinquency (SRD) Surveys in Europe.

The team for that workshop was composed by the following national correspondents and a general rapporteur:

- Lina Andersson (Department of Criminology, Stockholm University) for Sweden

- Cécile Carra (CESDIP – CNRS - IUFM du Nord/Pas de Calais – Université d'Artois) for France

- Thomas Görgen (German Police University, Münster) and Susann Rabold (Criminological Research Institute of Lower Saxony) for Germany

- Janne Kivivuori (National Research Institute of Legal Policy, Finland) for Finland

- Susan McVie (School of Law, University of Edinburgh, Scotland) for Ireland and the United Kingdom.

- Lieven Pauwels (Universiteit Gent) and Stefaan Pleysier (Research Centre in Security, Safety and Society’ at KATHO University College, Institute of Criminology, Catholic University of Leuven) for Belgium and the Netherlands.

- Simona Traverso, Giada Cartocci, and Giovanni Battista Traverso (University of  Siena, Italy) for Italy.

- Marcelo Aebi (University of Lausanne, Switzerland), general rapporteur.

The national correspondents prepared reports on SRD surveys in their countries. Then, a three days seminar took place in Paris from 17th to 19th January 2008 and was attended by the promoters of CRIMPREV, most of the national correspondents and the general rapporteur. Seven reports covering twelve countries were presented. The countries included were Belgium, England, Finland, France, Germany, Ireland, Italy, the Netherlands, Northern Ireland, Scotland, Sweden and Wales. The seminar gave the participants the opportunity of presenting and discussing the reports. Then the general rapporteur prepared an intermediate report of the situation and asked the national correspondents to introduce minor modifications to their papers according to the discussions that took place during the seminar. The final versions of the national reports were then handed over to the general rapporteur who established a first version of this final report that was discussed with the promoters of CRIMPREV in a meeting that took place in Bologna on 9 July 2008 and sent for comments to the national correspondents.

Accordingly, this report is based on the national reports, the discussions that took place during the Paris seminar and the Bologna meeting, and the comments of the national correspondents, as well as on bibliographical research conducted by the general rapporteur. As a consequence, it includes also references to SRD surveys conducted in countries not represented in the workshop.

This report is focused on general SRD studies, but surveys conducted in order to measure specific behaviours such as bullying or drug use are also covered. It includes a discussion of definitional issues, a short synthesis of the national reports, a historical overview of the development of SRD studies, a discussion on methodological issues related to that measure of delinquency, as well as an analysis of the impact of SRD surveys on criminological theories and criminal policies.

II - Definitional issues

Self-reported delinquency surveys are studies in which people - usually juveniles - are asked to reveal information about their delinquent behaviours. However, the terminology may be misleading for two reasons. First of all, because respondents give information not only about delinquency but also about their life-style in general, their attitudes toward different subjects, their families, their school, their friends, and many other socio-demographic factors. Thus, one could consider that delinquency is the dependent variable in such surveys, and that respondents give also a lot of information about independent variables that are supposed to be related to delinquency. The second reason is that the concept of delinquency may also be misleading. Needless to say that delinquency, as any other concept, is a social construct. However, it must be mentioned that the concept of delinquency used in the criminological literature is a very broad one. In fact, many of the behaviours included in a self-reported survey are not criminal offences in most European countries. Such behaviours include, for example, running away from home, skipping school or fare-dodging. In this report the term delinquency is used in that broad sense, and therefore it includes all sorts of antisocial or deviant behaviour, even if these behaviours are not defined as an offence in the criminal law.

In that context, it is interesting to point out that in countries that use languages derived from Latin (for example Catalan, French, Italian, Spanish, and Portuguese) the term delinquency has the same root as the word used to describe a criminal offence. For example, in Spanish we find “delincuencia” (delinquency) and “delito” (offence), in French “délinquance” and “délit”, and in Italian “delinquenza” and “delitto”. For that reason, in these countries the word delinquency is immediately related to behaviours forbidden by the criminal law. On the contrary, the Webster’s New Universal Unabridged Dictionary, defines delinquency as “wrongful, illegal, or antisocial behavior”. Thus, in English speaking countries, the common understanding of the concept of delinquency includes a wider series of acts ranging from deviant or antisocial behaviour to criminal offences.

This different evolution of the term delinquency - which is derived from the Latin delinquere (to do wrong)1 -, might be related to the fact that England developed a system of common law - based on custom or court decision instead of written law - and exported it to its former colonies, while Continental European countries followed typically the civil or Roman law system which is based on the written law and led to the introduction of criminal codes and a specific terminology related to them. In this context, the publication of Dei delitti e delle pene - translated into English as On Crimes and Punishments - by Beccaria in 1764 is particularly important. That book inspired many reforms in criminal justice throughout Europe and had a strong influence on the intellectual leaders of the French revolution. Thus, in 1795 - year IV according to the calendar introduced after the revolution -, the (French) National Convention approved its Code of Crimes and Punishments, whose name (Code des délits et des peines) took up the title of Beccaria’s book. Indeed, it was under that same name that the Criminal Code (Code pénal) was first introduced by Napoléon Bonaparte in 1810. That code is considered as the first criminal code and it inspired most of the European codes approved during the following years. In that way, in Continental Europe, the term delinquency was definitely linked to a violation of the criminal law.

  • 1  According to the Webster’s New Universal Unabrid(...)

Nevertheless, the term self-reported delinquencystudy or self-reported delinquencysurvey - coined in the United States of America, where the first surveys took place - was translated literally to the different European Latin languages2, creating thus some confusion about the contents of such a survey.

  • 2  For example, encuesta de delincuencia autorrevel(...)

This confusion was aggravated by the fact that some of the behaviours included since the first American3 SRD surveys are considered as (juvenile) status offences in most States of the United States of America. These are offences based on the personal condition (the status) of the offender. This means that the same behaviour is not an offence when it is done by an adult, but it is an offence when it is done by a minor. Status offences include behaviours such as truancy, running away, tobacco smoking or underage consumption of alcohol. As they constitute a sort of offence in the USA, American researchers are right when they state that the behaviours include in their surveys constitute violations of the criminal law, but the situation is completely different in Europe, where such behaviours are not considered as offences.

  • 3  In this report we follow the conventional use of(...)

III - On the development of SRD surveys in Europe

In this chapter, we present a brief summary of the national reports, an overview of the International self-reported delinquency study (ISRD) and the European School Survey Project on Alcohol and Other Drugs (ESPAD), as well as an outline of self-reported delinquency studies in other European countries.

    1 - SRD surveys in Finland4

    The fieldwork for first Finnish survey on SRD took place in 1962 and the results were published in the mid-1960s. That survey was part of the Nordic Draftee Research Program which included other Scandinavian countries in which similar surveys were conducted in the 1960s. The technique was seldom used from the mid-1960s to the 1990s. However, since then, several surveys took place. In particular, Finland was the only Nordic country to participate in the first ISRD in 1992 and has also participated in the second one in 2006. In both cases, the survey is based on a city sample (Helsinki), but in the meantime the country has developed a series of national surveys.

    • 4  This chapter is based on Kivivuori (2008 and for(...)

    Thus, in 1995, Finland launched the Finnish Self-Report Delinquency Study (FSRD) that is conducted periodically in schools with samples of 9th grade students. The latest available results refer to 2004 (The FSRD-2008 data has recently been connected but the results are not yet available). Moreover, since 2000/2001, the Finnish School Health Survey includes also some questions on SRD. This survey is a large-scale one and it includes results at the municipality level. An analysis of the trends from 1995 to 2004 shows a drastic decrease of property offending (especially of shoplifting), a relative stability of violent offences and an increase in soft drug use. Computer related offences were not included in the FSRD but results from the ISRD-2 suggest that there are relatively common. Taking into account that currently juveniles spend a lot of time in front of their computers, it is possible to imagine a displacement from property offences in public places to computer offences.

      Apart from that, the country has conducted in 2006 a Young Male Crime Survey (YMCS) based on the same questionnaire that was used in 1962 for the Nordic Draftee Research Program. The sample is composed by young males that attend their pre-military screening and therefore are slightly older than the adolescents included in most European samples. A comparison of the results obtained in 1962 and 2006 shows a decrease in workplace theft, buying or selling stolen goods, and theft from a car; an intriguing stability in shoplifting, and an increase in drunken disturbance in public places, drunken driving and bicycle theft. These trends can be explained by changes in the opportunity structure including a late arrival of juveniles to the working market, an improvement of the economic situation of the country, and an increasing in the availability of alcohol.

        Finland has also participated in the Mare Balticum Youth Victimisation Survey that took place in 2002/2003. The questionnaire used for that survey included a SRD scale. Finally, the country also participates in the ESPAD surveys.

          In sum, one can say that SRD surveys have been institutionalized in Finland and constitute a usual measure of delinquency. As a consequence, they are playing a role in the development of criminal policies. For example, the Ministry of Justice used self-reported data it in its estimation of the crime situation and crime trends in the country. In addition, the results of the FSRD were taken into account by the committees that reformed the law concerning young offenders and for the planning of the Finish national violence reduction program.

            2. SRD surveys in Sweden5

            As in the case of Finland, the origins of SRD surveys in Sweden are linked to the Nordic Draftee Research Program. The fieldwork for the first survey was conducted at the end of the 1950s and the first results were published in 1960. A few other surveys were conducted during the 1960s and the early 1970s. The technique was somehow abandoned in the mid-1970s but, since the beginning of the 1990s, the country is conducting regular SRD studies.

            • 5  This chapter is based on Andersson (2008 and for(...)

            Thus, in 1995, Sweden conducted a national SRD survey and since 1999 that survey is run every second year. The sample is national and varies from 5,000 to 10,000 juveniles attending the 9th grade. An analysis of the trends from 1995 to 2005 shows a decrease of theft and vandalism, and a relative stability of violent offences and drug use.

              A comparison of the first and the latest surveys shows a decrease in response rates from almost 100% in the 1960s to 80% in 2006. At the same time, it must be mentioned that Swedes researchers has paid a lot of attention to methodological issues, running reliability and validity tests that included measures of the effects of teacher and researcher supervision, anonymity and non-anonymity of the respondents, test-retest controls, observation of the attitudes of pupils during the filling of the questionnaire, tests of different versions of a questionnaire, and follow-up interviews.

                Sweden has also participated in the second ISRD with city samples. Apart from that, local or regional SRD surveys are conducted regularly with relatively large samples of high school students. Surveys on drug use among youth are also regularly conducted using the self-report technique.

                  All in all, Sweden is a country where SRD surveys are institutionalised and represent currently a typical measure of crime. As a consequence, results from national surveys are used in the political debate on crime and crime prevention. Their influence on public policies can be seen mainly at the local level.

                    3 - SRD surveys in the United Kingdom and Ireland6

                    Since the early 1960s, thirty major SRD studies - collecting data on more than 140,000 individuals - have been conducted in the United Kingdom and Ireland. Apart from that, there have been several local and regional studies. Northern Ireland, England and Wales participated in the first ISRD, and Ireland, Northern Ireland and Scotland participated in the second one.

                    • 6  This chapter is based on McVie (2008 and forthco(...)

                    The majority of the self-reported studies took place in Great Britain - more precisely in England - and probably the most well known is the Cambridge Study on Delinquent Development which spans over a 40 year period (1961-2004). Leaving aside that longitudinal study, two other major cross-sectional surveys were conducted in the 1960s in England, Wales and Scotland (1963) and London (1967). The technique was seldom used in the 1970s and the 1980s when surveys were conducted only in Sheffield (1975), England and Wales (1983), and Scotland (1989). Then, since the 1990s, there was a sudden explosion of the number of SRD surveys. Thus, new longitudinal studies such as the Peterborough Adolescent and Young Adult Development Study, the Edinburgh Study of Youth Transitions to Crime, and the Belfast Youth Development Study were launched. At the same time, in the 2000s, SRD surveys were introduced in the Republic of Ireland and institutionalized in England and Wales with the introduction of the Offending Crime and Justice Survey that is a repeated cross-sectional survey with a partial longitudinal panel design and a sample of 12,000 individuals. The same is true for Northern Ireland, where the Northern Ireland Crime and Justice Survey has a repeated cross-sectional design with a sample of 3,000 to 3,500 individuals including (in 2005) persons living in private households, offenders on probation, and offenders in custody.

                      In the United Kingdom, the increase in the number of SRD surveys may be due to the fact that the Central Government - that is the main provider of financial support for such studies - changed its attitude towards crime with campaigns such as “tough on crime, tough on the causes of crime”. The idea was to develop an “evidence-based” approach and, in that context, empirical data was clearly needed.

                        All in all, the number and diversity of SRD surveys in the United Kingdom is impressive. As a consequence, the information provided by the available surveys is a clear source of inspiration for public policies. Thus, the Cambridge Study on Delinquent Development had a strong influence on policy makers and inspired partially the reform of the juvenile justice system putting the accent on the early detection of problematic behaviours and ineffective parenting practices. Apart from that, SRD studies conducted in the United Kingdom have often been used for testing the validity and reliability of this measure of crime, as well as for the development and testing of criminological theories.

                          4 - SRD surveys in Germany7

                          In Germany, the first SRD surveys were conducted in the late 1960 s and the early 1970s with local or regional samples.

                          • 7  This chapter is based on Görgen and Rabold (2008(...)

                          Like in most countries, the vast majority of the German research based on SRD surveys, focuses on adolescent populations. However, since 1980, the German General Social Survey (ALLBUS) is conducted every two years and, from 1990 on, that survey includes four items on SRD. As the survey is based on a national sample of the German population, the respondents are mainly adults. Following the Swiss model, the self-report technique has also been used with adults for the evaluation of the involvement in delinquency of the participants in heroin prescription programs.

                            As the first SRD surveys were conducted in different regions and using different methodologies, their results were not easily comparable. Nevertheless, during the 1990s, the Criminological Research Institute of Lower Saxony (KFN) developed a SRD questionnaire that has been used since then in many German cities. Moreover, in some of these cities, the KFN survey is being used on a regular basis.

                              Germany is also one of the few European countries where longitudinal studies based on SRD surveys are available. Such studies are taking place at the local level in different cities, and some of them include comparisons with official data. Most of these studies started in the 2000s but between 1977 and 1996 a longitudinal survey followed a group of 399 children from 13 to 25 years old. Finally, Germany has also participated in the second ISRD.

                                In sum, Germany has a long tradition of SRD studies but the technique is not institutionalized yet. Surveys are organized at the local or regional level and, although the German Ministry of Interior has recently funded a big survey, not all the German States took part on it. Given the federal organization of the country, it is difficult to foresee if national surveys will be conducted in the future. As it happened in Switzerland, the positive results of the heroin prescription programs, measured through SRD surveys, had a strong influence on the German drug policy. Apart from that, SRD surveys did not have a strong impact on criminal policies yet, although they are regularly quoted in the Periodic Security Reports published by the German government.

                                  5 - SRD surveys in the Netherlands8

                                  SRD surveys were introduced in the Netherlands in the late 1960s and quite a few surveys were conducted in the following years. The country has also participated in the first and the second ISRD.

                                  • 8  This chapter is based on Pauwels, Pleysier (2008(...)

                                  Currently, the Scientific Research and Documentation Center (WODC) conducts systematic SRD surveys studies with representative samples of Dutch adolescents. The survey is called the WODC monitor and it takes place every second year. Apart from that, the Netherlands Institute for Social Research (SCP) and the Dutch Institute for Budget (NIBUD) are also financing SRD studies.

                                    Since the middle of the 1980s, SRD studies have been used to test different criminological theories including differential association, social disorganization theory, strain theory and social bonding theory. The effects of poverty, peers and neighbourhood were also studied. Apart from that, some researchers have tested the reliability and validity of SRD studies and their use with adult samples.

                                      As one can see, SRD surveys are institutionalized in the Netherlands and are having some influence not only in the academic field but also in the development of public policies.

                                        6 - SRD surveys in Belgium9

                                        In Belgium, the first SRD survey was conducted in 1976 with a local sample. In the 1980s two other surveys were conducted in both linguistic areas of the country. The country participated with a city sample (Liège) in the first ISRD and with samples from both linguistic regions in the second one.

                                        • 9  This chapter is based on Pauwels, Pleysier (2008(...)

                                        Since the 1990s no systematic large-scale representative self-reports have been conducted, but there is research usually based on urban samples. Thus, two surveys based on large scale samples have been conducted in the Flemish region and in Brussels in the 2000s.  However, the recent creation of a platform on adolescent research by three institutions may lead to a more systematic implementation of SRD studies.

                                          In sum, SRD studies are not yet institutionalized in Belgium and, apparently, they are not playing a major role in the development of criminal policies. Nevertheless, they have been used to test and develop criminological theories.

                                            7 - SRD surveys in Italy10

                                            Leaving aside a small SRD survey (N=198) conducted in Milano and published in 1980, the history of SRD surveys in Italy is strongly related to the ISRD project and the leadership of Umberto Gatti, professor at the University of Genoa, who has coordinated the participation of the country in the first two waves of that survey. Italy participated in the first ISRD with a sample of three cities (Genoa, Messina, and Siena), and the questionnaire was used again in Siena in the mid 1990s. The country has also participated in the second ISRD with a sample of 15 cities (N=7,278). In 2006, a self-reported survey was also used to measure the involvement in delinquency of foreign youth living in Italy.

                                            • 10  This chapter is based on Traverso, Cartocci, Tra(...)

                                            Research on bullying has been conducted in different cities and regions since the middle of the 1990s using self-reported surveys. In this context, sometimes the ISRD questionnaire was combined with a specific questionnaire on bullying. The self-report technique has also been used to measure drug and alcohol use among juveniles since the 1980s.

                                              In sum, SRD surveys are not institutionalized in Italy and their findings are mainly used by the scientific community. As a consequence, they do not play a role in the development of national public policies; nevertheless, there is evidence that they have been used at the local level in the city of Sienna.

                                                8 – SRD surveys in France11

                                                Among the countries included in this overview, France was the latest to introduce SRD studies. The first study of this kind was conducted in 1999 - and published in 2001 - with a sample of two cities. The questionnaire was based on the one used for the first ISRD. In 2003, a second survey was conducted using the same methodology in one of the two cities surveyed in 1999. Finally, in 2006, the country participated in the second ISRD.

                                                • 11  This chapter is based on Carra (2008, with refer(...)

                                                Violence at school has also been measured using the self-report technique since the mid-1990s. In that case, the survey was based on a questionnaire focused on victimisation, but it included also some questions on self-reported violence. The same questionnaire was later used in other countries, providing thus the possibility of performing some cross-national comparisons. A recent research is based on open questions about violence at school which are later recoded by the researcher.

                                                  One can easily see that SRD studies are not at all institutionalized in France. The findings from the very few surveys available are used by the scientific community, but they do not play a role in the development of public policies. The only exceptions are surveys on drug abuse which were quoted in the new law on crime prevention introduced in 2007.

                                                    9 - The International Self-Reported Delinquency Study (ISRD)

                                                    A - The ISRD-1

                                                    The origins of the ISRD project can be traced back to a meeting of experts that took place in 1988 in the Netherlands and led to the publication of a book on issues related to cross-national research based on the SRD technique (Klein, 1989). Some of the experts that assisted to that meeting decided to launch the project, which started formally in 1990 and was coordinated by the Research and Documentation Centre of the Dutch Ministry of Justice (WODC), headed at that time by Josine Junger-Tas.

                                                      The first survey (ISRD-1) had three main objectives: to compare prevalence and incidence of delinquent behaviours across countries; to contribute to the explanation of (differences in) delinquent behaviour; and to contribute to the solution of methodological problems related to cross-national research (Junger-Tas, 1994a, 2). It was based on a common questionnaire developed by a steering committee of researchers in criminology that was translated into each national language12. The theoretical framework of the questionnaire was heavily inspired by social bonding theory (Hirschi, 1969) while the questions on delinquency had similarities with the ones used in the National Youth Survey and the Denver Youth Study in the United States (Elliott, Huizinga, 1989)13.

                                                      • 12  The original English questionnaire can be found (...)
                                                      • 13  According to Moffitt et al. (1994, 358), 80% of (...)

                                                      Twelve countries participated in the study14. Four of them (England and Wales, the Netherlands, Portugal, and Switzerland) used national random samples; Spain participated with a stratified urban sample; Germany, Greece and Northern Ireland participated with city samples (Mannheim, Athens, and Belfast, respectively); Belgium (Liège), Finland (Helsinki), Italy (Genoa, Messina, and Siena) and the United States (Omaha) participated with school city samples. With the exception of the countries that used school samples, participants in the survey were aged 14 to 2115. Data collection took place in 1991 and 1992 and was based on face-to-face interviews with the exception of the countries that used school samples, where the questionnaire was self-administered16.

                                                      • 14  Our presentation of the ISRD-1 is based on the m(...)
                                                      • 15  Belgium applied a mixed approach combining a sch(...)
                                                      • 16  However, Italy used a school sample, but the que(...)

                                                      The ISRD-1 database is not freely available for researchers yet. A short synopsis of the preliminary results was presented by Junger-Tas (1994b, 379), but it took more than ten years to have a detailed presentation of the main findings in a publication by Junger-Tas, Haen Marshall, Ribeaud (2003) that are summarized here:

                                                        - Lifetime prevalence rates are surprisingly similar across countries. However, looking at the nature of the offences, there are clear cross-national disparities.

                                                          - Property offences are highest in the Northwest-European countries and the USA (Omaha) supporting the idea - based on opportunity theory - that these rates would be highest in the most prosperous countries.

                                                            - There were high rates of violent offences in the USA, England and Wales, Spain and Finland (Helsinki).

                                                              - High rates of mainly soft drug use were found in England and Wales, the USA, Northern Ireland (Belfast) and Switzerland. The Netherlands, that apply a tolerant policy on possession and use of soft drug use, occupied a medium position on the scale of drug use rates.

                                                                - Delinquent behaviour peak between age 14 and 18 across countries. The peak age was 15 in the Anglo-Saxon cluster and 16 in the other clusters.

                                                                  - The peak age vary according to the offence. For example, the peak age of vandalism was clearly lower (between age 14 and 15), while that of drug use was considerably higher.

                                                                    - Age of onset vary between clusters. The lowest age of onset for vandalism was found in Southern Europe, while the lowest age of onset for property offences was found in Northwest Europe. Violent offences and drug use had lowest age of onset in Anglo-Saxon countries.

                                                                      - Serious delinquents tend to start offending at an earlier age than their less seriously-involved counterparts.

                                                                        - Both prevalence and frequency of offending are lower for females than for males. However, gender disparities vary according to offense type and regional country-cluster. Gender disparity is smallest for property offending and drug use while it is largest for violent and serious offences. In addition, the disparity is smaller in Northwest Europe compared to other clusters, which may be related to the high rates of property offences and vandalism found in that cluster.

                                                                          - Family break-up has important effects only as far as it results in father absence. However, father absence is closely related with delinquent behaviour in the Anglo-Saxon cluster and in South Europe, but not in Northwest Europe.

                                                                            - Relationship with parents is not related to overall delinquency, but is related to serious delinquency and to drug use in all clusters.

                                                                              - In all countries there is tighter control on females than on males. Compared to males, the relationship of females with their parents deteriorates somewhat more with age. On the other hand, girls participate considerably more in family outing than do boys.

                                                                                - Truancy and disliking school are related to all types of delinquency in all country clusters. Having to repeat a grade, while not being related to petty delinquency, does appear to be related to violence, serious delinquency and drug use for males, but not for females.

                                                                                  - Peer group membership is related to age, enrolment in school, disliking school, parents’ informal control and not participating in family outings. Delinquents typically spend most of their leisure time with the peer group, while non-delinquents spend much more time with the family (Junger-Tas, Haen Marshall, Ribeaud, 2003, 139-142).

                                                                                    B - The ISRD-2

                                                                                    A first meeting of experts interested in participating in the second wave of the ISRD took place during the 2003 Annual Conference of the European Society of Criminology, in Helsinki. A steering committee was established and, during the following months, the research design and the revised questionnaire were established. The questionnaire was then translated into each national language.

                                                                                      Data collection took place in 2006 in thirty countries17 - of which twenty-four were European - and a first publication including national chapters is attended for 2009. In the meantime, some countries have already published their national findings.

                                                                                      • 17  Armenia, Aruba, Austria, Belgium, Bosnia-Herzego(...)

                                                                                      The ISRD-2 survey is based on school samples of at least 2,000 students - attending 7th to 9th grade - per country. The main sampling design is city based with a minimum of five cities including a metropolitan area, a medium sized city and three small rural towns. Such a design allows multi-level HLM analyses (Raudenbusch, Bryk, 2002) using students as first level units and cities as second level units. Nevertheless, some countries used national random samples drawn from lists of all school classes in the country. These countries oversampled some urban or rural areas in order to make comparisons possible with the rest of the countries.

                                                                                        The majority of countries used self-administered paper and pencil questionnaires filled by the students in their classrooms, however some countries (e.g. Switzerland) used computer assisted personal interview (CAWI) or computer assisted self-administered interview (CASI)18.

                                                                                        • 18  The different techniques for administering SRD s(...)

                                                                                        The ISRD-2 questionnaire can be seen as a revised version of the one used for the ISRD-1. This is completely logical, as one of the goals of the project is to achieve comparability between both surveys. However, there are some major modifications. As far as offences are concerned, the questionnaire includes new ones (e.g. computer related offences) and the wording of some questions has been modified. The instrument includes also a short victimisation survey. The theoretical framework keeps variables related to social bonding theory, but the new questionnaire includes a summary of the Grasmick self-control scale (Grasmick et al., 1993) that allows testing the general theory of delinquency - or self-control theory - developed by Gottfredson and Hirschi (1990), as well as questions related to social learning theory and lifestyle variables that allow testing opportunity-based theories. It also asks for information about life events as well as school and neighbourhood context. The survey included also a short questionnaire for the interviewers about the context in which the administration of the survey took place, as well as a questionnaire for the teachers of the classes being surveyed.

                                                                                          Most countries used Epidata software to input data from the questionnaires. In that context, the labelling of the variables was standardized, simplifying thus the construction of an international database that will be a major tool for European research in the future.

                                                                                            As a final consideration about the ISRD project, it must be mentioned that even if the survey has not been institutionalized at the European level - both the ISRD-1 and the ISRD-2 were financed in each country by local sources or eventually with the financial help of another European country19 -, the creation of a steering committee and a big consortium of researchers is a guarantee that the ISRD project will continue in the future.

                                                                                            • 19  For example, Switzerland funded partially the pa(...)

                                                                                            10 - The European School Survey Project on Alcohol and Other Drugs (ESPAD)

                                                                                            The success of self-reported studies on substance abuse (tobacco, drugs and alcohol) is surely due to the fact that even the most survey sceptic person must admit that official statistics do not provide any valid information on the extent of such use. In that perspective, in the mid 1980s, the Council of Europe created a group of experts that started working on a common questionnaire on substance abuse. Building on the experience of that group, in 1993, the Swedish Council for Information on Alcohol and Other Drugs (CAN) initiated the European School Survey Project on Alcohol and Other Drugs (ESPAD). The first survey took place in 1995 and data have been collected every four years since then. The latest available results refer to the 2003 survey. Results from the 2007-8 survey will only be available by the end of 2008.

                                                                                              The ESPAD project has received financial support from the Pompidou Group at the Council of Europe, the Swedish Ministry of Health and Social Affairs and the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA). However, data collection in individual countries is funded by national sources. The number of countries taking part in the ESPAD project has increased over the years. There were 23 countries in 1995, 30 in 1999, 35 in 2003 and 43 in 2007-8. Indeed, all the countries represented at the Paris workshop on SRD surveys are participating in it.

                                                                                                The survey is based on a common questionnaire that covers tobacco, alcohol, and illicit drugs use. Data are collected through group-administered questionnaires in schools. Samples are composed by 15-16 year old students, although some countries include also students aged 17-18. With only a few exceptions, samples are nationally representative.

                                                                                                  For researchers, one of the big advantages of the ESPAD project is that the main results are available on a free website [www.espad.org] where it is possible to download reports and generate key results graphs.

                                                                                                    Summarizing the main results of the 2003 SPADE survey, Hibell et al. (2004, 22) state:

                                                                                                      - The pattern of alcohol consumption reveals that frequent drinking is most prevalent among students in the western parts of Europe, such as the British Isles, the Netherlands, Belgium but also in Austria, the Czech Republic and Malta. Very few students in the northern parts of Europe drink that often.

                                                                                                        - Beer consumption is most prevalent in Bulgaria, Denmark, the Netherlands and Poland, while wine consumption is most prevalent in typical wine producing countries such as Austria, the Czech Republic, Greece, Italy, Malta and Slovenia.

                                                                                                          - The consumption of spirits is less uniform, with high prevalence rates in as disparate countries as the Faroe Islands, Greece, Ireland, Isle of Man, Malta and the United Kingdom.

                                                                                                            - The prevalence of drunkenness seem to be most concentrated to countries in the western parts of Europe, such as Denmark, Ireland, Isle of Man and the United Kingdom. Very few students report frequent drunkenness in Mediterranean countries such as Cyprus, France, Greece, Portugal, Romania and Turkey.

                                                                                                              - The illicit drug use is dominated by use of marijuana or hashish. Frequent use is mainly reported from countries in the central and western parts of Europe, where more than one third of the students have used it. The high prevalence countries include the Czech Republic, France, Ireland, Isle of Man, Switzerland and the United Kingdom. The low prevalence countries are found in the north as well as the south of Europe (Hibell et al., 2004, 22).

                                                                                                                - As far as the trends between 1995 and 2003 are concerned, Hibell et al. (2004, 128) conclude:

                                                                                                                  - To sum up, the trend development over the 8 years of the ESPAD history is indicative of the fact that smoking remains at about the same level or decreased in a majority of the countries.

                                                                                                                    - With regard to alcohol an unchanged or a somewhat decreasing consumption was observed in the western parts of Europe while increases mainly were found in the eastern parts.

                                                                                                                      - The use of drugs is still dominated by the use of cannabis. The high prevalence countries in 1999 are still at the top in 2003, but a clear increasing tendency can be observed in the eastern parts of Europe. It is also clear that an increasing number of students in many European countries find cannabis easily available.

                                                                                                                        4 - Historical overview of SRD surveys

                                                                                                                        After having presented an overview of SRD surveys in different European countries and some European projects based on the self-report technique, this chapter will provide a general overview on the historical development of that measure of crime.

                                                                                                                          In the criminological field, the first use of the self-report technique to measure delinquency can be traced back to the works of Porterfield in the United States in the 1940s. The technique was improved during the 1950s and a major step was taken when Nye and Short introduced the first Guttman delinquency scale constructed on the basis of an SRD survey, in 1957. As we have explained elsewhere (Aebi, 2006), in the context of The American Soldier in World War II project,Guttman (1950) developed a scalogram that allowed combining in a table with N columns and N lines the answers of each person to selected items of a questionnaire. Simplifying things a little bit, these items constitute what is commonly called a Guttman scale. Indeed the selection of the items is the key element in the construction of the scale because they should measure a single dimension and must be classified in a hierarchical order according to the principle that a positive answer to an item implies agreement with items appearing lower in the scale. For example, from a logical point of view, the person who gives an affirmative answer to the question “do you smoke more than 15 cigarettes per day?” should also answer yes to the questions “do you smoke more than 10 cigarettes per day” and “do you smoke more than 5 cigarettes per day”. Thus, knowing the number of positive answers, the researcher can identify the questions that have been answered in an affirmative way.

                                                                                                                            In the field of criminology, a Guttman scale classifies offences from the less serious to the most serious one. However, as the samples are usually composed by juveniles attending high school, it is very difficult to establish such a scale without including minor deviant behaviours at the bottom of the scale. Accordingly, Nye and Short (1957) selected the following seven items - presented here from the less serious to the most serious - from the twenty-one behaviours included in their questionnaire: driving without a license, taking little things (worth less than $2), buying or drinking beer, wine, or liquor, skipping school without a legitimate excuse, having sex relations with a person of the opposite sex, purposely damaging or destroying public or private property, and defying parents’ authority to their face. The researchers were obliged to drop items such as running away from home and taking things of medium value (between $2 and $50) because they were committed by less than 10 percent of the sample. Their scalogram took also into account the frequency of the behaviour according to four categories (never, once or twice, several times, and very often).

                                                                                                                              Actually, one can see here a nice example of the interrelationship between science and technology. In the 1950s, the analysis of the data took a lot of time and it was almost impossible to conduct sophisticated statistical analysis. The Guttman scale was a technical innovation that allowed a relatively cheap and quick way of improving the quality of the analysis. Indeed, the article by Nye and Short (1957) can be seen as a statement in support of the use of such a scale in criminological research in order to break away from the simple dichotomies between delinquents and non-delinquents. Most researchers followed their advice and, in the same way that since the mid-1990s it is very difficult to find an empirical article that does not include a logistic regression - and in the 2000s it seems that a multilevel HLM analysis is a must for every research -, in the 1960s most articles included analysis based on Guttman scales. The problem comes from the fact that the scores in Guttman scales were used to classify respondents in different groups but, many times, one has the impression that the ones classified as heavily involved in delinquency are not really offenders but merely deviant youth.

                                                                                                                                In any case, only a few years later, SRD surveys were introduced in Europe. In Scandinavian countries, the first SRD surveys were part of the Nordic Draftee Research (NDR) Program and the fieldwork took place in Sweden already in 1959 and in Finland in 1962. In the United Kingdom, the technique was adopted in 1961 by the Cambridge Study in Delinquent Development (1961-2004), which was the first European longitudinal study of this kind, and it was also used in cross-sectional studies since 1963. In Germany and in the Netherlands, the first SRD surveys were conducted in the late 1960s, while in Belgium the first survey took place in 1976. In Italy, a small survey was conducted in 1980 but it was only with the participation of the country to the first ISRD in 1992 that the technique was used with larger samples. Finally, in France, the first ISRD was conducted in the late 1990s20.

                                                                                                                                • 20  It is difficult to write the history of SRD surv(...)

                                                                                                                                It is possible to analyze the evolution of SRD surveys in Europe using the categories proposed by Kivivuori (2008) for Scandinavian countries. One can thus distinguish three periods: the first one runs from the 1960s to the mid-1970s and corresponds to the moment when the first surveys were conducted in many countries. A second period runs from the mid-1970s to the end of the 1980s and is characterized by the fact that the technique was seldom used. The third and current period corresponds to a kind of golden age of such surveys that includes the development of the ISRD project at the European level in 1992 - that allowed countries such as Italy, Portugal, Spain and Switzerland to conduct their first large scale or national surveys -, a clear increase in the financing of such projects in the United Kingdom, and their institutionalization in countries such as Finland, Sweden, the Netherlands, England and Wales and Northern Ireland.

                                                                                                                                  In order to understand this evolution, one should take into account once more the interrelationship between science and technology. This link is clearer in the natural sciences. For example, the Galilean revolution would have been impossible without the help of the first telescopes (Morin, 1990), and Pasteur would have been unable as Semmelweiss, a few years before him (Hempel, 1966) to develop his theories without the microscope. More recently, the development of very powerful telescopes allowed researchers to discover galaxies and planets that could not be seen before and led to major changes in astronomy theories. At the same time, in order to refine their theories, researchers ask for more powerful instruments and help developing them. This leads to a “loop phenomenon” between science and technology (Morin 1990: 60) that can be appreciated when one analyzes the evolution of crime measures and crime theories (Aebi, 1999, 2006; Aebi, Jaquier, 2008). For example, the development of victimization surveys in the 1970s allowed the development of life-style theory (Hindelang, Gottfredson, Garofalo, 1978); but, before that, the development of SRD surveys had also play a major role on the evolution of criminological theory.

                                                                                                                                    As we have mentioned before, the first SRD surveys included a lot of problematic behaviours and trivial offences. These items were admitted by the great majority of the respondents, and such a result challenged previous criminological theories. The typical criminal was no longer a young urban male belonging to an ethnic minority. Indeed criminality seemed normal21. Moreover, if everyone was engaged in delinquency, the overrepresentation of some categories of the population found in official statistics could only be due to a differential reaction of the criminal justice system. This interpretation was also supported by a very influent article by Kitsuse and Cicourel (1963) on the limits and biases of official statistics. Such findings played a major role on the development of labelling theory in the 1960s and Marxist criminology in the early 1970s.

                                                                                                                                    • 21  The hypothesis of the normality of delinquency h(...)

                                                                                                                                    However, the self-report technique was being constantly improved and very soon it became clear that serious delinquency was not normal. As Kivivuori (2008) states:

                                                                                                                                      However, the NDR data harboured findings that could potentially contradict the “crime as normality” interpretation. For example, it was found that police detection likelihood reflected offending intensity (Christie, Andenaes, Skirbekk, 1965). Nils Christie, one of the pioneers of Nordic self-report research, explicitly warned that the crime-as-normality rhetoric, appealing as it was, should not be pressed into absurdity (Christie 1966 [1964], 59). However, he himself kept on using the normality argument (Christie 1975, 73).

                                                                                                                                        The stubborn existence of the chronic offender thus haunted the early NDR researchers.

                                                                                                                                          At the same time:

                                                                                                                                            The spirit of the times took an anti-positivist turn. The early NDR design had reflected the influence of empirically oriented American social science. This emphasis lost much of its appeal towards the end of the 1960s as researchers wanted to engage politically and started to criticise quantitative methods (Kivivuori, 2008).

                                                                                                                                              The Scandinavian situation described by Kivivuori was probably not very different than the one that could be found in the rest of Western Europe. At the beginning of the 1970s, a vast majority of the new generation of European social researchers was fascinated by the emergency of the paradigm of the social reaction. Thus, they concentrated their efforts on the study of the social reaction to crime and the process of labelling. As a consequence, even if one can find some noteworthy exceptions, research based on SRD surveys was fairly uncommon. Curiously enough, in 1977, an analysis made by Farrington using the SRD data from the Cambridge study was giving some support to labelling theory, even if it was also showing that police activity was not as arbitrary as some researchers thought because juveniles contacted by the police where the ones that were more engaged in delinquency according to self-report measures (Farrington, 1977). The latter result had also been found by Christie, Andenaes, Skirbekk (1965) some years before with a Nordic sample, but it has seldom been quoted by Christie himself.

                                                                                                                                                The situation changed drastically at the beginning of the 1990s. Since then, the number of SRD surveys conducted in Europe has increased at an incredible pace. Currently, they are being conducted in all Western European countries and also, through the ISRD-2 project, in some Central and Eastern European ones. Several explanations can be given to this evolution, but we think that the following three are the main ones:

                                                                                                                                                  * Improvements in computer technology: Once more we insist here on the link between science and technology. Since the 1990s it has been possible to conduct all the statistical analysis of a big SRD database using a single personal computer. This implies a reduction of the costs of a research and also of the time needed to produce a final report on it.

                                                                                                                                                    * A change in attitudes towards crime and security in developed nations22: There is currently no doubt that most developed countries became more punitive since the 1990s. Moreover, in many of these countries there are unending debates - leading sometimes to modifications of the Criminal Code - on juvenile delinquency and the ways in which juveniles should be treated by the criminal justice system. The causes of this change are probably twofold: on the one hand, there has been a real increase in delinquency during the second half of the 20th Century in most developed nations (Aebi, 2004, with references); on the other hand, politicians reintroduced the debate on crime and security in their agendas, amplifying thus the phenomenon, and asking for (and funding research on) data on crime and delinquency. The idea was to introduce policy reforms based on an evidence-based approach, although it seems that in many cases it was guided by merely electoral goals.

                                                                                                                                                    • 22  An interesting analysis of some aspects of this (...)

                                                                                                                                                    * A change of the main paradigm in criminological research: Thomas Kuhn (1970, 15) had doubts about the existence of paradigms in the social sciences. As a consequence, his concept of scientific revolutions is probably impossible to apply to them. In the natural sciences, a new paradigm replaces the old one completely (e.g. no serious scientist would defend today that the conception of Newton is superior to the one of Einstein). In the social sciences, on the contrary, there is no replacement but addition of paradigms. By the end of the 1960s, the paradigm of social reaction was introduced in the criminological scene, but it never replaced the former paradigms and nowadays is only one paradigm among others. Nevertheless, during the 1970s and early 1980s most European criminologist adhered to the social reaction paradigm and where quite reluctant to the types of empirical research that could be labelled as positivist23. The situation changed in the 1990s when a new generation of criminologists, trained in the use of personal computers, entered the scene and started to put the accent on alternative explanations of delinquency developed by researches that did not belong to the mainstream in the 1980s and that could be tested through SRD surveys.

                                                                                                                                                    • 23  The depletion in SRD surveys during that period (...)

                                                                                                                                                    Thus, in the 1990s research was relatively cheaper than before, in some countries politicians were willing to finance it, and there was a new generation of criminologists ready to conduct it. As a consequence, SRD surveys became an established measure of juvenile delinquency in Europe and therefore it is necessary to discuss their methodology.

                                                                                                                                                      V - Methodological issues

                                                                                                                                                      Several authors have identified the main methodological problems related to the use of SRD surveys. At the same time, many of the national reports presented before include considerations on that topic. This chapter offers a short summary of such problems based on the national chapters as well as on the reviews of Hindelang, Hirschi, Weiss (1981); Aebi (1999, 2006); Junger-Tas, Haen Marshall (1999); Thornberry, Krohn (2000); Aebi, Jaquier (2008).

                                                                                                                                                        1 - The content of SRD surveys

                                                                                                                                                        A typical self-reported delinquency survey usually includes questions on problematic behaviours, property offences, violent offences, substance abuse, fraud and dishonesty.

                                                                                                                                                          A - Problematic behaviours

                                                                                                                                                          These are behaviours that are wrongful or antisocial but that are usually not included in criminal law. Typical examples are truancy, running away from home, tobacco smoking, underage consumption of alcohol, and defying parental authority. The first American surveys included also having sex with a person of the opposite sex as an item that was included in self-reported delinquency scales (see Nye, Short, 1957).

                                                                                                                                                            As mentioned before, when self-reported delinquency surveys were introduced in the United States of America, most States considered these behaviours as status offences. As a consequence, such behaviours were included in self-reported delinquency surveys and, later, in the delinquency scales constructed on the basis of such surveys. We have already seen that the delinquency scale proposed by Nye and Short (1957) included mainly problematic behaviour and trivial offences. Later, the delinquency scale developed by Erickson (1972) encompassed eight behaviours including the following four: Smoking, buying tobacco, drinking beer, wine or liquors, and buying beer, wine or liquors. Of course, these behaviours are bad for the health of an adolescent, but from a European point of view, they cannot be considered as offences. However, as European self-reported delinquency surveys were inspired by American research, some of these behaviours were - and are still being included - in many European questionnaires. As a consequence, researchers should pay special attention when analyzing the results in order to avoid mixing problematic behaviours with serious delinquent act.

                                                                                                                                                              B - Property offences

                                                                                                                                                              SRD surveys regularly include questions on theft and many times they introduce a distinction between minor theft (less than a certain amount of money) and serious theft. Burglary is also generally included and there are questions about theft of a car - as well as about theft of motorcycles or bicycles - and theft from a car too. Other property offences frequently included in the survey are graffiti, vandalism and arson.

                                                                                                                                                                Curiously enough, questions on intellectual property theft are rare. Only in recent years, there has been an interest in this behaviour but it is usually measured in an indirect way through the use of computers to download illegal files. It is a pity that such questions were not included in the first SRD surveys, because one can imagine that the rates of intellectual property theft in the 1970s and 1980s (through the recording of cassettes) and in the 1990s (through the burning of compact discs) were not very different to the ones found in the 2000s (through the downloading of MP3 files). Also, theft of intellectual property through Xerox copies or - more recently - scanning of books and other printed material are regularly excluded from the survey.

                                                                                                                                                                  C - Violent offences

                                                                                                                                                                  Violent offences typically include robbery - which in many European continental countries is defined as theft with violence and therefore considered a property offence -, assault, bullying and use of weapons. The questions on assault have been improved with time in order to avoid including minor incidents such as slapping. Bullying has been included in recent years and therefore it is impossible to compare the current situation with the one that existed in the 1960s and 1970s.

                                                                                                                                                                    D - Substance abuse

                                                                                                                                                                    Questions on tobacco, alcohol and drug use were already included in the first SRD surveys. Currently, there is an international specialized survey - the ESPAD project presented previously - on that topic. In European countries that are wine producers, consumption of wine by adolescents during the meals was fairly common until some years ago and it was not seen as problematic. The same was true for beer. Our personal impression, based on research conducted in Southern Spain and informal interviews in Switzerland, is that there has been a displacement from consumption at home - where supervision by adults was assured - to consumption while going out with peers. It also seems that there has been a displacement from consumption of wine to consumption of liquors such as whisky, gin or vodka. Unfortunately, as many European SRD questionnaires were inspired by the American questionnaires that - according to the American laws - considered the use of alcohol by adolescents as an offence, SRD surveys have seldom measured the context in which consumption was taking place and the nuances between the different types of alcohol being used.

                                                                                                                                                                      E - Fraud and dishonesty

                                                                                                                                                                      SRD surveys frequently include questions on fare-dodging as well as on receiving and handling stolen goods. In countries where checks are a typical way of paying, there are usually questions on the misuse of them. Also, in some cases, one can find questions related to the abuse of credit cards and, more recently, computer related offences has been added in most questionnaires although sometimes their wording is doubtful. For example, in the ISRD-2 questionnaire, it is not clear whether the download that is being measured is legal or illegal.

                                                                                                                                                                        F - The issue of trivial offences

                                                                                                                                                                        As a rule, some of the behaviours included in SRD surveys are criminal or administrative offences but, as they are not serious, they would probably never lead to a criminal procedure. A typical example of such offence is fare-dodging.

                                                                                                                                                                          Both trivial offences and problematic behaviours - discussed above - can be seen, under certain circumstances, as good predictors of future delinquency. Therefore we are not advocating for the removal of all of them from self-reported delinquency scales. However, as suggested by Junger-Tas (1989), it is very important to avoid mixing them with serious offences when constructing delinquency scales. The inclusion of such items in general delinquency scales - especially when lifetime prevalence is studied - usually leads to results that show little differences in the distribution of delinquency by sex, socioeconomic status and other socio-demographic factors. As mentioned before, such indexes led in the 1950s and 1960s to the erroneous idea that delinquency was distributed homogeneously across the population.

                                                                                                                                                                            2 - Delinquency scales

                                                                                                                                                                            The first SRD surveys were mainly interested in the lifetime prevalence of different behaviours (“Have you ever…?”). For example, in the first Scandinavian SRD surveys, measures of incidence/frequency of delinquency were non-existent or rudimentary (Kivivuori, 2008). However, as indicated by McVie (2008), nowadays most SRD surveys include a general prevalence questions (i.e. have you done…?), for the period ‘ever’ and ‘last year’, followed by an incidence/frequency question (i.e. how many times have you done it?). Both ISRD surveys and most current SRD survey include also follow-up questions about the characteristics of the latest delinquent behaviour. However, in many cases these questions - and many others included in the questionnaire - have not been fully exploited in the available publications on SRD surveys.

                                                                                                                                                                              Prevalence and incidence questions are used to construct delinquency scales that include different behaviours. Variety (i.e. number of different offences committed) is also an interesting index that can easily be calculated on the basis of the questions on lifetime and last year prevalence. It is particularly interesting for testing versatility and specialization in delinquency. For example, Aebi (1999, 2006) found that most heavy heroin-addicted adult offenders had committed a large range of offences during their lifespan, but that they tend to specialize themselves in one or two types of offences - small scale drug-trafficking and shoplifting - when shorter periods of time were studied. Their situation can be compared to the one of the smoker that prefers a certain brand but has tasted most of the available ones. Moreover, variety of offending was the best predictor of the risk of being arrested by the police - the greater the range of offence, the greater the risk of being arrested -, while incidence was not linearly correlated to that risk. However, frequency is seldom used in publications on SRD surveys. An exception can be found in Junger-Tas, Marshall, Ribaux (2003) that used it as a good index for international comparisons.

                                                                                                                                                                                3 - Sampling

                                                                                                                                                                                In recent years, many cross-sectional SRD surveys are being conducted at school with students attending the ninth grade. Sometimes for example in the ISRD-2 the sample is enlarged in order to cover 7th, 8th and 9th grade students. Only exceptionally are post-adolescents included in the sample; for example, the ISRD-1 included juveniles 14 to 21 years old. Finally, one can find a few exceptions of surveys conducted with adults using samples of drug users or prisoners or in the context of longitudinal studies.

                                                                                                                                                                                  One of the typical critics to school samples is that juveniles that most engaged in deviant behaviours are probably not at school. Also, sometimes special education classes are not included in the sample. As a consequence, SRD surveys tend to underestimate the most serious types of delinquency.

                                                                                                                                                                                    Moreover, taking into account that the influence of some factors - such as the family of the school - is likely to diminish with time, limiting samples to the 9th grade may lead to an overestimation of the influence of such factors in future delinquency. For example, a study based on the Swiss ISRD-1 sample showed that family structure had an influence on drug consumption for adolescents aged 14-17 - males from broken homes were more involved in soft drugs use than the ones coming from traditional families -, but this influence disappeared for the ones aged 18-21 (Aebi, 1997).

                                                                                                                                                                                      4 - Response rate

                                                                                                                                                                                      Response rates vary widely across time and space, ranging typically from 70 to 90%. In Sweden, a decreased trend was observed in recent years (Andersson, 2008), but in other countries the evolution is not linear. One advantage in some European countries is that parents do not have to authorize explicitly their children to participate in the survey. Typically, parents are informed that, without opposition from their part, their children are going to participate in a survey. The experience of the ISRD shows that, under such circumstances, few parents refuse the participation of their children in the survey.

                                                                                                                                                                                        5 - Method of administration

                                                                                                                                                                                        The first SRD surveys were based on personal interviews. Paper-pencil questionnaires were introduced relatively soon. Sometimes –for example in the ISRD-1 both techniques are combined, with juveniles answering in a written form to the screen questions on delinquency (using the sealed-envelope technique). Since the 1990s, CAPI (computer assisted personal interview) and CASI (computer assisted self-administered interview) techniques are being used. In the first case, the interviewer administers the questionnaire, while in the second one the respondent reads the questions on the screen and enters directly the answers on the computer. Switzerland used CAWI (computer assisted web interview) for the ISRD-2. The only difference between CASI and CAWI is that, in the first one, the answers are stocked on the computer used for the interview, while on the second one they are stocked directly on the main database which is accessible through the Web (in a secured webpage only accessible through a password given to the interviewers). Finally, it is also possible to use ACASI (audio computer assisted audio self-interviewing), a technique in which the answers are recorded and the respondent can listen to them through the computer.

                                                                                                                                                                                          As far as the validity of the answers is concerned, comparisons of these techniques show only small differences from one to the other. However, the use of ACASI may solve at least partially the problems of illiteracy of some of the respondents even if, when the research is conducted at schools, this issue does not seem a serious threat to validity. At the same time, the use of CAPI, CASI and CAWI improves reliability by reducing the risk of introducing wrongful information in the database while coding the paper-pencil questionnaires. It also allows the use of more complex questionnaires - especially as follow-up questions are concerned - because the computer will lead the respondent directly to the next relevant question. This is a major problem in paper-pencil questionnaires, where the respondent has to look by himself/herself for the next question according to his/her answers (e.g. “if yes, please go to question No 55”).

                                                                                                                                                                                            Finally, comparisons of the presence of teachers or researchers during the administration of the questionnaire as well as anonymous versus non-anonymous versions of a questionnaire show that lower prevalence delinquency rates are found when the teacher is present and when anonymity is not assured.

                                                                                                                                                                                              6 - Validity of SRD surveys

                                                                                                                                                                                              Every instrument designed to measure a phenomenon can be considered as a measure of that phenomenon. The validity of a measure can be defined as its capacity to measure efficiently the phenomenon under investigation. In particular, the validity of a crime measure can be defined as its capacity to measure efficiently the crime phenomenon.

                                                                                                                                                                                                There are several classifications of the different types of validity. In the following chapters we will present these types according to the classification that we proposed a few years ago (Aebi, 1999, 2006).

                                                                                                                                                                                                  * Content validity

                                                                                                                                                                                                    - Face validity (apparent validity)

                                                                                                                                                                                                      - Logical validity

                                                                                                                                                                                                        * Pragmatic validity

                                                                                                                                                                                                          - Concurrent validity

                                                                                                                                                                                                            Known group validity

                                                                                                                                                                                                              Correlational validity

                                                                                                                                                                                                                - Predictive validity

                                                                                                                                                                                                                  * Construct validity

                                                                                                                                                                                                                    Before starting the presentation of these types of validity, it is necessary to point out that it is possible to test the validity of the whole survey or the validity of the delinquency scales derived from it.

                                                                                                                                                                                                                      A - Content validity

                                                                                                                                                                                                                      Content validity refers to question of whether the content of the survey - i.e. the offences included in it - corresponds to its label (is this survey really a self-reported delinquency survey?). As such, it entails examining if the behaviours included in the SRD survey are actually delinquent behaviours (face validity) and if they constitute a representative sample of the universe of delinquent behaviours (logical validity). From that definition it can be seen that there are two subtypes of content validity that we will discuss in the following chapters.

                                                                                                                                                                                                                        a - Face validity

                                                                                                                                                                                                                        By definition, a self-reported delinquency survey is supposed to measure delinquency. However, as we have explained at the beginning of this paper, some of the behaviours included in SRD surveys are not considered as criminal offences in Continental European countries. As mentioned before, the reason of this apparent paradox is that the concept of delinquency is broader in English speaking countries - i.e. it includes wrongful and antisocial behaviour - than in Continental European countries.

                                                                                                                                                                                                                          The only empirical way of testing face validity is to compare the items included in the survey with the offences included in the criminal laws of the country. American researchers include under such heading the statutory laws that foreseen the status offences mentioned above and usually conclude that the survey meets the requirements of face validity. European researchers usually include such behaviours under the heading of problematic behaviour and conclude that the survey measures delinquency and other forms of antisocial behaviour. Once more it must be mentioned that, in order to overcome this threat to face validity, the researcher must avoid mixing serious offences with problematic behaviours or trivial offences in a general delinquency scale.

                                                                                                                                                                                                                            b - Logical validity

                                                                                                                                                                                                                            We have already discussed the content of SRD surveys indicating that typically they include problematic behaviours, property offences, violent offences, substance abuse, fraud and dishonesty. From that list, it is obvious that many criminal offences are not included in the surveys, and therefore one can conclude that the behaviours included in it do not constitute a representative sample of all criminal offences. This implies a challenge to the logical validity of the survey. However, it has generally been considered that most SRD surveys include the most typical offences committed by juveniles and that therefore their logical validity is acceptable.

                                                                                                                                                                                                                              In that context, we have already regretted the fact that intellectual property offences were not included until recent years, and the same can be said about bullying and the context in which alcohol is being consumed. From that point of view, it is possible to affirm that the logical validity of most SRD surveys has been improved in recent years.

                                                                                                                                                                                                                                In SRD surveys conducted with adults, the behaviours included in the survey must be modified. For example, in Switzerland and Germany, SRD surveys were used for measuring delinquency of drug-addicts following heroin prescription programs and the offences included in such surveys were adapted to those related to hard-drug addict life-style.

                                                                                                                                                                                                                                  B - Pragmatic validity

                                                                                                                                                                                                                                  Pragmatic validity refers to question of whether an instrument allows establishing the current state (concurrent validity) or the future state (predictive validity) of the concept that is being measured. In the case of SRD surveys, concurrent validity entails examining if the survey measures the involvement in delinquency of the respondents; while predictive validity requires examining if the survey predicts correctly the future involvement in delinquency of them.

                                                                                                                                                                                                                                    Indeed, the only way to establish if an instrument measures correctly a phenomenon is to compare its results with the results provided by an instrument whose validity has already been proven. For example, a researcher that develops a new balance can test its measures with the one provided by the balances that are already available. The problem comes from the fact that there is no universally accepted measure of delinquency. Furthermore, all existing measures of delinquency are considered imperfect. As a consequence, it is possible to compare different measures of delinquency, but it is not clear how to interpret the differences found between these different measures. In that context, the results of SRD surveys have usually been compared with the results of official measures of crime, such as police or court files.

                                                                                                                                                                                                                                      a - Predictive validity

                                                                                                                                                                                                                                      The correct way of testing predictive validity implies using an SRD survey at time 1 (t1) and, on the basis of the analysis of the survey, making a prediction about the future implication in delinquency of the respondents to the survey. After some time (t2), a second SRD survey must be conducted with the same sample and a comparison of the predictions with the observations should be performed.

                                                                                                                                                                                                                                        In spite of the existence of quite a few SRD longitudinal surveys, predictive validity is never tested that way. Usually, the researchers do not make a prediction. Instead of that, they compare the results at t1 and at t2 in order to establish retrospectively which would have been the best predictive factors at t1.

                                                                                                                                                                                                                                          b - Concurrent validity

                                                                                                                                                                                                                                          In order to estimate concurrent validity, one needs to compare the results obtained with two instruments. In that perspective, researchers distinguish two subtypes of concurrent validity: Known group validity and correlational validity.

                                                                                                                                                                                                                                            c - Known group validity

                                                                                                                                                                                                                                            In order to test known group validity, one compares the scores in self-reported delinquency of two or more groups whose implication in delinquency is different according to another crime measure. For example, using official records of crime - such as police or court records -, one can compare the SRD scores of the group of persons officially delinquent (known by the police or convicted by a court) with the ones that are officially non-delinquent. The logic of the comparison implies that the groups that are different according to official measures of crime should also be different according to SRD measures. Thus, officially delinquent youth should disclose more offences than officially non-delinquent youth. If that is not the case, there is a problem of validity with one of the measures.

                                                                                                                                                                                                                                              The alternative measures used to test the validity of SRD surveys vary widely. Some are close to official measures of crime (contacts with the police, contact with a court, institutionalisation), some use non-official measures (reports from the teachers, personality tests, class play role assignment, etc.) and some combine both types of measures24. In that context, the typical way of conducting known group validity tests is to compare self-reported police contacts with the police with contacts registered by the police. If respondents do not admit contacts registered by the police, it is clear that they are lying.

                                                                                                                                                                                                                                              • 24  For a detailed review of these studies, see Aebi(...)

                                                                                                                                                                                                                                              The results of the studies that tested known group validity show that the group that was supposed to be more involved in delinquency did present higher SRD scores - in incidence and variety scales - than the group that was supposed to be less involved in it.

                                                                                                                                                                                                                                                d - Correlational validity

                                                                                                                                                                                                                                                According to Hindeland, Hirschi and Weiss (1981, ch. 5) correlational validity can be seen as a refined form of known group validity. Both of them compare two measures of crime, but the difference comes from the fact that in known group validity one uses the crime measures as a criterion to assign individuals to different groups, while in correlational validity the goal is to calculate the correlation between both measures. Thus, studies that tested correlational validity have also compared SRD surveys to official and non-official measures of crime.

                                                                                                                                                                                                                                                  A review of the studies that conducted comparisons of SRD surveys and official measures of crime (Aebi, 2006)25 shows that they have usually found weak positive correlations. In particular, the size of the correlation is heavily influenced by the scale used to measure delinquency. The strongest correlations are found with the widest scales of delinquency in terms of the number of offences included and the period studied. In addition, the correlations are sensitive to the sex, age and ethnic background of the sample, suggesting a differential validity of SRD surveys. In that context the validity of SRD surveys seems stronger for juveniles and for autochtones.

                                                                                                                                                                                                                                                  • 25  For a detailed review of these studies, see Aebi(...)

                                                                                                                                                                                                                                                  The fact that the correlations are positive shows that both measures are measuring the same concept. As this concept was defined as delinquency, it is possible to conclude that both measures are relatively valid as measures of crime. However, as we will see in the following chapter, it is not easy to interpret the size of the correlations found.

                                                                                                                                                                                                                                                    Some reflexions on the logical problems related to the comparisons of SRD surveys and official measures of crime

                                                                                                                                                                                                                                                      The history of criminology can be seen as the history of the search for reliable and valid measures of crime. In that context, Hindelang, Hirschi, Weis (1981, 97s.) pointed out that the fact that researchers would turn to official measures of crime as a criterion of validity of SRD surveys was seen as a paradox or at least as an admission of the superiority of such measures by some researchers. Indeed, if official measures of crime were considered as non-valid and SRD survey were developed as an alternative to them, it is not logic at all to use official measures of crime to test the new measure.

                                                                                                                                                                                                                                                        Trying to find a solution to that situation, Hindelang, Hirschi, Weis (1981, 97s.) pointed out that a good survey should at least identify all those who had been identified by the official measure, because the latter is only considered completely non-valid by those who are unconditional supporters of the SRD surveys. This is precisely the kind of study that is conducted when testing known group validity, and we have seen that SRD surveys usually meet the requirements of such test.

                                                                                                                                                                                                                                                          In that context, the fact that persons identified as delinquent by the police or the courts show higher scores in SRD than those not identified as such has been interpreted as a corroboration of the validity of SRD surveys. However, it can also be interpreted as a corroboration of the validity of official measures of crime. Indeed, if a person that admits having committed a lot of offences has already been contacted by the police while someone who has not admitted offences has never been contacted by them, one could conclude that such a result shows that the police are doing a relatively good work.

                                                                                                                                                                                                                                                            The problem with this kind of reasoning is that it is deductive in nature and it starts with an axiom. Indeed, researchers have adapted the classic syllogism Man is mortal, Socratesis aman, therefore, Socratesismortal in this way: “Official measures of crime are relatively valid; SRD surveys match official measures; therefore, SRD surveys are relatively valid”. However, this method - severely criticized by Russell (1961 [1912]) - can only lead to a valid conclusion when the major premise can be accepted without any restriction (i.e. when it constitutes an axiom), and that is not the case as far as the validity of official measures is concerned.

                                                                                                                                                                                                                                                              From a logical point of view, the problem cannot be solved because no measure of delinquency is considered as completely valid. From an empirical point of view, the only possibility is to test both measures in both ways. This implies testing whether SRD surveys can differentiate official delinquents, and also whether official measures can differentiate the respondents that defined themselves in the SRD survey as delinquents.

                                                                                                                                                                                                                                                                The combination of both types of tests is interesting because, on the one hand, if persons identified by the police do not admit offences, one can conclude that SRD surveys are not valid; while, on the other hand, if the police (courts) do not identify the persons that admit offences in the SRD survey, one can conclude that official records are not valid. Finally, if the two measures coincide, the validity of both of them is corroborated.

                                                                                                                                                                                                                                                                  In practice, the second type of test - creating groups of delinquents according to the SRD survey and comparing their scores in official records - has seldom been used because usually SRD studies have been conducted mainly with adolescents that are not heavily involved in delinquency and therefore it is difficult to establish a group of delinquents. One European exception is our research with adult hard-drug addicts following a heroin prescription treatment in Switzerland (Aebi, 2006). In that case, both measures were associated even if SRD surveys allowed identifying more offenders and many more offences than police and court records and therefore they were considered as relatively valid.

                                                                                                                                                                                                                                                                    C - Construct validity

                                                                                                                                                                                                                                                                    The degree of correspondence between the empirical results of the research and the theoretical previsions is called construct validity. However, when the results do not follow the theoretical previsions, sometimes it is difficult to know whether the theory is not correct or whether it is correct but the instruments used to test it are not valid (Selltiz, Whrigtsman, Cook, 1977). Of course, the same problem also exists when the theoretical previsions are corroborated by the results. In that case, one can wonder whether the results are not due to the research design or the instruments used.

                                                                                                                                                                                                                                                                      Construct validity has been studied mainly in the field of psychology, in order to test the efficiency of some tests. In the mid-1980s, Huizinga and Elliott (1986) considered that it had seldom been used in delinquency research because it was too complicate to examine within the same study both tests of theory and validity issues. However they concluded that there was some indication of the construct validity of SRD measures because, when correlations were not the ones specified by a given theory, researchers have concluded that the theory was wrong and not that the SRD measure was invalid (Huizinga, Elliott, 1986).

                                                                                                                                                                                                                                                                        Even if in recent years many theories have been tested using SRD studies, it is difficult to establish clearly the construct validity of the surveys because in many cases they do not include questions that would allow testing alternative theories. However, the importance of construct validity should not be underestimated because, as it has been shown in the field of domestic violence, research based on the self-report technique (Straus, Gelles, 1990) leads to completely different conclusions to the one based on victimization surveys.

                                                                                                                                                                                                                                                                          7 - Reliability of SRD surveys

                                                                                                                                                                                                                                                                          Reliability refers to the extent to which a measuring instrument - in our case, the SRD survey - would produce identical scores if it were used to make multiple measures of the same object (Huizinga, Elliott, 1986). According to Hagan (2005, 298) there are two types of reliability: stability and consistency. Stability implies testing whether, assuming that conditions have not changed, a respondent gives the same answer to the same question on second testing; while consistency implies testing whether the set of item used to measure a phenomenon are highly related and measure the same concept (Hagan, 2005, 298).

                                                                                                                                                                                                                                                                            As with validity, it is possible to test the reliability of the whole survey or the reliability of the delinquency scales constructed on the basis of the survey. Reliability of SRD surveys has been tested using the test-retest method (i.e. administering the survey twice to the same population26), the split-half technique (in which the instrument is divided in two halves and each half is analyzed separately and then compared to the other in order to see whether the scores are similar), and the split-ballot technique (in which the sample is divided randomly in two halves an alternate forms of the instrument are administered to each half before testing whether the scores are comparable).

                                                                                                                                                                                                                                                                            • 26  It is also possible to test alternate forms of t(...)

                                                                                                                                                                                                                                                                            VI - Self-reported delinquency surveys and criminological theory

                                                                                                                                                                                                                                                                            SRD surveys are a major tool for the development and testing of criminological theories, in particular of those that try to explain juvenile delinquency. In fact, since the 1960s, most theories in criminology have been developed on the basis of the results of SRD surveys. We have already mentioned the influence that the first SRD studies had on development of the labelling approach. Other paradigmatic examples in the United States are social bonding theory (Hirschi, 1969) and self-control theory (Gottfredson, Hirschi, 1990). In Europe, these surveys played a major role in the development of the integrated cognitive antisocial potential theory (Farrington, 2005), and Situational Action Theory (Wikström, 2005). Apart from that, SRD surveys are regularly used for testing existing theories. For example, as we have already mentioned, the ISRD-1 questionnaire was inspired mainly by control theory, and the second one includes different theoretical approaches. The later approach - a combination of different theories - is used by the majority of the recent studies, including the Edinburg Study of Youth Transitions and Crime, the Belfast Youth Development Study and the KFN surveys in Germany.

                                                                                                                                                                                                                                                                              From a theoretical point of view, the ideal SRD survey should include questions referring to different theories in order to allow comparisons of the explanatory power of each of these theories. However, leaving aside the classical problems related to the operationalization of the concepts of each theory - that are at the origins of many debates in criminology -, it must be mentioned that some theories are harder to test than others. This is typically the case for labelling theory that requires a longitudinal design in order to compare at time t2 the implication in delinquency of persons who had contacts with the criminal justice system and persons who did not have such contact, but controlling that both groups had with a similar implication in delinquency at time t1. Situational-based theories create also problems when it comes to measures the opportunities to commit offences. Usually, the questionnaires include questions on the lifestyle of the person that are then used to infer his/her exposition to the risk of committing an offence but, with only a few exceptions, the concrete occasions of committing offences are seldom taken into account.

                                                                                                                                                                                                                                                                                VII - Self-reported delinquency surveys and criminal policies

                                                                                                                                                                                                                                                                                At a national level, the influence of SRD surveys on criminal policies is clearly related to the influence of this indicator in each country. In countries with a weak tradition of SRD studies, it seems that these studies are not playing a major role on the development of criminal policies or are having only some influence at the local level (e.g. Italy). On the other hand, when this crime measure becomes part of the criminological scene, it is often taken into account for such policies. This is the case mainly in the United Kingdom, where the Cambridge Study in Delinquent Development has inspired some legal reforms. In countries such as Finland, Germany and the Netherlands, SRD surveys are also playing a role in the political debate on crime and crime prevention.

                                                                                                                                                                                                                                                                                  VIII - Conclusion

                                                                                                                                                                                                                                                                                  SRD surveys have become a standard measure of delinquency in Europe. However, their validity cannot be established on an abstract basis. On the contrary, it is necessary to analyze each survey - paying particular attention to the sampling, the conditions of administration of the survey and the construction of the questionnaire - in order to establish its degree of validity.

                                                                                                                                                                                                                                                                                    SRD surveys do not measure the most serious types of offences. However, they provide extremely useful information for minor and less-serious types of offences.

                                                                                                                                                                                                                                                                                      Recommended good practices include:

                                                                                                                                                                                                                                                                                        - When using school samples, include special education classes and find a way of adding youth non attending school to the sample.

                                                                                                                                                                                                                                                                                          - Improving a common questionnaire, such as the ISRD-2, in order to allow comparisons across time and space.

                                                                                                                                                                                                                                                                                            - Clearly separate problematic behaviours and trivial offences from the rest of the offences.

                                                                                                                                                                                                                                                                                              - Include more serious offences in the questionnaire (e.g. sexual abuse).

                                                                                                                                                                                                                                                                                                - Improve measures of socio-demographic variables.

                                                                                                                                                                                                                                                                                                  - Use CAPI, CASI and CAWI in order to reduce the costs of the survey and the risks of introducing mistakes while entering the information in the database.

                                                                                                                                                                                                                                                                                                    - Include questions on victimization in order to have a more complete picture of the sample.

                                                                                                                                                                                                                                                                                                      The national reports summarized in this article show that in countries such as England and Wales, Finland, the Netherlands, Northern Ireland and Sweden, SRD surveys have been institutionalized and national surveys are run on a regular basis that allows the development of time series. In particular, the United Kingdom has a strong tradition of SRD studies and currently an impressive number of studies - including longitudinal ones - are taking place there. Germany is also running longitudinal as well as cross-sectional regional and local surveys on a more or less regular basis. Finally, in Belgium, France, Italy and Ireland, SRD surveys are not institutionalized but surveys are taking place punctually and the four countries have taken part in the second ISRD.

                                                                                                                                                                                                                                                                                                        The reports also show that the use of SRD studies decreased since the mid-1970s to the end of the 1980s and clearly increased since the beginning of the 1990s.

                                                                                                                                                                                                                                                                                                          At the general European level, the positive experience of the ISRD-1 in 1992 - in which eleven of the twelve participant countries were European - showed that it was possible to develop joint and comparable research and provided a common questionnaire that was later used in many studies. Finally, the creation of the European Society of Criminology (ESC) in 2000 provided a forum where European criminologists could meet and build up joint projects and clearly helped the development of the ISRD-2, in which 24 of the 30 participating countries are European, and their national correspondents meet regularly at the ESC Annual Conferences.

                                                                                                                                                                                                                                                                                                            In brief, not only SRD surveys seem to have found their place in European criminology as a major measure of juvenile delinquency, but it seems also clear that their use will probably increase in the future.

                                                                                                                                                                                                                                                                                                              Bibliography

                                                                                                                                                                                                                                                                                                              Aebi M. F., Famille dissociée et criminalité : le cas suisse, Kriminologisches Bulletin de Criminologie, 1997, 23, 1, 53–80.

                                                                                                                                                                                                                                                                                                              Aebi M. F., La validité des indicateurs de la criminalité : les sondages de délinquance autoreportée face aux données de police et du casier judiciaire, thèse de doctorat en criminologie, Lausanne, Université de Lausanne, Faculté de Droit, Institut de police scientifique et de criminologie, 1999.

                                                                                                                                                                                                                                                                                                              Aebi M. F., Crime Trends in Western Europe from 1990 to 2000, European Journal on Criminal Policy and Research, 2004, 10, 2-3, 163-186.

                                                                                                                                                                                                                                                                                                              Aebi M. F., Comment mesurer la délinquance ?, Paris, Armand Colin, 2006.

                                                                                                                                                                                                                                                                                                              Aebi M. F., Enquêtes de délinquance autoreportée en Europe, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Aebi M. F., Jaquier V., Les sondages de délinquance autoreportée : origines, fiabilité et validité, Déviance et société, 2008, 32, 2, 205-227.

                                                                                                                                                                                                                                                                                                              Andersson L., Self-Reported Delinquency Surveys in Sweden, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              Andersonn L., Délinquance autoreportée en Suède, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Beccaria C. (1991). Des délits et des peines. Paris: Garnier-Flammarion (original ed.: Dei delitti e delle pene, 1764).

                                                                                                                                                                                                                                                                                                              Carra C.,Self-Reported Delinquency Surveys in France, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              Carra C., Les enquêtes de délinquance autoreportée en France, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Christie  N., Hvorfor blif unge mennesker forbrytere?, in Andenaes J., Bratholm A., Christie N., Kriminalitet og samfunn, Bergen, J. W. Eides Boktrykkeri, 1966 [1964.

                                                                                                                                                                                                                                                                                                              Christie N., Hvor tett ett samfunn?,Oslo, Christian Ejler’s forlag, 1975.

                                                                                                                                                                                                                                                                                                              Christie N., Andenaes J., Skirbekk S., A study of self reported crime, Scandinavian Studies in Criminology, 1965, 1, 86-116.

                                                                                                                                                                                                                                                                                                              Durkheim É., Les règles de la méthode sociologique, Paris, Flammarion, 1988 (Original ed. : 1895).

                                                                                                                                                                                                                                                                                                              Elliott D. S., Huizinga D., Improving self-reported measures of delinquency, in Klein M. W. (Ed.), Cross-national research in self-reported crime and delinquency, Dordrecht [etc.], Kluwer Academic Publishers, 1989, 155-186.

                                                                                                                                                                                                                                                                                                              Erickson M., The changing relationship between official and self-reported measures of delinquency: An exploratory-predictive study, The journal of Criminal Law, Criminology and Police Science, 1972, 63, 3), 388-395.

                                                                                                                                                                                                                                                                                                              Farrington D. P., The effects of public labeling, The British Journal of Criminology, 1977, 17, 2, 112-125.

                                                                                                                                                                                                                                                                                                              Farrington D.P., The integrated Cognitive Antisocial Potential (ICAP) theory, in Farrington D.P. (Ed.), Integrated Developmental and Life-course theories of offending,  New Brunswick, NJ, Transaction, 2005.

                                                                                                                                                                                                                                                                                                              Garland D., The culture of control: Crime and social order in contemporary society, Oxford, Oxford University Press, 2001.

                                                                                                                                                                                                                                                                                                              Görgen T., Rabold S.,Self-Reported Delinquency Surveys in Germany, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              Görgen T., Rabold S., Enquêtes de délinquance autoreportée en Allemagne, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Gottfredson M. R., Hirschi T., A General Theory of Crime, Stanford, CA, Stanford University Press, 1990.

                                                                                                                                                                                                                                                                                                              Grasmick H.G., Tittle C.R., Bursik R.J., Arneklev B. J., Testing the Core Empirical Implications of Gottfredson and Hirschi's General Theory of Crime, Journal of Research in Crime and Delinquency, 1993, 30, 1, 5-29.

                                                                                                                                                                                                                                                                                                              Guttman L., The basis for scalogram analysis, in Stouffer S. A., Guttman L., Suchman E. A., Lazarsfeld P. F., Star S. A., Clausen J. A.(Eds), The American Soldier: Studies in Social Psychology in World War II. Vol. 4:Measurement and Prediction, Princeton, NJ, Princeton University Press, 1950, 60-90.

                                                                                                                                                                                                                                                                                                              Hagan F. E.,Research methods in criminal justice and criminology (7th ed.), Boston [etc.], Allyn and Bacon, 2005.

                                                                                                                                                                                                                                                                                                              HempelC., Philosophy of Natural Science, Englewood Cliffs, NJ, Prentice Hall, 1966.

                                                                                                                                                                                                                                                                                                              Hibell B., Anderson B., Bjarnason T., Ahlstrom S., Bakalireva O, Kokkevi A., Morgan M., The 2003 ESPAD report: Alcohol and other drug use among students in 35 European countries,Stockholm, Swedish Council for Information on Alcohol and Other Drugs, Strasbourg, Council of Europe, 2004.

                                                                                                                                                                                                                                                                                                              Hindelang M. J., Gottfredson M. R., Garofalo J.,Victims of personal crime: An empirical foundation for a theory of personal victimization, Cambridge, MA, Ballinger, 1978.

                                                                                                                                                                                                                                                                                                              Hindelang M. J., Hirschi T., Weiss J. G.,Correlates of delinquency: The illusion of discrepancy between self–report and official measures, American Sociological Review, 1979, 44, 995-1014.

                                                                                                                                                                                                                                                                                                              Hindelang M. J., Hirschi T., Weiss J. G., Measuring delinquency, Beverly Hills, CA, Sage, 1981.

                                                                                                                                                                                                                                                                                                              Hirschi T.,Causes of Delinquency, Berkeley, CA, University of California Press, 1969.

                                                                                                                                                                                                                                                                                                              Huizinga D., Elliott D. S., Reassessing the reliability and validity of self-report delinquency measures. Journal of Quantitative Criminology, 1986, 2, 4, 293-327.

                                                                                                                                                                                                                                                                                                              Junger-Tas J., Self-report delinquency research in Holland with a perspective on international comparison, in Klein M. W. (Ed.), Cross-national research in self-reported crime and delinquency, Dordrecht [etc.], Kluwer, 1989, 17-41.

                                                                                                                                                                                                                                                                                                              Junger-Tas J., The international self-report delinquency study: Some methodological and theoretical issues, in J. Junger-Tas, G.-J. Terlouw, Klein M. W. (Eds), Delinquent behavior among young people in the western world: First results of the international self-report delinquency study, Amsterdam [etc.], Kugler, 1994a, 1-13.

                                                                                                                                                                                                                                                                                                              Junger-Tas J., Delinquency in thirteen western countries: Some preliminary conclusions, in J. Junger-Tas, G.-J. Terlouw, Klein M. W. (Eds), Delinquent behavior among young people in the western world: First results of the international self-report delinquency study, Amsterdam [etc.], Kugler, 1994b, 370-380.

                                                                                                                                                                                                                                                                                                              Junger-Tas J., Haen Marshall I.,The self-report methodology in crime research, Crime and Justice: A Review of Research, 2003, 25, 291-367.

                                                                                                                                                                                                                                                                                                              Junger-Tas J., Haen Marshall I., Ribeaud D., Delinquency in an International Perspective, Monsey NY: Criminal Justice Press, The Hague, Kugler, 1999.

                                                                                                                                                                                                                                                                                                              Junger-Tas J., Terlouw G., Klein M.,. Delinquent Behavior Among Young People in the Western World: First Results of the International Self-Report Delinquency Study, Amsterdam [etc.], Kugler, 1994.

                                                                                                                                                                                                                                                                                                              Kivivuori J., Self-Reported Delinquency Surveys in Finland, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              Kivivuori J., Les enquêtes de délinquance autoreportée en Finlande, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Kitsuse J., Cicourel A., A note on the use of official statistics, Social Problems, 1963, 11, 2, 131-139.

                                                                                                                                                                                                                                                                                                              Klein M. W. (Ed.), Cross-national research in self-reported crime and delinquency, Dordrecht, Boston, London, Kluwer Academic, 1989.

                                                                                                                                                                                                                                                                                                              Kuhn T. S., The structure of scientific revolutions (2nd ed.), Chicago, University of Chicago Press, 1970.

                                                                                                                                                                                                                                                                                                              McVie S. Self-Reported Delinquency Surveys in the United Kingdom and Ireland, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              McVie S., Enquêtes de criminalité et de délinquance autoreportées en Grande-Bretagne et en Irlande, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Moffitt T.E., Silva P.A., Lynam D.R., Henry B., Self-reported delinquency at age 18: New Zealand’s Dunedin multidisciplinary health and development study, in Junger-Tas J., Terlouw G. J., Klein M. W. (Eds), Delinquent behavior among young people in the western world: First results of the international self-report delinquency study, Amsterdam [etc.], Kugler, 1994, 354-369.

                                                                                                                                                                                                                                                                                                              Morin E., Science avec conscience, Paris, Fayard, 1990.

                                                                                                                                                                                                                                                                                                              Nye F. I., Short J. F. Jr., Scaling delinquent behavior, American Sociological Review, 1957, 22, 326-331.

                                                                                                                                                                                                                                                                                                              Pauwels L., Pleysier S.,Self-Reported Delinquency Surveys in Belgium and the Netherlands, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              Pauwels L., Pleysier S., Enquêtes de délinquance autoreportée en Belgique et aux Pays-Bas, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Porterfield A. L., Youth in trouble, Fort Worth, Leo Potisham Foundation, 1946.

                                                                                                                                                                                                                                                                                                              Raudenbush, S. W.,Bryk, A. S., Hierarchical Linear Models: Applications and Data Analysis. Method (2nd. Edition), Newbury Park, CA, Sage, 2002.

                                                                                                                                                                                                                                                                                                              Russell B., The problems of philosophy, in Egner E., Denonn L. E. (Eds), The basic writings of Bertrand Russell, New York, Simon and Schuster, 1961 (article published originally in 1912).

                                                                                                                                                                                                                                                                                                              Selltiz C., Wrightsman L. S., Cook S. W., Les méthodes de recherche en sciences sociales, Montréal, HRW, 1977.

                                                                                                                                                                                                                                                                                                              Straus M. A., Gelles R. J. (Eds), Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8,145 Families, New Brunswick, London, Transaction, 1990.

                                                                                                                                                                                                                                                                                                              Thornberry T. P., Krohn M. D., The self-report method for measuring delinquency and crime, in Duffee D. (Ed.), Criminal Justice 2000 Volume 4: Measurement and analysis of crime and justice, Washington, DC, 2000, 33-83.

                                                                                                                                                                                                                                                                                                              Traverso S., Cartocci G., Traverso G. B., Self-Reported Delinquency Surveys in Italy, Paper presented at the Crimprev Seminar on Self-Reported Delinquency Surveys in Europe, Paris, 17-19/01/2008.

                                                                                                                                                                                                                                                                                                              Traverso S., Cartocci G., Traverso G.B., Enquêtes de délinquance autoreportée en Italie, in Zauberman R. (dir.), Les enquêtes de délinquance et de déviance autoreportées en Europe. État des savoirs et bilan des usages, Paris, l’Harmattan, forthcoming.

                                                                                                                                                                                                                                                                                                              Wikström P. O., The Social Origins of Pathways in Crime: Towards a Developmental Ecological Action Theory of Crime Involvement and Its Changes, in Farrington D.P. (Ed)., Integrated Developmental and Life-course theories of offending,  New Brunswick, NJ, Transaction, 2005.

                                                                                                                                                                                                                                                                                                              Footnotes

                                                                                                                                                                                                                                                                                                              1  According to the Webster’s New Universal Unabridged Dictionary, the Latin delinquere (to do wrong) led to the Late Latin delinquentia (fault, crime).

                                                                                                                                                                                                                                                                                                              2  For example, encuesta de delincuencia autorrevelada in Spanish, sondage de délinquance autoreportée in French, and indagine sulla delinquenza autorivelata in Italian.

                                                                                                                                                                                                                                                                                                              3  In this report we follow the conventional use of the term American to refer to the United States of America, even if we are fully aware that America is a continent and the United States are only one of the countries of that continent.

                                                                                                                                                                                                                                                                                                              4  This chapter is based on Kivivuori (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              5  This chapter is based on Andersson (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              6  This chapter is based on McVie (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              7  This chapter is based on Görgen and Rabold (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              8  This chapter is based on Pauwels, Pleysier (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              9  This chapter is based on Pauwels, Pleysier (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              10  This chapter is based on Traverso, Cartocci, Traverso (2008 and forthcoming).

                                                                                                                                                                                                                                                                                                              11  This chapter is based on Carra (2008, with references).

                                                                                                                                                                                                                                                                                                              12  The original English questionnaire can be found in Junger-Tas, Terlouw, Klein (1994, 387-441).

                                                                                                                                                                                                                                                                                                              13  According to Moffitt et al. (1994, 358), 80% of the delinquency items included in both instruments (ISRD-1 and National Youth Survey) overlapped.

                                                                                                                                                                                                                                                                                                              14  Our presentation of the ISRD-1 is based on the main publication about it (Junger-Tas, Terlouw, Klein, 1994). The authors of that publication mention thirteen countries because they include New Zealand. However, data from New Zealand refer to people aged 18 and comes from a longitudinal study that did not use the ISRD questionnaire, but the SRD instrument developed by Elliott and Huizinga (1989) for the National Youth Survey and the Denver Youth Study in the United States (Moffitt et al., 1994, 357), that we have mentioned in the previous footnote.

                                                                                                                                                                                                                                                                                                              15  Belgium applied a mixed approach combining a school sample with a random sample of persons aged 18 to 21.

                                                                                                                                                                                                                                                                                                              16  However, Italy used a school sample, but the questionnaire was administered through face-to-face interviews.

                                                                                                                                                                                                                                                                                                              17  Armenia, Aruba, Austria, Belgium, Bosnia-Herzegovina, Canada, Cyprus, Czech Republic, Denmark, Estonia, Germany, Hungary, Iceland, Ireland, Italy, Lithuania, the Netherlands, the Netherlands Antilles, Northern Ireland, Norway, Poland, Portugal, Russia, Scotland, Slovenia, Surinam, Sweden, Switzerland, the United States of America, and Venezuela.

                                                                                                                                                                                                                                                                                                              18  The different techniques for administering SRD surveys will be discussed later, in the chapter on Methods of administration.

                                                                                                                                                                                                                                                                                                              19  For example, Switzerland funded partially the participation in the survey of Armenia and Bosnia-Herzegovina.

                                                                                                                                                                                                                                                                                                              20  It is difficult to write the history of SRD surveys in Europe without mentioning the name of Josine Junger-Tas, who conducted the first survey in Belgium in 1976, promoted such surveys very actively in the Netherlands while she was Head of the WODC during the 1980s and was one of the main promoters of both ISRD surveys in 1992 and 2006.

                                                                                                                                                                                                                                                                                                              21  The hypothesis of the normality of delinquency had already been posed by Durkheim (1988 [1895]: 160) in the 19th Century. However, Durkheim considered crime as normal in the sense that it is hard to imagine a society without deviants; but, as research has shown, crime is not normal in the sense that anyone commits serious offences.

                                                                                                                                                                                                                                                                                                              22  An interesting analysis of some aspects of this change in England and the United States has been done by Garland (2001).

                                                                                                                                                                                                                                                                                                              23  The depletion in SRD surveys during that period (late 70s et early 80s) may be also have been caused by the concurrent development of the first important European programs of victimisation surveys, in Britain, the Netherlands, Sweden...: the energy invested in this innovations may have been diverted from SRD surveys.

                                                                                                                                                                                                                                                                                                              24  For a detailed review of these studies, see Aebi (2006, 162-164 and forthcoming).

                                                                                                                                                                                                                                                                                                              25  For a detailed review of these studies, see Aebi (2006, 169-177 and forthcoming).

                                                                                                                                                                                                                                                                                                              26  It is also possible to test alternate forms of the instrument on the same group. This technique is known as multiple forms (Hagan, 2005, 299).

                                                                                                                                                                                                                                                                                                              Date of publishing :

                                                                                                                                                                                                                                                                                                              01 march 2009

                                                                                                                                                                                                                                                                                                              Paper ISBN :

                                                                                                                                                                                                                                                                                                              978 2 917565 27 8

                                                                                                                                                                                                                                                                                                              To cite this document

                                                                                                                                                                                                                                                                                                              Marcelo F. Aebi, «Brochure n°2bis - Self-Reported Delinquency Surveys in Europe», CRIMPREV [En ligne], CRIMPREV programme, Brochures, URL : http://lodel.irevues.inist.fr/crimprev/index.php?id=177

                                                                                                                                                                                                                                                                                                              Contacts :

                                                                                                                                                                                                                                                                                                              Marcelo F. Aebi - Université de Lausanne - École de sciences criminelles - Institut de Criminologie et de Droit Pénal - Quartier UNIL Sorge-Batochime - CH 1015 LAUSANNE. E-mail : Marcelo.Aebi@unil.ch