Antisocial Behaviour and Violent Delinquency Among Boys with a Migration Background: A German Panel Study

Abstract

Immigrant boys show higher rates of antisocial behaviour. However, results of previous studies showed some contradictory findings in terms of intercultural differences in antisocial behaviour. In our study, we used an intercultural comparison of antisocial behaviour based on two different definitions of migration status (nationality vs. migration background). According to migration background, however not according to nationality, significant but small differences were found only in violent delinquency. A longitudinal mediator model based on the disintegration approach (Anhut & Heitmeyer, 2000) was examined in order to contribute to our understanding of the background of violent delinquency in immigrant boys. The data were from a German panel study conducted in the cities of Dortmund and Nuremberg. The results of the mediator model showed that perceived discrimination and negative parenting affect violent delinquency indirectly through violence attitudes, self-control, and peer delinquency. The findings suggest that preventive measures against violent delinquency should focus on these indirect effects and migrant-specific variables.

Keywords

Antisocial behaviour violence immigrant adolescents social disintegration approach longitudinal mediator analysis

In most societies, higher rates of antisocial behavior are attributed to immigration. According to the 2014 German General Social Survey (ALLBUS), nearly half of the respondents (49.7%) believed that immigration increases the crime rates (Naplava, 2018). Individual violent crimes of migrants have received great attention by media and society, thus strengthening the public image of “criminal foreigners” (Uysal, Link, & Weiss, 2016).

In contrast to subjective judgments about the antisocial behavior of foreigners, scientific research should contribute to an objective and more detailed evaluation of this issue, taking into account causal mechanisms and differential effects.

Previous studies on antisocial behaviour which included ethnic comparisons have yielded contradictory results for German-speaking countries (Strohmeier, 2007). Some studies pointed to a higher level of violence among adolescents with migration background (Baier, Pfeiffer, & Windzio, 2006; Rabold & Baier, 2011), while others either showed completely opposite results or found no clear differences between local and foreign adolescents or adolescents with a migration background (Boers, Walburg, & Reinecke, 2006; Othold & Schumann, 2003). Strohmeier (2007) reported that the ethnic background is usually dealt with as a ‘secondary issue’ in studies of antisocial behavior in German-speaking countries and that the samples of these studies included a low proportion of migrant adolescents or adolescents with a migration background.

Furthermore, data on crime rates and immigration vary depending on the definition of ‘migrant background’; e.g., German official data (Bundeskriminalamt, 2017) only take into account nationality, thus excluding all persons with a migration background who have acquired German citizenship. Up to date, there is a lack of detailed migrant-specific analyses in criminological research.

Disintegration Approach

The disintegration approach aims to explain antisocial behaviour and is based on the approach of Anhut and Heitmeyer (2000). It does not deal specifically with migrant criminal behaviour, but it can be effectively applied to this topic and provides the theoretical background for several criminological studies in Germany (Baier, 2005; Heitmeyer et al., 2012; Rippl, Baier, Kindervater, & Boehnke, 2005).

Heitmeyer, one of the two German sociologists who developed the social disintegration theory, worked with the Institute for Interdisciplinary Research on Conflict and Violence (IKG)-Youth Panel data on the basis of this theory, and this group investigated the violent behaviour of Turkish, German, and German resettler adolescents. This research will be discussed in the following pages, but first, disintegration theory will be explained briefly.

According to the disintegration approach, experience or fear of disintegration increases conflicts and weakens the society’s regulation capabilities. Thus, the disintegration approach explains violent behaviour as a result of deficient integrative capacities of modern societies. On the other hand, successful integration leads to positional, moral, and emotional recognition, resulting in a self-definition of being part of the social collective and willingness to accept social norms.

Well-functioning social integration requires coping with tasks on the structural, institutional, and social-emotional level. The social-structural level refers to ensuring participation in socially generated material and cultural goods (e.g. work and education). On the institutional level, the right of political participation, which means the determination of socially relevant procedures, is essential. The social-emotional level means self-actualization in emotional relationships, such as those involving family, friend groups, or the neighbourhood (Anhut & Heitmeyer, 2000, 2009; Babka von Gostomski, 2003a).

The Model of the Study

Our study was based on Babka von Gostomski’s model (2003b). This model applies the disintegration approach to the question of migrant criminality, operationalizing its core concepts as follows: The social-structural dimension is represented by the type of school attended by the adolescent and the father’s highest educational degree. The institutional level is represented by nationality, discrimination experiences, and (dis)trust in the justice system. The social-emotional dimension is represented by the adolescents’ relationship with their parents and by experienced parental education.

Anhut and Heitmeyer (2000) emphasized that the result of disintegration experiences depends on additional influential factors, such as attribution styles or affiliations with specific milieus (Anhut, 2005). Therefore, Babka von Gostomski (2003b) added two factors that mediate the relationship between disintegration and violent delinquency: firstly, retribution-oriented conflict style; secondly, a highly cohesive peer network.

Babka von Gostomski (2003b) tested this model using IKG-Youth Panel data. This study was carried out with 11,000 pupils in the tenth grade of all school types in the Federal State of North Rhine-Westphalia. Babka von Gostomski (2003b) examined the violent behaviour of Turkish, German, and ethnic German resettler (German: ‘Aussiedler’) boys. According to the study, the type of school visited, as an indicator of the social-structural dimension, influenced the violence level of young people, as students of lower-track schools showed a higher risk of violent delinquency than students of higher-track schools. As indicators of the institutional dimension, citizenship had no influence on violent crime, whereas discrimination experiences and (dis)trust in law treatment increased the risk of violent action. Taking the indicators of the social-emotional dimension into account, lack of trust in parents and a harsh and inconsistent parenting style increase the risk of violence. The mediation factors availability of retribution-oriented conflict resolution strategies and membership in highly cohesive cliques enhanced the risk of violent behaviour.

In our study, we reassessed the model proposed by Babka von Gostomski (2003b), making adjustments as follows: as the disintegration approach stresses the relevance of social competence as another mediating factor (Anhut & Heitmeyer, 2000), the present study aimed to include this factor in the operationalization of the disintegration approach. As self-control is considered a relevant aspect of social competence (Kanning, 2009), we focused on this aspect. The remaining mediating factors violence attitudes and delinquency of friends take a supplementary role in the retribution-oriented conflict resolution strategy and membership in a highly cohesive clique. Another difference in the operationalisation between the current study and Babka von Gostomski’s (2003b) is that the variable trust in the justice system could not be examined as an indicator of the institutional level in our study. Moreover, the model of the current study is examined longitudinally using structural equation modelling (SEM) with two cohorts, whereas Babka von Gostomski tested his model cross-sectionally by using regression analyses in one cohort that was comparable with the older cohort of this study. Figure 1 presents an overview of all variables that were examined in this study.

Figure 1

Operationalisation of current study.

In order to achieve comparability with bright field data, as well as to avoid the loss of the diversity of the ethnic origins, two categories (nationality and migration background) were examined separately in this work. Migrant background is defined according to Babka von Gostomski, considering the nationality of the respondents and their parents, as the country of birth of the respondents and their parents and also the language spoken at home.

Only male adolescents were included in the analyses because there is increasing evidence that female delinquency follows its own developmental processes and should be addressed as a distinct phenomenon (Hoyt & Scherer, 1998; Oberwittler, 2007). Therefore, the model was tested only with boys with a migration background.

Research questions:

As a result of the indefinite state of research findings as stated above, the question whether crime rates are different for boys with German vs. other nationalities, or boys with vs. boys without migration background is to be examined.

In a second step, our revised version of the social disintegration model (Babka von Gostomski, 2003b) as shown in Figure 1 will be tested using mediator analyses. The model will be tested separately for the younger and older cohorts, because they supposedly follow different developmental trajectories of antisocial behaviour (Moffitt, 1993). We chose violent delinquency as the dependent variable of our mediator models as in the study of Babka von Gostomski. With our model, we aimed to contribute to a better understanding of the background of immigrant boys’ violent and delinquent behavior.

Method

The sample was recruited via three survey waves (with yearly follow-ups) from the longitudinal subproject A2 “Chances and Risks in the Life Course” of the Collaborative Research Centre (CRC) “From Heterogeneities to Inequalities” at Bielefeld University, which was funded by the German Research Foundation. The study used a cohort-sequential design with two cohorts (5th and 9th grade) and data collection in two different German major cities (Nuremberg and Dortmund). Therefore, two cohorts of students in Nuremberg and Dortmund were assessed once a year in the 5th, 6th and 7th grades (cohort 1) and the 9th, 10th and 11th grades (cohort 2), respectively.

Sample

The initial sample of this subproject consisted of nearly 3,000 students from the 5th and 9th grade in Dortmund and Nuremberg in 2012. This article reports data from the second and third measurement points (2013 and 2014); at these measurement points, cohort 1 attended 6th and 7th grade, and cohort 2 attended 10th and 11th grade, respectively. Table 1 illustrates descriptive statistics for the cross-sectional and longitudinal samples. The cross-sectional samples are larger than the longitudinal sample for two reasons: 1. New participants could enter and leave the study at every wave; and 2. Especially in cohort 2, dropout was an issue as a significant number of students left school after 9th or 10th grade and could only be contacted by mail which led to lower response rates. However, students who dropped out of the study did not differ from those who continued in terms of violent delinquency (t(290) = – 0.831, p = 0.41, M = 0.26/0.32, SD = 0.69/0.65). In general, delinquency was not a significant predictor of drop-out in our study (for more information on drop-out analysis, see Weiss & Link, 2019). For more information on the sampling procedures, see Schepers and Uysal (2014), and Meinert and Uysal (2015).

Table 1

Descriptive Statistics of Cross-sectional and Longitudinal Samples

	Cross-sectional Samples			Longitudinal Samples
	6. G.	10. G.	7. G.	11. G.	6. & 7. G.	10. & 11. G.
Total	1609	1304	2180	1005	1252	795
Male	829	606	1070	423	607	315
♂Mean age (SD)	11.95 (0.71)	16.32 (0.82)	13.12 (0.74)	17.27 (0.78)	12.61 (0.94)	16.73 (0.99)
♂With migration background	504	309	682	203	348	147

Note. 1. (G): Grade; 2. due to missing data, the sum of male and female participants sometimes does not correspond to the total number; and 3. When determining the migration status, only those were included in the categories that gave the same information for two waves.

The sample that was used for the mediator analyses – longitudinal data on boys with a migration background – was limited. This is partly due to the fact that the target sample of this study was very specific. Table 1 compares the sample sizes of the A2 subproject and the sample of the current study in order to make the numbers more comprehensible. Due to the limited number of longitudinal data on immigrant boys, further migrant-specific considerations were not possible.

Survey Instruments

Both cohorts filled in similar questionnaires at every measurement point with only minor differences in wording; furthermore, some constructs were assessed in more detail in cohort 2 in order to keep the questionnaire shorter for the younger cohort (Reinecke et al., 2013).

In regard to the social-structural level of the disintegration approach, the type of school attended by the adolescents and (in cohort 2) the highest level of education of their parents were taken into account. This item was adopted from the Progress in International Reading Literacy Study (PIRLS) (Gonzalez & Kennedy, 2001).

As indicators of the institutional level, the citizenship of the respondents and their experiences of discrimination were noted. Our modified version of the perceived discrimination scale (Dogan & Strohmeier, 2013; Skrobanek, 2007) consisted of items on perceived discrimination within school (e.g. “I get the impression that teachers like me less than German students”; “In class, I got laughed at or insulted more often than German students”), and perceived discrimination outside school (e.g. “In public, I got laughed at or insulted more often than Germans”; “I feel discriminated by offices and authorities compared to Germans”). These items were assigned to subscales following exploratory factor analyses (see Table 2).

Table 2

Features of Perceived Discrimination Scale

Scale	Item number	Subscales	Reliability	Rating Scale	References
Perceived	4 (y), 4 (o)	pd within the school (f)	α (y) = 0.78, α (o) = 0.83	5-point Likert	Dogan &
discrimination	3 (y), 5 (o)	pd outside the school (f)	α (y) = 0.73, α (o) = 0.86		Strohmeier (2013);
(pd)	3 (y), 2 (o)	pd by teacher (s)	α (y) = 0.92, α (o) = 0.91		Skrobanek (2007)
	3 (y), 4 (o)	pd by classmates (s)	α (y) = 0.90, α (o) = 0.91
	3 (y), 5 (o)	pd outside the school (s)	α (y) = 0.78, α (o) = 0.87

Note. 1. (f) first survey wave, (s) second survey, (o) older cohort, (y) cohort; 2. the subscales are assigned according to retained factors after exploratory factor analyses; 3. Since older cohort respondents were expected to experience more discrimination outside the school, more items were surveyed in this group; 4. Since peers play a major role in adolescents and 11th graders are no longer in school at all times, four items for discrimination by classmates and only two items for discrimination by teachers were included in this group.

As indicators of the social-emotional level, parental education and attachment to parents were assessed. Parental education was measured by three subscales of the Alabama Parenting Questionnaire (Essau, Sasagawa, & Frick, 2006; Lösel et al., 2003). The subscale corporal punishment (4 items, α= 0.88– 0.91) contained questions about experienced violent education practices, e.g. “My parents give me a slap in the face”. The subscale inconsistent discipline (3 items, α= 0.55– 0.69) assessed lack of consequence in perceived parental education, e.g. “My parents threaten to punish me and then do not do it”. The subscale low supervision/control (4/5 items, α= 0.59– 0.78) contained questions on lack of parental supervision, e.g. “My parents get so busy that they forget where I am and what I am doing”. For the mediation models, an aggregate latent construct was formed from the three subscales of parental education, low control, inconsistent discipline, and corporal punishment, which fulfilled the prerequisites of preliminary analyses. This construct was named negative parenting. Attachment to parents was assessed by two scales from the Inventory of Parent and Peer Attachment (IPPA) (Armsden & Greenberg, 1987; Rollett, Werneck, & Hanfstingl, 2005): trust (4 items, α= 0.90– 0.92, e.g. “My parents trust me”) and communication (4 items, α= 0.55– 0.69, e.g. “My parents talk to me when they realize that something is depressing me”).

Further, the mediators acceptance of violence, self-control, and delinquency of friends were assessed (see Fig. 1). Acceptance of violence (Boers & Reinecke, 2007; Dünkel & Geng, 2003) was measured by 5 (cohort 1) and 9 (cohort 2) items, e.g. “A little bit of violence is just part of having fun” (α= 0.69– 0.73). The Grasmick Scale (Eifler & Seipel, 2001; Grasmick, Tittle, Bursik, & Arneklev, 1993) was used to measure different aspects of self-control: risky behaviour, impulsivity, irritability, and a preference for simple tasks. In this study, overall self-control was assessed by 10 (cohort 1) and 12 (cohort 2) items, e.g. “Others should rather leave me alone when I’m in a temper” (α= 0.74– 0.78). Lastly, peer delinquency was measured by seven items that were derived from the Peterborough Adolescent and Young Adult Development Study (PADS+) (Wikström, Oberwittler, Treiber, & Hardie, 2012) and the Crime in the Modern City Study (CRIMOC) (Boers & Reinecke, 2007), e.g. “What do you think, how often do your friends commit the following deed: Steal a bicycle?” (α= 0.84– 0.86).

As a dependent variable, participants were asked if they had committed one or more offenses in the last 12 months. The delinquency items of this survey were adapted from established German delinquency surveys (Boers & Reinecke, 2007; Lösel, 1975). We assessed 16 different offenses in cohort 1 and 19 different offenses in cohort 2.

For research question 1, we calculated 1-year offender rates for overall delinquency and its subscales violent delinquency (e.g., assault, robbery), property crime (e.g., theft from classmates, burglary), and vandalism (e.g., graffiti, scratching) for boys with German and other nationalities and boys with and without migration background, respectively.

In order to test our model of the disintegration approach, we used the accumulated 1-year prevalences (versatility) of violent delinquency, since the examination of the mediator model required continuous manifest variables (Hayes, 2017).

Statistical Analyses

In order to analyse the model in the study, preliminary analyses were necessary. First, crime prevalence rates of male youths with and without migration background, and with and without German nationality, were compared using χ2 tests. As preparation of model testing, we used the often-cited four steps of Baron and Kenney (1986), although there are several limitations to this approach (see MacKinnon, Fairchild, & Fritz, 2007).

The mediator analyses of migrant male respondents were performed longitudinally in both cohorts with the dependent variable violent delinquency. The independent variables and mediator variables were from the second survey wave, and the dependent variable was from the third survey wave.

Due to the violation of the normal distribution assumption of the dependent variable violent delinquency, weighted least squares mean and variance adjusted (WLSMV) was used as estimation method in Mplus, which enables robust parameter estimates (Flora & Curran, 2004, Reinecke, 2014). The analyses were conducted using IBM SPSS Statistics 23 software (IBM, Armonk, NY, USA), and the models were estimated with Mplus (Version 7.1) (Muthén & Muthén, 1998– 2012).

Results

Prevalence Rates

First of all, we examined whether German and non-German adolescents differed in terms of perpetration in the various areas of crime. The one year prevalence rates of overall delinquency for all male respondents were 32 % (6th grade), 30 % (7th grade), 34 % (10th grade), and 24 % (11th grade). There was no significant difference (p > 0.05) in delinquency in terms of nationality (data not shown). Differences were seen only on the basis of migration background, and they were limited to the 6th and 10th graders (p < 0.05) and only to violent crimes. However, these effect sizes were extremely small (ES < 0.10). The percentages of perpetrator rates of adolescents with and without migration background are demonstrated in Table 3.

Table 3

Percentages of Perpetrator Rates of Male Respondents with and Without a Migration Background in Various Offenses

	6 Grade		10 Grade		7 Grade		11 Grade
Delinquency Form	Native	WMB	Native	WMB	Native	WMB	Native	WMB
Vandalism	14	16	17	18	16	13	8	10
Property crime	21	22	27	25	23	24	15	20
Violent delinquency	11	18	14	20	13	17	9	12
Overall delinquency	31	35	36	37	32	36	24	27

Note. WMB: with migration background.

Examination of the Models

Following the procedure by Baron and Kenny (1986), the variables highest education level of parents, attended school type, attachment to parents, and citizenship of the respondents did not meet the requirements of the causal conditions for the models, which is why they were eliminated from the models. The mediator analyses of migrant male respondents were performed longitudinally in both cohorts with the dependent variable ‘violent delinquency’.

The model fit of the mediator model was good in the older cohort. The model of violent delinquency in the older cohort explained 36.1% of variance. The direct effect of the independent variables discrimination (β= 0.18) and negative education (β= –.29) were not significant. They both correlated positively and significantly with each other (r = 0.30, p < 0.001). Peer delinquency (β= 0.32, p < 0.001) and self-control (β= 0.32, p < 0.05) had the strongest influence on violent delinquency in the older cohort. Peer delinquency mediated a significant influence of negative education (β= 0.18, p < 0.05), and perceived discrimination (β= 0.14, p < 0.05) did so for violent delinquency. Self-control also mediated the effect of negative education (β= 0.48, p < 0.001) on violent delinquency. Violence acceptance was influenced by self-control (β= 0.23, p < 0.01) and had a significant effect on peer delinquency (β= 0.24, p < 0.01) and on violent delinquency (β= 0.25, p < 0.01). Negative education (β= 0.23, p < 0.01) and perceived discrimination (β= 0.18, p < 0.01) were both significantly predictive of violence acceptance. Figure 2 presents these results. In the models shown in the figures, the values given are all standardized coefficients. However, coefficients that are under.10 are not mentioned.

Figure 2

Longitudinal mediator model of violent delinquency in boys with migration background in the older cohort.

The model of violent delinquency in the younger cohort had poor model fits and explained lower variance (14.2%). In contrast to the model in the older cohort, negative parenting had a marginally significant direct influence on violent delinquency (β= 0.19, p < 0.10). Moreover, there was no mediation effect of peer delinquency. The influence of self-control on violent delinquency was not significant (β= –.02). The acceptance of violence mediated significant influences of discrimination (β= 0.26, p < 0.001) and negative parenting (β= 0.20, p < 0.001) on violent delinquency (β= 0.28, p < 0.05). The models in the younger and older cohorts differed in terms of influences of coefficients of variables, explained variances, and model fits. For more details about the model of the younger cohort see Fig. 3.

Figure 3

Longitudinal mediator model of violent delinquency in boys with migration background in the younger cohort.

Discussion

The present study deals with two research questions: First, we examined differences in crime rates depending on nationality and migration background. Second, our main aim was to test a revised version of the disintegration model as a theoretical framework of violent criminality in adolescents with a migration background.

In order to eliminate a probable reason for conflicting results of previous research (Baier, Pfeiffer, & Windzio, 2006; Boers, Walburg, & Reinecke, 2006), the differentiation of native and non-native groups was provided in two ways: via nationality and via an expanded definition of migration background. On the basis of nationality, the comparison of German and non-German adolescents revealed no significant differences in any of the areas of delinquency, regardless of the age of the subjects. Similar results were found by Othold and Schumann (2003) who reported identical violent perpetrator rates. However, this result contradicts the Police Crime Statistics of Germany (PCS). According to the latest PCS data, the proportion of non-German suspects is 30.4%, even if the foreign-specific offenses such as unauthorized entry are excluded (Bundeskriminalamt, 2017).

As no differences were found on the basis of nationality, this variable was not considered as an identifying feature of the groups in further analyses. However, the categorization based on the migration background resulted in significant differences in violent delinquency in the 6th and 10th grades. The results confirmed the findings of the Second International Self-Reported Delinquency Study (ISRD-2) for Germany (Enzmann, 2010). At this point, it should be noted that the effect sizes were very small (<.10) for violent delinquency. This means that there were significant but small differences between immigrant and non-immigrant boys in violent delinquency. This is consistent with the results of Boers et al. (2006).

Considering our mediation models, the fit with the data of the causal model was not good in the younger cohort, while the model fit was very good in the older one. The proportion of explained total variance in the dependent variable was also higher in the older cohort (36.1% vs. 14.2%). The mediator analyses showed that the explanation of violent delinquency was determined by the indirect effect of perceived discrimination and negative parenting via the mediator variables. Simons and colleagues (2003) examined the cross-sectional and longitudinal effects of perceived ethnic discrimination on delinquency in the USA using SEM. Their results, however, do not support our results. However, it should be considered that their model tested fewer variables, and their sample consisted of different migrant groups in different countries. However, Babka von Gostomski (2003b) found that discrimination experiences increased violence risk in his cross-sectional study in Germany.

In the tested model independent variables showed almost non-direct effects. However, the independent variables showed their effects through mediator variables. Anhut and Heitmeyer’s (2000) study confirmed that not every disintegration experience links with antisocial behaviour, depending on further influencing factors. Our results support their assumption that other mechanisms, such as social competence or active coping strategies of the person, play a determinant role in the effect of the indicators of three disintegration levels. In the mediation model (in the older cohort), the institutional level (perceived discrimination) and the social-emotional level (negative education) have no direct influence on the antisocial behaviour, but they have significant effects via the mediator variables.

In the present study, the strongest effect of negative education was on self-control in both cohorts. This result corroborated the assumptions by Gottfredson and Hirschi (1990) that parenting affects the self-control of children. Hay and Forrest (2006) found that parents influence the self-control of their children during puberty. Another important indirect effect of negative parenting is mediated through the acceptance of violence. Inconsistent education (Uslucan, 2009) and violence in the family (Baier et al., 2006) promote the acceptance of violence. Negative education also significantly influenced peer delinquency in both cohorts. Past studies have confirmed this result that parental violence (Baier et al., 2006; Pfeiffer, Wetzels, & Enzmann, 1999) and poor parental control (Baier, 2005) promote joining delinquent peers. While considering the indirect effect of discrimination, the strongest effect is mediated by the acceptance of violence on violent delinquency in both cohorts. Also Möller (2010) emphasized that ethnic discrimination can cause marginalized masculinity.

The mediator variables had some direct effects on the dependent variable, which has been supported in the literature. Several studies reported that acceptance of violence is related with higher rates of violence (Babka von Gostomski, 2003b; Baier et al., 2009). When considering the effect of self-control on delinquency, the General Theory of Crime of Gottfredson and Hirschi (1990) should be mentioned. This theory views the lack of self-control as the main source of crime, and this has been supported in several studies (e.g. McCullough & Willoughby, 2009; Schulz, Eifler, & Baier, 2011). Furthermore, in the older cohort, there was a highly significant direct effect of peer delinquency on violent delinquency. Many studies have also documented the effect of peer delinquency on violent delinquency in the literature (Lösel & Bliesener, 2003; Rabold, Baier, & Pfeiffer, 2008). This path, which was only significant in the older cohort, may indicate that the importance of peers is enhanced in older adolescence. The third mediator variable, peer delinquency, also had a highly significant effect on violent delinquency, which has also been confirmed many times in the literature (Baier & Pfeiffer, 2007; Lösel & Bliesener, 2003; Rabold et al., 2008).

Methodological Strengths and Weaknesses of the Present Study

The sample of the present study assessed two cohorts at two different locations. Thanks to the cohort sequence design, individual development could be observed over time, as well as a comparison of the younger and older cohorts. The combination of sociological and psychological aspects regarding antisocial behavior is another unique feature of this present study. Since previous analyses were almost exclusively performed using cross-sectional data, the present study contributed significantly to research of antisocial behavior in Germany.

One methodological limitation concerns the operationalisation of the theoretical model. Because of the variables measured in the study, the disintegration approach could not be operationalised to its full potential. For instance, consideration of the lack of fairness or equal opportunity (Endrikat, Schaefer, Mansel, & Heitmeyer, 2002), (dis)trust in law treatment (Babka von Gostomski, 2003b) at the institutional level or future prospects for the economic situation (Rippl & Baier, 2005) for the social-structural dimension had to be omitted from our analyses. There was a particular concern about not grasping the social-structural dimension appropriately, so that this dimension could not be represented in the statistical model after preliminary analysis. Nevertheless, important parts of the theory could be tested and confirmed.

Some further methodological limitations are concerned with sampling. First, our sampling procedure did not ensure representativeness with the German population. Therefore, the prevalence rates are to be interpreted with caution. Second, although we analysed only two measurement points, we had to deal with significant dropout, especially in the older cohort. This was due to a change in assessment modes: in the 10th and 11th grades, some of our subjects had already left school and therefore received the questionnaires via mail instead of during school lessons, which led to higher dropout rates in this study (Weiss & Link, 2019); however as mentioned before the dropout did not lead to a biased sample.

Conclusion

Our data support those studies that reported no differences in the delinquency rates of adolescents with or without migration background except for violent delinquency, however, the differences in violent delinquency were not high (ES < 0.10). The application of the disintegration model on migrant violence showed good model fit in late adolescents, whereas different factors seem relevant for younger adolescents.

Author Note

This manuscript was compiled based on a dissertation study which was published as a book thanks to the contributions of Prof. Dr. Mark Stemmler and Prof. Dr. Jost Reinecke (Uysal, 2017).

References

Anhut,

(2005). Die Konflikttheorie der Desintegrationstheorie [The conflict theory of disintegration theory]. In T. Bonacker (Ed.), Sozialwissenschaftliche Konflikttheorien [Conflict theories in the social sciences] (pp. 381–408). Wiesbaden: VS Verlag für Sozialwissenschaften. doi: 10.1007/978-3-322-94989-9_19

Anhut,

, & Heitmeyer,

(2000). Desintegration, Konflikt und Ethnisierung. Eine Problemanalyse und theoretische Rahmenkonzeption [Disintegration, conflict and ethnicization. A problem analysis and theoretical framework]. In W. Heitmeyer & R. Anhut (Eds.), Bedrohte Stadtgesellschaft: Soziale Desintegrationsprozesse und ethnisch-kulturelle Konfliktkonstellationen [Endangered urban societies: Social disintegration and ethnic-cultural conflict constellations] (pp. 17–75). Weinheim und München: Juventa. Retrieved from https://pub.uni-bielefeld.de/record/1861557

Anhut,

, & Heitmeyer,

(2009). Desintegration, Anerkennungsbilanzen und die Rolle sozialer Vergleichsprozesse für unterschiedliche Verarbeitungsmuster [Disintegration, recognition budgets and the role of social comparison processes for different processing patterns]. In G. Preyer (Ed.), Neuer Mensch und kollektive Identität in der Kommunikationsgesellschaft [New human and collective identity in the communication society] (pp. 212–236). Wiesbaden: VS Verlag für Sozialwissenschaften. doi: 10.1007/978-3-531-91471-8_12

Armsden,

G. C.

, & Greenberg,

M. T.

(1987). The inventory of parent and peer attachment: Individual differences and their relationship to psychological well-being in adolescence. Journal of Youth and Adolescence, 16, 427–454

10.1007/BF02202939

Babka von Gostomski,

, (2003a). Einflussfaktoren inter-und intraethnischen Gewalthandelns bei männlichen deutschen, türkischen und Aussiedler-Jugendlichen [Determinants of inter- and intra-ethnic violence among male German, Turkish and eastern resettler adolescents]. ZSE: Zeitschrift für Soziologie der Erziehung und Sozialisation, 23, 399–415.

Babka von Gostomski,

, (2003b). Gewalt als Reaktion auf Anerkennungsdefizite? Eine Analyse bei männlichen deutschen, türkischen und Aussiedler-Jugendlichen mit dem IKG-Jugendpanel 2001 [Violence as a reaction to lack of recognition? Male adolescents of Turkish descent, late migrants from Russia and Poland and Native Germans IKG Youth Panel 2001]. KZfSS Kölner Zeitschrift für Soziologie und Sozialpsychologie, 2, 253–277. doi: 10.1007/s11577-003-0050-4

Baier,

(2005). Abweichendes Verhalten im Jugendalter. Ein empirischer Vergleich verschiedener Erklärungsansätze [Deviant behavior in adolescence. An empirical comparison of different theoretical explanations]. ZSE: Zeitschrift für Soziologie der Erziehung und Sozialisation, 25, 381–398.

Baier,

, & Pfeiffer,

(2007). Gewalttätigkeit bei deutschen und nichtdeutschen Jugendlichen. Befunde der Schülerbefragung 2005 und Folgerungen für die Prävention [Violence among German and non-German youth. Findings from the 2005 student survey and implications for prevention] (KFN Forschungsberichte Nr. 100). Hannover: Kriminologisches Forschungsinstitut Niedersachsen. Retrieved from https://berlit.de/JKX_Baier_Pfeiffer_Gewalttaetigkeiten_2007_KFNForschungsbericht_100_2007.pdf

Baier,

, Pfeiffer,

, Simonson,

, & Rabold,

(2009). Jugendliche in Deutschland als Täter und Opfer von Gewalt [Adolescents as perpetrators and victims of violence] (Forschungsbericht Nr. 107). Hannover: Kriminologisches Forschungsinstitut Niedersachsen. Retrieved from https://kfn.de/wp-content/uploads/Forschungsberichte/FB_107.pdf

10.

Baier,

, Pfeiffer,

, & Windzio,

(2006). Jugendliche mit Migrationshintergrund als Opfer und Täter [Adolescents with a migration background as victims and perpetrators]. In W. Heitmeyer & M. Schröttle (Eds.), Gewalt. Beschreibungen, Analysen, Prävention [Violence. Descriptions, analyses, prevention] (pp. 240–268). Bonn: Bundeszentrale für Politische Bildung. doi: 10.1037//0022-3514.51.6.1173

11.

Baron,

R. M.

, & Kenny,

D. A.

(1986). The moderator– mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–10.1037//0022-3514.51.6.1173

12.

Boers,

, & Reinecke,

J. (Eds.).

(2007). Delinquenz im Jugendalter. Erkenntnisse einer Münsteraner Längsschnittstudie [Delinquency during adolescence. Findings from a longitudinal study in Münster]. Münster: Waxmann. Retrieved from https://pub.uni-bielefeld.de/record/1870168

13.

Boers,

, Walburg,

, & Reinecke,

(2006). Jugendkriminalität - Keine Zunahme im Dunkelfeld, kaum Unterschiede zwischen Einheimischen und Migranten. Befunde aus Duisburger und Münsteraner Längsschnittstudien [Juvenile delinquency - No increase in the dark field, little difference between natives and migrants. Findings from Duisburg and Münster longitudinal studies]. Monatsschrift für Kriminologie und Strafrechtsre-form, 89, 1–26.

14.

Bundeskriminalamt (2017). Polizeiliche Kriminalstatistik. Bundesrepublik Deutschland Jahrbuch 2017 [Police crime statistics. Federal Republic of Germany Yearbook 2017]. Wiesbaden: Bundeskriminalamt. Retrieved from https://www.bka.de/DE/AktuelleInformationen/StatistikenLagebilder/PolizeilicheKriminalstatistik/PKS2017/pks2017_node.html

15.

Dogan,

, & Strohmeier,

(2013). Perceived discrimination, multi-ethnic identity development, and acculturation orientations among Turkish immigrant youth living in Austria. 16th European Conference on Developmental Psychology (ECDP), Lausanne, Switzerland.

16.

Dünkel,

, & Geng,

B. (Eds.).

(2003). Jugendgewalt und Krimi- nalprävention. Empirische Befunde zu Gewalterfahrungen von Jugendlichen in Greifswald und Usedom/ Vorpommern und ihre Auswirkungen für die kommunale Kriminalprävention [Youth violence and crime prevention. Empirical findings on adolescents’ experiences of violence in Greifswald and Usedom/Vorpommern, and their impacts for municipal crime prevention]. Mönchengladbach: Forum-Verlag Godesberg.

17.

Eifler,

, & Seipel,

(2001). Eine deutsche Version einer Skala zur Erfassung von self-control. Bielefeld: Universität Biele-feld, Fakultät für Soziologie [A German version of a scale for assessing self-control]. Retrieved from https://pub.unibielefeld.de/record/1885540

18.

Endrikat,

, Schaefer,

, Mansel,

, & Heitmeyer,

(2002). Soziale Desintegration. Die riskanten Folgen negativer Anerkennungsbilanzen [Social disintegration. Lack of recognition and its risks]. In W. Heitmeyer (Ed.), Deutsche Zustände [German States]. Folge 1, (pp. 37–58). Frankfurt am Main: Suhrkamp. Retrieved from https://pub.uni-bielefeld.de/record/1861663

19.

Enzmann (2010). Germany. In J. Junger-Tas, I. H. Marshall, D. Enzmann, M. Killias, M. Steketee, & B. Gruszczyńska (Eds.), Juvenile delinquency in Europe and beyond: Results of the Second International Self-Report Delinquency Study (pp. 47–64). New York, NY: Springer.

20.

Essau,

C. A.

, Sasagawa,

, & Frick,

P. J.

(2006). Psychometric properties of the Alabama parenting questionnaire. Journal of Child and Family Studies, 15, 595–614

10.1007/s10826-006-9036-y

21.

Flora,

D. B.

, & Curran,

P. J.

(2004). An empirical evaluation of alternative methods of estimation for confirmatory factor analysis with ordinal data. Psychological Methods, 9, 466–491. doi: 10.1037/1082-989X.9.4.466

22.

Gonzalez,

E. J.

, & Kennedy,

A. M.

(2001). PIRLS 2001 user guide for the international database. Boston College, MA: International Association for the Evaluation of Educational Achievement (IEA). Retrieved from https://timssandpirls.bc.edu/pirls2001i/pdf/UserGuide.pdf

23.

Gottfredson,

M. R.

, & Hirschi,

(1990), Ageneral theory of crime. Stanford, CA: Stanford University Press.

24.

Grasmick,

H. G.

, Tittle,

C. R.

, Bursik,

R. J.

, & Arneklev,

B. J.

(1993). Testing the core empirical implications of Gottfredson and Hirschi’s general theory of crime. Journal of Research in Crime and Delinquency, 30, 5–29. doi: 10.1177/0022427893030001002

25.

Hay,

, & Forrest,

(2006). The development of self-control: Examining self-control theory’s stability thesis. Criminology, 44, 739–774. doi: 10.1111/j.1745-9125.2006.00062.x

26.

Hayes,

A. F.

(2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guilford Press.

27.

Heitmeyer,

, Kock,

, Marth,

, Schroth,

, Thome,

, & van de Wetering,

(2012). Situationen in den Stadtteilen im Vergleich. Ergebnisse quantitativer Beschreibungen [Comparison of situations in the districts. Results of quantitative descriptions]. In W. Heitmeyer, S. Kock, J. Marth, U.Thöle, H. Thome, A.,... Schroth (Eds.), Gewalt in öffentlichen Räumen [Violence in public spaces] (pp. 55–107). Wiesbaden: VS Verlag für Sozialwissenschaften. doi: 10.1007/978-3-531-18724-2_3

28.

Hoyt,

, & Scherer,

(1998). Female juvenile delinquency: Misunderstood by the juvenile justice system, neglected by social science. Law and Human Behavior, 22, 81–107. doi: 10.1023/A:1025728822468

29.

Kanning,

U. P.

(2009). Diagnostik sozialer Kompetenzen [Diagnostics of social competences]. Göttingen: Hogrefe.

30.

Lösel,

(1975). Handlungskontrolle und Jugenddelinquenz [Action control and juvenile delinquency]. Stuttgart: Enke.

31.

Lösel,

, Beelmann,

, Jaursch,

, Scherer,

, Stemmler,

, & Wallner,

(2003). Skalen zur Messung elterlichen Erziehungsverhaltens bei Vorschul- und Grundschulkindern. Die Deutschen Versionen der Parenting Sense of Competence Scale (PSOC) und der Parenting Scale (PARS) [Scales for assesing parenting behavior among preschool and primary school children. The German versions of the Parenting Sense of Competence Scale (PSOC) and the Parenting Scale (PARS)]. Universität Erlangen-Nürnberg: Institut für Psychologie.

32.

Lösel,

, & Bliesener,

(2003). Aggression und Delinquenz unter Jugendlichen. Untersuchungen von kognitiven und sozialen Bedingungen [Aggression and delinquency among adolescents. Studies on cognitive and social origins]. München, Neuwied: Luchterhand.

33.

MacKinnon,

D. P.

, Fairchild,

A. J.

, & Fritz,

M. S.

(2007). Mediation analysis. Annual Review in Psychology58, 593–614. doi: 10.1146/annurev.psych.58.110405.085542

34.

McCullough,

M. E.

, & Willoughby,

B. L. B.

(2009). Religion, self-regulation, and self-control: Associations, explanations, and implications. Psychological Bulletin, 135, 69–93. doi: 10.1037/a0014213

35.

Meinert,

, & Uysal,

(2015). Die Entstehung und Entwicklung devianten und delinquenten Verhaltens im Lebensverlauf und ihre Bedeutung für soziale Ungleichheitsprozesse: Methodendokumentation der Schülerbefragungen in Dortmund und Nürnberg. Dritte Erhebung 2014 SFB 882 Technical Report Series, 22 [The onset and development of deviant and delinquent behavior in the life-course and it‘s significance for social inequality processes: methodical documentation of pupil-surveys in Dortmund and Nuremberg. Third Survey 2014]. Bielefeld: DFG Research Center (SFB) 882 - From heterogeneities to inequalities. Retrieved from https://pub.uni-bielefeld.de/record/2782416

36.

Moffitt,

T. E.

(1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674–701. doi: 10.1037/0033-295X.100.4.674

37.

Möller,

(2010). Männlichkeit, Migration und Gewalt [Masculinity, migration and violence]. In H. Prömper, M. M. Jansen, A. Ruffling, & H. Nagel (Eds.), Was macht die Migration mit Männlichkeit. Kontexte und Erfahrungen zur Bildung und Sozialen Arbeit mit Migranten [What does migration do with masculinity. Contexts and experiences of education asnd social work with migrants] (pp. 51–72). Opladen und Farmington Hills: Barbara Budrich.

38.

Muthén, L. K., & Muthén, B. O. (1998–2012). Mplus user’s guide (7th ed.). Los Angeles: CA: Muthén & Muthén.

39.

Naplava,

(2018). Jugenddelinquenz im interethnischen Vergleich. In B. Dollinger, & H. Schmidt-Semisch [Juvenile delinquency in a multi-ethnic comparison] (Eds.), Handbuch Jugendkriminalität [Handbook of juvenile crime] (pp. 317–336). Wiesbaden: Springer VS. doi: 10.1007/978-3-531-19953-5_16

40.

Oberwittler,

(2007). The effects of neighbourhood poverty on adolescent problem behaviours: A multi-level analysis differentiated by gender and ethnicity. Housing Studies, 22, 781–803. 10.1080/02673030701474727

41.

Othold,

, & Schumann,

K. F.

(2003). Delinquenzverläufe nach Alter, Geschlecht und Nationalitätenstatus [Delinquent pathways by age, gender, and nationality status]. In Schumann

K. F.

(Eds.), Delinquenz im Lebensverlauf. Bremer Längsschnittstudie zum Übergang von der Schule in den Beruf bei ehemaligen Hauptschülern [Delinquency over the life course. Bremer longitudinal study on the transition from school to work with former lower-track students]. (Bd. 2, pp. 67–94). Weinheim und München: Juventa.

42.

Pfeiffer,

, Wetzels,

, & Enzmann,

(1999). Inner-familiäre Gewalt gegen Kinder und Jugendliche und ihre Auswirkungen [Intra-family violence against children and adolescents and its effects] (Forschungsberichte Nr. 80). Hannover: Kriminologisches Forschungsinstitut Niedersachsen. Retrieved from https://kfn.de/wp-content/uploads/Forschungsberichte/FB_80.pdf

43.

Rabold,

, & Baier,

(2011). Why are some ethnic groups more violent than others? The role of friendship network’s ethnic composition. Journal of Interpersonal Violence26, 3127–3156. doi: 10.1177/0886260510390944

44.

Rabold,

, Baier,

, & Pfeiffer,

(2008). Jugendgewalt und Jugenddelinquenz in Hannover. Aktuelle Befunde und Entwicklungen seit 1998 [Youth violence and juve-nile delinquency in Hannover. Recent findings and developments since 1998] (Forschungsbericht Nr. 105). Hannover: Kriminologisches Forschungsinstitut Nieder- sachsen. Retrieved from https://kfn.de/wp-content/uploads/Forschungsberichte/FB_105.pdf

45.

Reinecke,

(2014). Strukturgleichungsmodelle in den Sozialwissenschaften (2. Aufl.) [Structural equation models in the social sciences (2nd ed.)]. München: Oldenbourg Wissenschaftsverlag GmbH. Retrieved from https://pub.unibielefeld.de/record/2633406

46.

Reinecke,

, Stemmler,

, Arnis,

, El-Kayed,

, Meinert,

, Pöge,

, ... Wittenberg,

(2013). Entstehung und Entwicklung von Kinderund Jugenddelinquenz: erste Ergebnisse einer Längsschnittstudie [Onset and development of children and teenagers delinquency: First results of a longitudinal study]. Neue Kriminalpolitik, 25, 207–228. Retrieved from https://pub.uni-bielefeld.de/record/2633406

47.

Rippl,

, & Baier,

(2005). Das Deprivationskonzept in der Rechtsextremismusforschung [The concept of deprivation in the tradition of right-wing extremism research]. KZfSS Kölner Zeitschrift für Soziologie und Sozialpsychologie, 57, 644–666. doi: 10.1007/s11577-005-0219-0

48.

Rippl,

, Baier,

, Kindervater,

, & Boehnke,

(2005). DieEU-Osterweiterung als Mobilisierungsschub für ethnozentrische Einstellungen? Die Rolle von Bedrohungsgefühlen im Kontext situativer und dispositioneller Faktoren [Does the eastern Europe enlargement of the European Union mobilize ethnocentric attitudes? The importance of threats vis-à-vis situational and personality factors]. Zeitschrift für Soziologie, 34, 288–310. doi: 10.1515/zfsoz-2005-0403

49.

Rollett,

, Werneck,

, & Hanfstingl,

(2005). Das Längsschnittprojekt Familienentwicklung im Lebenslauf (FIL): Familienentwicklung beim Schulübertritt [The longitudinal Project of Family Development in the Life Course (FIL): Family Development at School Transition]. Forschungsbericht. Fakultät für Psychologie, Universität Wien: Institut für Entwicklungspsychologie und Psychologische Diagnostik. Retrieved from https://homepage.univie.ac.at/harald.werneck/FIL/Forschungsbericht t5.pdf

50.

Schepers,

, & Uysal,

(2014). Die Entstehung und Entwicklung devianten und delinquenten Verhaltens im Lebensverlauf und ihre Bedeutung für soziale Ungleichheitsprozesse: Methodendokumentation der Schülerbefragungen in Dortmund und Nürnberg. Zweite Erhebung 2013 [The onset and development of deviant and delinquent behavior over the life course and it‘s significance for social inequality processes: Methodical documentation of pupil-surveys in Dortmund and Nuremberg. Second survey 2013], SFB 882 Technical Report Series 05. Bielefeld: DFG Research Center (SFB) 882 - From Heterogeneities to Inequalities. Retrieved from https://pub.unibielefeld.de/download/2678841/2678842/SFB_882_Technical_Report_05_A2.pdf

51.

Schulz,

, Eifler,

, & Baier,

(2011). Wer Wind sät, wird Sturm ernten. Die Transmission von Gewalt im empirischen Theorienvergleich [They sow the wind and reap the whirlwind. An empirical theory comparison of explanatory mechanisms in the cycle of violence]. KZfSS Kölner Zeitschrift für Soziologie und Sozialpsychologie, 63, 111–145. doi: 10.1007/s11577-010-0127-9

52.

Simons,

R. L.

, Chen,

Y.-F.

, Stewart,

E. A.

, & Brody,

G. H.

(2003). Incidents of discrimination and risk for delinquency: A longitudinal test of strain theory with an African American sample. Justice Quarterly, 20, 827–854. doi: 10.1080/07418820300095711

53.

Skrobanek,

(2007). Wahrgenommene Diskriminierung und (Re) Ethnisierung bei Jugendlichen mit türkischem Migrationshintergrund und jungen Aussiedlern [Perceived discrimination and (Re)ethnicisation among young Turks and Resettlers]. ZSE: Zeitschrift für Soziologie der Erziehung und Sozialisation, 27, 265–284

54.

Strohmeier,

(2007). Soziale Beziehungen in multikulturellen Schulklassen. Wo liegen die Chancen, wo die Risiken? [Social relations in multicultural classes. Where are the opportunities, where are the risks?]Erziehung und Unterricht, 157, (9–10), 796–809.

55.

Uslucan,

H.-H.

(2009). Gewalterfahrungen, Erziehung im Elternhaus und Wohlbefinden bei deutschen und türkischen Jugendlichen [Domestic violence, parenting styles and well-being of German and Turkish juveniles]. Praxis der Kinderpsychologie und Kinderpsychiatrie, 5, 278–296. doi: 10.13109/prkk.2009.58.4.278

56.

Uysal,

(2017). Interkulturelle Spezifika bei antisozialem Verhalten männlicher Jugendlicher. Längsschnittliche Auswirkungen der Diskriminierung auf Gewalt- und Delinquenzverhalten [Intercultural specifics of antisocial behaviour of male adolescents. Longitudinal effects of discrimination on violence and delinquent behaviour]. Wiesbaden: Springer.

57.

Uysal,

, Link,

, & Weiss,

(2016). Migrationshintergrund und Jugendkriminalität [Migration background and youth delinquency]. In J. Reinecke, M. Stemmler & J. Wittenberg (Eds.), Devianz und Delinquenz im Kindes- und Jugendalter. Ungleichheitsdimensionen und Risikofaktoren [Deviance and delinquency in childhood and adolescence. Inequality and risk factors] (pp. 75–83).Wiesbaden: VS Verlag für Sozialwissenschaften. doi: 10.1007/978-3-658-08135-5_5

58.

Weiss,

, & Link,

(2019). Analysen zur Panelmortalität [Analyzes of panel mortality]. In S. Wallner, M. Weiss, J. Reinecke, & M. Stemmler (Eds.), Die Entwicklung von Devianz und Delinquenz in Kindheit und Jugend. Neue Ansätze der kriminologischen Forschung [The development of deviance and delinquency in childhood and adolescence. New approaches to criminological research] (pp. 63–82). Wiesbaden: Springer VS. doi: 10.1007/978-3-658-21234-6_6

59.

Wikström,

P.-O. H.

, Oberwittler,

, Treiber,

, & Hardie,

(2012). Breaking rules. The social and situational dynamics of young people’s urban crime. Oxford, UK: Oxford University Press. doi: 10.1057/sj.2013.5