Self-Reported Offending Research Paper

This sample Self-Reported Offending Research Paper is published for educational and informational purposes only. If you need help writing your assignment, please use our research paper writing service and buy a paper on any topic at affordable price. Also check our tips on how to write a research paper, see the lists of criminal justice research paper topics, and browse research paper examples.

Offending is commonly measured by asking people to admit whether they have committed each of a specified number of delinquent acts such as burglary, theft, robbery, assault, and vandalism. A key question is: How reliable and valid are these admissions? Given that people may conceal, exaggerate, or forget their offenses, how accurate are self-reports as a measure of actual delinquent behavior? This research paper will focus on self-reported offending (SRO) surveys of large community samples (at least several hundred), to address this question. The extent to which these reports are differentially valid by gender and race will also be explored. There is no space to review self-report surveys of prisoners or the literature on self-reported drug use.

Self-reports of offending have been used for a number of years, initially to uncover “hidden delinquency,” but more recently as a complementary measure to indexes of official offending (such as arrests or convictions). Self-reports of offending are reliable, since measures administered at different times produce similar results. The validity of self-reports is usually measured by comparing them with official arrests or convictions. Individuals who self-report offenses are more likely to have official records than those who do not report offenses, and individuals with officially recorded crimes usually admit these crimes. Crucially, self-reported offenses predict future arrests and convictions for the same crimes, even among people with no official record. Overall, the research suggests that self-reports provide a valid measure of offending for white males. Validity is sometimes lower for other demographic categories (e.g., females, Asians, and African-Americans), but these findings are far from consistent. More methodological research on self-reported offending is needed, especially on the differential validity of self-reported offending (SRO) by various demographic categories.

History Of The Use Of SRO

For many years, people in surveys have been asked to report offenses that they have committed that have not necessarily been detected by the police, in order to obtain information about “hidden delinquency.” For example, Burt (1929) in England and the Glueck and Glueck (1968) in the United States obtained this information from their interviewees and from other informants such as parents and teachers. However, the use of structured SRO questionnaires began in the 1940s. One of the most influential early studies was by Wallerstein and Wyle (1947), who distributed a 49-item mail questionnaire to nearly 1,700 adults in New York. The main finding of this research was that offending was very common: 99 % of adults admitted at least one offense, and even ministers of the church admitted an average of 8 offenses each. Reflecting on this, however, Wallerstein and Wyle (1947) concluded that acts that were technically offenses in this study were often quite trivial.

The major influence of SRO research on criminology really began in the 1950s with the research of Short and Nye (1958). Perhaps the main reason why their SRO survey was influential was because they found no relationship between socioeconomic status (SES) and self-reported delinquency. However, when they compared training school boys with high school boys, they found that the institutionalized delinquents came from much lower SES backgrounds. The implication, therefore, was that official processing was biased against low SES people – a finding that fitted well with the prevailing theorizing at the time. This result is credited with triggering the “self-report revolution,” with criminologists enthusiastically embracing the self-report method, often using the same items as Short and Nye.

Gold (1970) carried out a very important SRO survey in Michigan. This study is noteworthy because of the care taken in addressing methodological issues such as random sampling, efforts to reduce attrition, and the comparison of self-reports with informant reports of offending. The major findings of this study were that three-quarters of offenses were committed with others (most commonly peers) and that getting caught by the police was followed by an increase in offending (in agreement with labeling theory). Hirschi (1969 was the first to use the self-report method to test a theory about the causes of delinquency, and this seminal and highly cited study inspired a number of similar cross-sectional surveys designed to investigate causes using self-reports.

National SRO surveys quickly followed in the United States (e.g., Williams and Gold 1972), and these provided an alternative measure of juvenile crime rates to the official arrest records. These surveys were in turn followed by the ambitious longitudinal US National Youth Survey (NYS), beginning in 1977 and continuing to the present day (Elliott et al. 1985). This survey is one of the most important sources of self-report data on criminal careers and the causes of offending, but there are numerous other influential American prospective longitudinal studies which have used the self-report methodology, such as the Causes and Correlates projects in Pittsburgh, Denver, and Rochester; the Seattle Social Development Project; and the Oregon Youth Study (for details about all these surveys, see Farrington and Welsh 2007, Chap. 2). SRO surveys have been used in many other countries, with the Pan-European International Self-Report Delinquency Study (Junger-Tas 2010) being particularly noteworthy. The most comprehensive methodological work on these surveys has been conducted in the United States and England.

History Of Research On Reliability And Validity

The first attempt to review the reliability and validity of SRO surveys was in a conference report by Hardt and Bodine (1965). They noted (p. 15) that little was known about these topics because sociologists preferred substantive research to methodological examination. The first systematic assessment of SRO surveys on standard psychometric criteria such as questionnaire construction, administration procedure, objective scoring, norms for various populations, internal consistency, retest stability, and concurrent and predictive validity was completed by Farrington (1973). He published the first demonstration that an SRO survey had predictive validity. The next important methodological and substantive assessment of SRO surveys was completed by Hindelang et al. (1981), who concluded that the reliability and validity of SRO surveys was quite good in comparison with other methods and did not vary much with the method of administration (questionnaire or interview, anonymous or not). Unfortunately, this influential and highly cited book caused a great decrease in methodological SRO research, because it was now assumed that the validity of SRO surveys had been established for all time and that methodological research was not needed. From then on, the focus was on obtaining substantive results, using self-reports as the main outcome measure in criminology.

Reliability Of SRO Surveys

Reliability refers to the consistency of a measure in providing similar results under similar conditions, but this concept is generally more relevant when attempting to assess attitudes or other difficult-to-observe social constructs, as opposed to the recall of actual behavior. However, there are two criteria by which the reliability of SRO surveys can be assessed: test-retest stability (the degree to which results are consistent from one questionnaire administration to the next) and internal consistency (the degree to which items on a questionnaire all measure the same underlying construct), with the former arguably being more important than the latter. This is because it would reflect poorly on SRO surveys if the same person reported a different profile of offending during a specified time period when asked on two separate occasions (low test-retest reliability). When evaluating the test-retest reliability of SRO surveys, the time lag between the two administrations is important. If this time period is too short (e.g., a few days), participants could simply be recalling what they had said previously, whereas if the time lag is too long (e.g., several weeks), participants may forget the offenses that they had committed or may report new offenses committed in the interim.

A number of studies have evaluated the testretest reliability of SRO measures (with various time lags). Although the results varied depending on the number and types of offenses enquired about (and the methods of scoring), they suggest that these measures are at least as reliable as measures of attitudes (Thornberry and Krohn 2003). For example, Huizinga and Elliott (1986 reinterviewed a random selection of 177 participants from the NYS 4 weeks after the initial completion of the SRO survey. They found that crime index offenses (which are more serious and less frequent) had very high reliability, while less serious and more frequent offenses (e.g., public disorder offenses) had lower but still acceptable reliability. They also did not find any consistent variation in test-retest reliability over gender, race, social class, or involvement in delinquency.

Validity Of SRO Surveys

The key question for SRO surveys is validity: To what extent do self-reports produce an accurate estimate of the true number of offenses committed? And how accurately do self-reports measure the prevalence, frequency, and seriousness of offending? Setting aside the less important issues of content and construct validity, the validity of self-reports is usually assessed by comparing them with some external criterion of offending. The comparison can be concurrent (measure and criterion at the same time) or predictive (measure before criterion).

The main problem centers on what to use as an accurate external criterion of offending. Unlike drug use, for example (Harrison and Hughes 1997), there are no physical traces of burglary or shoplifting in hair, blood, or urine. Some researchers have compared SRO in the usual conditions and when people are told that their lying will be detected physiologically, and generally admissions increase in these physiological conditions. There is some evidence that admissions of problematic behavior are also greater in anonymous conditions, with audio computer-assisted interviews yielding the highest response rate (Tourangeau and Smith 1996).

Notwithstanding the fact that self-reports were intended to overcome some of the perceived deficiencies of official records, SRO results are usually validated against arrests or convictions. It is possible to compare self-reports with parent, teacher, or peer reports of offending and also to compare reports of offending with direct behavioral measures of actual offending, but there have been relatively few of these types of studies (see Farrington et al. 1980). Generally, the most important validity checks are in relation to official records.

There are two general ways in which the concurrent validity of SRO has been evaluated. The first is to compare self-reported offending (e.g., reporting breaking into a house with the intention of stealing) with official records (e.g., an arrest or conviction for burglary). The second is to examine the fraction of those who are known to have an official record who self-report official offending or to compare self-reported arrests or convictions with official records of arrests or convictions.

The level of correspondence between SRO and official records is influenced by the severity and frequency of the delinquent behavior, with the expectation that more severe and more frequent behaviors would be more likely to result in an official record and more likely to be self-reported. For example, Maxfield et al. (2000) found that 59 % of those with one arrest self-reported that they had been arrested, compared with 73 % of those with 2–4 arrests and 85 % of those with five or more arrests (vs. 21 % of those with no arrests).

Concurrent Validity Of SRO

Overall, the correlations between self-reports of offending and official records of arrests or convictions are generally substantial and statistically significant. However, correlations are not very good assessments of the validity of SRO surveys because most self-reported offenses will not result in an official record. The correspondence between self-reported offending and official records (both dichotomized) can be illustrated using a 2×2 table (Table 1).

Self-Reported Offending Research Paper

In cell a (SR Yes, OR Yes) and cell d (SR No, OR No), the two measures of offending correspond, and this is treated as accurate responding. Cell b (SR No, OR Yes) is usually treated as concealment of offending by the individual and as such is commonly referred to as underreporting. There are of course other reasons why someone might not report an offense that was officially recorded, including forgetting and being innocent of that particular offense (e.g., because of plea bargaining). Cell c (SR Yes, OR No) represents one of the key purposes of SRO surveys, to identify offenders and offenses that have not been officially recorded. A lack of correspondence here can therefore not necessarily be attributed to a lack of validity, although this cell count might be influenced by exaggerated reporting. However, when self-reports of arrests or convictions are compared to official records, cell c might indicate over-reporting (or faulty records).

When examining the validity of SRO, it might be expected that the probability of an arrest or conviction would increase with the intensity (e.g., frequency or seriousness) of self-reported delinquency. For example, Jolliffe et al. (2003) investigated the validity of self-reports by comparing these to the court referrals of over 800 boys and girls (age 11–17) in the prospective longitudinal Seattle Social Development Project (SSDP). A total of 626 youths self-reported at least one offense during this age range, and 230 of these (37 %) had a court referral. In contrast, 101 reported that they had not committed any offense, and only 16 of these (16 %) had a court referral. This comparison was statistically significant and suggested that court referrals were much more common among those who self-reported an offense than among those who denied offending. This measure of concurrent validity was highest for those who self-reported burglary (a serious offense) or drug offenses (the most frequent). A number of other studies have also demonstrated the concordance between SRO and official records.

Some researchers have questioned, however, whether self-reports of offending are equally valid across demographic categories. Generally, male/female and black/white ratios in American SRO surveys are lower than in official records. One possible reason for this is that male and black youth are more likely to conceal their offenses than females and white youths, but there are also a number of other potential explanations. Since the prevalence of offending is usually much greater in self-reports than in official records, there may be a “ceiling effect”; for example, if 50 % of white youth self-reported offending, the maximum possible black/white ratio in self-reports would be 2:1, whereas if 20 % of white youth were convicted, the maximum possible black/white ratio in convictions would be 5:1. In addition, self-reported offenses may not be comparable to (may be less serious than) official offenses.

Methods of scoring may influence race and gender ratios. Elliott and Ageton (1980) found that these ratios were much greater when the scoring allowed for large numbers of admitted crimes, rather than the maximum category being “three or more,” for example. They concluded (p. 107) that race and gender differences were “more extreme at the high end of the frequency continuum, that part of the delinquency continuum where police contacts are more likely.” Another possibility is that official records may reflect police and court bias against male and black youth, and more recorded male and black youth might in fact be innocent. Alternatively, male and black youth may be disproportionately arrested because of their aggressive demeanor. In a validity test, Hindelang et al. (1981) investigated how many offenses known to the police were self-reported and found that white males failed to report 20 % of serious offenses, compared with 57 % for black males, 50 % for white females, and 59 % for black females.

Similar findings on differential validity of SRO have been reported in other studies, but the pattern is far from consistent. For example, using data from the Pittsburgh Youth Study, Farrington et al. (1996) discovered that, although the self-reports of both Caucasian and African-American boys were significantly associated with their court referrals, the relationship was stronger for Caucasians. However, in the SSDP, Jolliffe et al. (2003) found that concurrent validity was lowest for Asians and females and that African-American males had the highest concurrent validity.

The comparison between SRO and official records has also been studied for racial and ethnic minorities in other countries. In a longitudinal study in New Zealand, Fergusson et al. (1993) found that children of Maori or Pacific Island descent were 2.9 times more likely to have a police record than children of European descent, but were only 1.7 times more likely to offend according to parent and self-reports. The researchers attributed these findings to ethnic biases in citizen reporting and police recording. In a national SRO survey in the Netherlands, Junger (1989) discovered that Moroccan and Turkish boys with police records were less likely to admit delinquency than indigenous Dutch boys or Surinamese boys. She pointed out that concurrent validity was not necessarily lower for ethnic minorities, because the Surinamese boys were black.

Another measure of the concurrent validity of SRO is the percentage of those with a court record for a particular offense who admit committing the same offense. For example, of those with a court referral for burglary, what percentage admit that they have committed a burglary? The association between self-reports and official records in this type of comparison is generally quite high. For example, in the SSDP, Jolliffe et al. (2003) found that, of the 246 youth referred to court, 94 % admitted committing at least one offense. The probability for specific offenses was lowest for robbery (38 %) and highest for marijuana use (100 %). Figures ranging from 54 % to 90 % have been discovered in similar studies, suggesting that most officially identified offenders self-report their particular offenses, but some differential validity by gender and race has been found. For example, in the study by Jolliffe et al. (2003), males were more likely to admit than females, and Asians were less likely to admit than either African-Americans or Caucasians. There were no differences in concurrent validity between African-Americans and Caucasians.

Concurrent Validity Of Self-Reports Of Arrests Or Convictions

A more direct method of investigating validity is to compare self-reports of arrests or convictions with official records of arrests or convictions. For example, West and Farrington (1977) found that, at age 18, 94 % of convicted boys admitted that they had been convicted, while only 2 % of unconvicted boys claimed to have been convicted. Of offenses leading to conviction, 53 % were reported accurately, 34 % were reported but minimized, 7 % were reported but exaggerated, and 6 % were not reported. The probability of arrested or convicted offenders admitting their crimes is also high in other studies, with those who have a greater number of convictions being generally more willing to self-report offenses (e.g., Brame et al. 2004).

Again, there is some indication of differential validity by gender and race in the self-reporting of official contacts. Hindelang et al. (1981) found that 76 % of white male official delinquents reported that they had been picked up by the police, compared with 52 % of white females, 50 % of black males, and only 30 % of black females. These results fostered the widespread belief that the self-reports of black youths (especially those who were the most delinquent) were less valid because of concealment. Other studies, however, have not found race differences in reporting arrests. Maxfield et al. (2000) discovered that 75 % of convicted white youths admitted being arrested, compared with 70 % of black youths (not significantly different), and 76 % of convicted males admitted being arrested, compared to 60 % of convicted females (significantly different). More recently, in a large sample of Dutch adolescents, Van Batenburg-Eddes et al. (2012) found that 62 % of those with official police contacts self-reported a police contact (interrogation in the police station), but this was somewhat higher for older (age 14–15) adolescents compared to younger (age 12–13) ones: 65 % compared to 57 % (significantly different). They also found that indigenous Dutch (65 %) and Surinamese (70 %) boys and girls were significantly more likely than Moroccan boys and girls (46 %) to report their police contacts.

The differential validity of the self-reporting of arrests was explored by Krohn et al. (2013) using data from the Rochester Youth Development Study. Overall, 61 % of arrests were admitted. Arrested males (64 %) were more likely to admit their arrests than arrested females (46 %). Caucasian males were likely to over-report arrests, while African-American males were likely to underreport arrests. However, Krohn et al. (2013) also demonstrated that those who were arrested more frequently were least likely to report each arrest, regardless of race or gender. It could be that frequently arrested offenders had more difficulty in remembering all of their arrests.

There have been relatively few studies of how concurrent validity varies with age. However, Kirk (2006) compared age-crime curves based on official and self-reported arrests. The two curves were generally similar, but the self-report curve had a higher and earlier peak (at ages 18–19) than the official curve (at age 21). Kirk also found that the black/white prevalence ratio was higher for official arrests (2.3–1) than for self-reported arrests (1.3–1), but this was because whites over-reported their arrests; 21 % self-reported an arrest compared to 13 % who had an official arrest.

Predictive Validity

Predictive validity is more impressive than concurrent validity, because being convicted may itself lead to an increase in the probability of admitting delinquent acts (Farrington 1977), perhaps because a person may assume that the researcher will know about convictions and therefore that concealment is futile. It is to be expected that current self-reported offending (of past acts) will predict future convictions, because many current self-reported offenders are also convicted and because future convictions are of course predicted by past convictions. A better test of the predictive validity of SRO surveys is to investigate the extent to which self-reports predict future convictions among currently unconvicted people.

There have only been four investigations (using three prospective longitudinal studies) of the predictive validity of SRO. In the Cambridge Study in Delinquent Development (a survey of over 400 London boys), Farrington (1973) showed that, for unconvicted boys, a measure of self-reported variety of offending at age 14 significantly predicted their probability of conviction in the following 3 years. Farrington (1989) later repeated this analysis for specific types of offenses. For example, among boys not convicted of burglary up to age 18, significantly more of those who self-reported burglary up to age 18 were subsequently convicted of burglary up to age 32, in comparison with the remaining boys who denied committing burglary up to age 18 (20 % as opposed to 2 %). Similar results were obtained for vehicle theft, assault, vandalism, and drug use, but not for shoplifting.

In the first test of predictive validity using American data, Farrington et al. (1996) classified the self-reported offending of Pittsburgh Youth Study boys into four categories based on seriousness of offending and compared this with their later petitions to court. For both samples, the predictive validity of self-reports was high and statistically significant (except for drug offenses in the oldest sample). Of the boys who reported the most serious forms of delinquency, 29 % were petitioned to court for Index Violence, compared to only 7 % of the boys who reported no delinquency. This difference was statistically significant, again showing that SRO had predictive validity. Farrington et al. (1996) also examined the extent to which predictive validity varied according to race, but found that predictive validity was similar for Caucasians and African-Americans. Perhaps the most important finding with regard to the validity of SRO was that there was a substantial increase in predictive validity by combining the reports of the boys with those from other sources (mothers and teachers).

Jolliffe et al. (2003) examined the predictive validity of SRO for males and females of different racial backgrounds. In the SSDP, they found that predictive validity was high and statistically significant for most offenses (except robbery and vehicle theft). It was generally higher for males than for females and highest for Caucasians and lowest for Asians (predominantly Chinese and Filipinos). The self-reports of Asian and African-American females had the lowest predictive validity.

Summary And Future Directions For Research

Self-reports of offending were initially developed to reveal the dark figure of crime, but they have evolved to be one of the most common outcome measures used in criminological research. On psychometric criteria, self-reports provide a reliable measure of offending. In addition, most research suggests that self-reports are concurrently valid, in that those who self-report offenses or arrests are more likely to have official records than those who do not self-report, and those who have been officially recorded tend to admit their offenses. Importantly, self-reports have predictive validity; among those with no official record, those who self-report offenses are more likely to be officially recorded in the future than those who do not self-report. This is good news for criminological researchers. The bad news is that the overall validity of self-reports conceals the fact that they appear to be more valid for some demographic categories than for others. Self-reports are clearly valid for white males, but they are less consistently valid for females, for Asians, and for African-Americans. Little is known about validity at older ages. That said, there are indications that other factors (e.g., number of arrests) might be more relevant to validity than demographic categories. More research is needed on the validity of self-reports of offending, and of self-reported arrests and convictions, to establish the optimal conditions for enquiring about these topics. Randomized experiments, such as that conducted by Enzmann (2013), are needed to study the effects of different methods of administration on the reliability and validity of different demographic categories. It is also important to establish to what extent differential validity might reflect biases in official processing. Since most criminological research is based on self-reports, it is very important that their reliability and validity should be measured.

Bibliography:

Brame R, Fagan J, Piquero AR, Schubert CA, Steinberg L (2004) Criminal careers of serious delinquents in two cities. Youth Violence Juv Justice 2:256–272
Burt C (1929) The young delinquent. Appleton, New York
Elliott DS, Ageton SS (1980) Reconciling race and class differences in self-reported and official delinquency. Am Sociol Rev 45:95–110
Elliott DS, Huizinga D, Ageton SS (1985) Explaining delinquency and drug use. Sage, Beverly Hills
Enzmann D (2013) The impact of questionnaire design on prevalence and incidence rates of self-reported delinquency: results of an experiment modifying the ISRD-2 questionnaire. J Contemp Crim Justice (in press)
Farrington DP (1973) Self-reports of deviant behavior: predictive and stable? J Crim Law Criminol 64:99–110
Farrington DP (1977) The effects of public labelling. Br J Criminol 17:112–125
Farrington DP (1989) Self-reported and official offending from adolescence to adulthood. In: Klein MW (ed) Cross-national research in self-reported crime and delinquency. Kluwer, Dordrecht, pp 399–423
Farrington DP, Welsh BC (2007) Saving children from a life of crime: early risk factors and effective interventions. Oxford University Press, Oxford, UK
Farrington DP, Knapp WS, Erickson BE, Knight BJ (1980) Words and deeds in the study of stealing. J Adolesc 3:35–49
Farrington DP, Loeber R, Stouthamer-Loeber M, Van Kammen WB, Schmidt L (1996) Self-reported delinquency and a combined delinquency seriousness scale based on boys, mothers and teachers: concurrent and predictive validity for African Americans and Caucasians. Criminology 34:493–517
Fergusson DM, Horwood LJ, Lynskey MT (1993) Ethnicity and bias in police contact statistics. Aust N Z J Criminol 26:193–206
Glueck S, Glueck ET (1968) Delinquents and non- delinquents in perspective. Harvard University Press, Cambridge, MA
Gold M (1970) Delinquent behavior in an American city. Brooks/Cole, Belmont
Hardt RH, Bodine GE (1965) Development of self-report instruments in delinquency research. Youth Development Center, Syracuse University, Syracuse
Harrison L, Hughes A (eds) (1997) The validity of self-reported drug use. US National Institute on Drug Abuse, Rockville
Hindelang MJ, Hirschi T, Weis JG (1981) Measuring delinquency. Sage, Beverly Hills
Hirschi T (1969) Causes of delinquency. University of California Press, Berkeley
Huizinga D, Elliott DS (1986) Reassessing the reliability and validity of self-report measures. J Quant Criminol 2:293–327
Jolliffe D, Farrington DP, Hawkins JD, Catalano RF, Hill KG, Kosterman R (2003) Predictive, concurrent, prospective and retrospective validity of self-reported delinquency. Crim Behav Ment Health 13:179–197
Junger M (1989) Discrepancies between police and self-report data for Dutch racial minorities. Br J Criminol 29:273–284
Junger-Tas J (2010) The significance of the international self-report delinquency study (ISRD). Eur J Crim Policy Res 16:71–87
Junger-Tas J, Marshall IH (1999) The self-report methodology in crime research: strengths and weaknesses. In: Tonry M (ed) Crime and justice: a review of research, vol. 25. University of Chicago Press, Chicago, pp 291–367
Kirk DS (2006) Examining the divergence across self-report and official data sources on inferences about the adolescent life-course of crime. J Quant Criminol 22:107–129
Krohn MD, Thornberry TP, Gibson CL, Baldwin JM (2010) The development and impact of self-report measures of crime and delinquency. J Quant Criminol 26:509–525
Krohn MD, Lizotte AJ, Phillips MD, Thornberry TP, Bell KA (2013) Explaining systematic bias in self-reported measures: factors that affect the under and over-reporting of self-reported arrests. Justice Quart (in press)
Maxfield MG, Weiler BL, Widom CS (2000) Comparing self-reports and official records of arrests. J Quant Criminol 16:87–110
Short JF, Nye FI (1958) Extent of unrecorded juvenile delinquency: tentative conclusions. J Crim Law Criminol Police Sci 49:296–302
Thornberry TP, Krohn MD (2003) Comparison of self-report and official data for measuring crime. In: Pepper JV, Petrie CV (eds) Measurement problems in criminal justice research: workshop summary. National Academy Press, Washington, DC, pp 43–94
Tourangeau R, Smith TW (1996) Asking sensitive questions: the impact of data collection, mode, question format, and question context. Public Opin Q 60:275–304
Van Batenburg-Eddes T, Butte D, Van De Looij-Jansen P, Schiethart W, Raat H, De Waart F, Jansen W (2012) Measuring juvenile delinquency: how do self-reports compare with official police statistics? Eur J Criminol 9:23–37
Wallerstein JS, Wyle CJ (1947) Our law-abiding lawbreakers. Probation 25:107–112
West DJ, Farrington DP (1977) The delinquent way of life. Heinemann, London
Williams JR, Gold M (1972) From delinquent behavior to official delinquency. Soc Probl 20:209–229