An empirical test of self-control theory in Roma adolescents

Journal of Criminal Justice 44 (2016) 66–76

Contents lists available at ScienceDirect

Journal of Criminal Justice

An empirical test of self-control theory in Roma adolescents Alexander T. Vazsonyi a,⁎, Gabriela Ksinan Jiskrova a, Albert J. Ksinan a, Marek Blatný b a b

Department of Family Sciences, University of Kentucky, 160 Funkhouser Drive, Lexington, KY 40506, United States Institute of Psychology, Academy of Sciences of the Czech Republic, Veveří 97, Brno 602 00, Czech Republic

a r t i c l e

i n f o

Article history: Received 21 September 2015 Received in revised form 20 December 2015 Accepted 21 December 2015 Available online xxxx Keywords: Roma Romani Self-control Cross-cultural Deviance Parenting

a b s t r a c t Purpose: The current study tested the applicability of self-control theory in Roma adolescents, one of the largest ethnic minorities in Europe; it compared mean levels in deviance measures and correlates (parenting and low self-control) in Roma versus non-Roma Czech youth. Methods: Questionnaire data were collected from Roma (n = 239, 47.5% female, Mage = 14.02) and non-Roma (n = 130, 47.7% female, Mage = 14.71) adolescents residing in the Czech Republic. Measures included maternal parenting processes, low self-control, and deviance (alcohol use, school misconduct, and theft). Results: Findings from SEM analyses provided evidence that perceived maternal support predicted lower deviance both directly and indirectly via low self-control, while perceived maternal conflict predicted lower self-control and higher deviance. No differences were found between Roma and non-Roma adolescents in mean deviance scores or in the links between parenting, low self-control, and deviance. Conclusion: Roma adolescents did not differ from ethnic Czech adolescents in rates of deviance or the developmental processes focused on age-appropriate indicators of deviance. Thus, this evidence further extends the reach of Gottfredson and Hirschi's seminal theoretical work to ethnic minorities outside of North America. © 2015 Elsevier Ltd. All rights reserved.

The General Theory of Crime or Self-control Theory (Gottfredson & Hirschi, 1990) has generated a substantial amount of empirical research on how an individual's level of self-control interacts with environmental opportunities to foretell deviant behaviors in both adolescents and adults (Grasmick, Tittle, Bursik, & Arneklev, 1993; Hay & Forrest, 2006; Vaske, Ward, Boisvert, & Wright, 2012). Early socialization is a key in the development of self-control. Ineffective child rearing results in lower levels of self-control, which is linked to subsequent norm-violations. Empirical studies on these links have provided evidence of both direct and indirect effects of parenting on deviance, mediated through low self-control (Gibbs, Giever, & Higgins, 2003; Vazsonyi & Belliston, 2007); some work has also found limited support for indirect effects only (e.g., Hay, 2001). Based on longitudinal data, Vazsonyi and Huang (2010) concluded that variability in self-control among children at 4.5 years in age explained almost half of the variance of developmental changes in deviance over time. Although Gottfredson and Hirschi (1990) emphasized the importance of early socialization, they acknowledge in their original work that self-control has an individual difference or biological component (Vazsonyi, Roberts, & Huang, 2015). Indeed, recent behavior genetic evidence has revealed that between 55% and 66% of variability in selfcontrol is heritable (Beaver, Wright, DeLisi, & Vaughn, 2008; Wright & Beaver, 2005) and that the stability of self-control over time also has a

strong genetic basis (Beaver, Connolly, Schwartz, Al-Ghamdi, & Kobeisy, 2013; Coyne & Wright, 2014). This does not mean, however, that socialization processes, including effective parenting, are not meaningful in the development of self-control. Rather, biological factors predict deviance in an interaction with environmental factors (Moffitt, 2005; Vazsonyi & Huang, 2010; Vazsonyi et al., 2015), much like shown in experimental intervention work (Piquero, Farrington, Welsh, Tremblay, & Jennings, 2009; Piquero, Jennings, & Farrington, 2010). Individuals who do not develop adequate self-control are more prone to risky or deviant behaviors, including smoking, excessive drinking, risky sex behavior, drug abuse, or gambling (Pratt & Cullen, 2000), or in other words, acts that provide immediate and easy gratification. Numerous studies in criminology but also the developmental sciences have established self-control as a key individual difference that affects levels of adjustment and developmental outcomes throughout the life course (Casey et al., 2001; Mischel, Shoda, & Rodriguez, 1989; Moffitt et al., 2011; Vazsonyi & Huang, 2010). The analogous importance of impulse control or self-regulation in predicting adjustment outcomes has also been recognized in psychology (Baumeister, Heatherton, & Tice, 1994; Tangney, Baumeister, & Boone, 2004), education (Duckworth & Seligman, 2005), or health and health-risk research (Griffin, Scheier, Acevedo, Grenard, & Botvin, 2011; Miller, Barnes, & Beaver, 2011). The far reach of self-control theory

⁎ Corresponding author at: University of Kentucky, Department of Family Sciences, 316 Funkhouser Building, Lexington, KY 40506, United States. E-mail address: [email protected] (A.T. Vazsonyi).

http://dx.doi.org/10.1016/j.jcrimjus.2015.12.004 0047-2352/© 2015 Elsevier Ltd. All rights reserved.

The relationship between low self-control and various deviant behaviors has been consistently found in a number of studies utilizing

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

variety of methods from nationally representative longitudinal data analysis (Perrone, Sullivan, Pratt, & Margaryan, 2004; Vazsonyi & Huang, 2010) to meta-analysis (Pratt & Cullen, 2000) and in diverse adolescent samples. Previous research provided some evidence supporting the applicability of the theory across ethnic and cultural contexts. Low self-control was positively and consistently associated with delinquency in both European American and American Indian high school students (Morris, Wood, & Dunaway, 2007); additionally, virtually identical associations were found in a sample of Latino youth residing in the United States (Shekarkhar & Gibson, 2011), in Puerto Rican youth (Miller, Jennings, Alvarez-Rivera, & Lanza-Kaduce, 2009). In Macanese and Hong Kong adolescents, low self-control was associated with both violent and non-violent delinquency (Chan & Chui, 2015), and in Macanese adolescents with bullying perpetration (Chui & Chan, 2013). Additionally, the associations between low self-control and deviant behaviors among youth were invariant across a number of different cultures (Vazsonyi, Pickering, Junger, & Hessing, 2001), as was the link between low self-control and cyberbullying across 25 countries (Vazsonyi, Machackova, Sevcikova, Smahel, & Cerna, 2012). On the other hand, Cheung and Cheung (2008) found that low self-control was unrelated to delinquency in a sample of Hong Kong youth, above and beyond the effects by social factors. While the evidence of cross-cultural and cross-national generalizability of the link between self-control and delinquency is abundant, the role of parenting in predicting self-control and deviance in less clear. Cheung and Cheung (2008) found links between coercive parenting and both low self-control and delinquency. Similarly, Miller et al. (2009) found maternal attachment to be positively related to selfcontrol, and both self-control and maternal attachment to independently predict deviant behaviors. In contrast, Shekarkhar and Gibson (2011) did not find links between parenting and self-control. With regards to comparative studies, some found the links between parenting, low self-control, and deviance were invariant across cultural contexts (Vazsonyi & Belliston, 2007; Vazsonyi & Crosswhite, 2004; Vazsonyi, Wittekind, Belliston, & Loh, 2004), while others did not (Morris et al., 2007). Although the majority of authors found either direct links of parenting on deviance or indirect links via low self-control (e.g. Miller et al., 2009; Vazsonyi & Belliston, 2007; Vazsonyi et al., 2004), much work remains to be done in establishing whether these links vary across cultural environments or not, specifically outside of the United States. The current study builds on this previous work by testing the links between parenting, low self-control, and deviance in samples of Roma and non-Roma youth from the Czech Republic. To date, this study is among the few to test the applicability of self-control theory in minority versus majority populations outside of the United States. The developmental context of Roma youth in Central and Eastern Europe is particularly interesting for testing the theory as Roma adolescents develop in unique developmental contexts as compared to majority youth. One would expect that this might alter developmental processes, and thus, potentially limit the applicability of self-control theory to Roma youth. The following section further describes and contextualizes the current situation of Roma minority in Central and Eastern Europe, including challenges that Roma youth face. The developmental context of Roma adolescents in the Central and Eastern Europe The Roma people are a distinctive ethnic group that has had a long historical presence in Europe. They originally migrated from Central Northern India and entered Europe around the 14–15th century (Simko & Ginter, 2009). Since then, the use of the term “gypsy” to refer to any and all “nomadic and socially unstable” (Simko & Ginter, 2009, p. 16) groups of people has been used. The exact number of Roma living in Europe is unclear due to a lack of reliable, official data. This is a result of migration as well as a reluctance of Roma to report their ethnic origin (Simko & Ginter, 2009). However, it is estimated

67

that more than 5 million Roma live in the countries that once made up the Eastern bloc (e.g., Bulgaria, Czech Republic, Hungary, Romania, and Slovakia; Ginter, Krajcovicova-Kudlackova, Kacala, Kovacic, & Valachovicova, 2001). They are also the largest ethnic minority in Central and Eastern Europe. Roma have always had difficulties to fit into the majority culture, partly because European countries were not very welcoming towards these “dark-skinned people with completely different beliefs, odd traditions and incomprehensible language” (Historie a puvod Romu, n.d., para. 1). They have often been persecuted, and they have experienced chronic stigmatization (Kabakchieva et al., 2006). Thus, during the Second World War, about a half a million Roma were exterminated in Nazi concentration camps (Kabakchieva et al., 2006). During the second half of the 20th century, the Roma situation seemed to have been improving; however, Roma continue to experience discrimination in “schools, health service access, employment, housing access, and the judicial system” (de Cortazar, Leon, Garcia, & Nunez, 2009, p. 606). They suffer from a much higher level of poverty compared to the majority population, high levels of unemployment (Kabakchieva et al., 2006), and lower educational attainment (Kolarcik, Geckova, Orosova, van Dijk, & Reijneveld, 2010). After the expansion of European Union (EU) at the beginning of 21st century, the “Roma problem,” previously seen as rather a local issue, has quickly become an EU-wide concern, as the newly admitted members were mostly Central and Eastern European countries with significant Roma minority populations (Pogány, 2006). Several European countries, including the Czech Republic, have been criticized for their inadequate treatment of the Roma minority (e.g. segregation of Roma youth into “special schools”) by the European Union itself, the United Nations, as well as many non-governmental organizations (Parekh & Rose, 2011). Sigona and Trehan (2011) argue that in addition to racism and xenophobia across Europe (Eurobarometer survey, 2012), the Roma people have faced new challenges because of their inability to renegotiate a new social contract following the economic and social changes in Central and Eastern Europe, driven by the establishment of the European Union. Despite these concerns and challenges, research on Roma minority remains scarce. Most of the existing scholarship is predominantly descriptive with focus on adults (e.g. Ginter et al., 2001; Janevic, Jankovic, & Bradley, 2012). Only a few studies focused on Roma adolescents or comparisons of Roma and non-Roma youth (de Cortazar et al., 2009; Gerevich, Bacskai, Czobor, & Szabo, 2010; Kolarcik et al., 2010). Evidence from previous studies indicates that the developmental context of Roma children and youth in the Czech Republic is characterized by: (a) exclusionary practices in the educational process (The Ministry of Education, Youth and Sport of the Czech Republic, 2012; Vomastková, Souralová, & Nekorjak, 2011) and low levels of educational attainment in general (e.g. Ginter et al., 2001); (b) lower socioeconomic status and a higher likelihood of living in poverty than majority populations (Sirovátka & Mareš, 2006); (c) worse health outcomes and higher levels of health-compromising behaviors (Nesvadbová, Rutsch, Kroupa, & Sojka, 2000); and (d) fairly substantial negative views about Roma by members of majority populations (Eurobarometer survey, 2012) and pervasive stereotypes about Roma youth as prevalent perpetrators of deviant behaviors (Gatenio Gabel, 2009). The majority of ethnic Czechs share an opinion that Roma themselves are responsible for their current situation, and they see problems they face as caused by their ethnicity (being Roma, which they perceive as strikingly different from majority population regarding cultural values, customs, attitudes towards law, community and work; Dunbar & Simonova, 2003). The majority population reacts by excluding Roma from full participation in the majority culture and is willing to accept as members only those Roma who conform completely to majority values and expectations (Slavíčková & Zvagulis, 2014). The stigma associated with being members of this ethnicity is reflected in the fact that Roma are reluctant to identify with their

68

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

own ethnicity. In the Czech Republic, only 11,746 citizens officially self-identified themselves as Roma during the 2001 Census; however the actual number is estimated to be much higher, about 160,000–350,000 individuals (Stejskalova, 2013; see Fawn (2001) for an in-depth analysis of the situation of Roma minority in the Czech context). These challenges that Roma youth face suggest that problem behaviors in Czech youth might vary as function of being a member of the Roma versus non-Roma population. First, some scholarship found that low SES combined with minority status is a risk factor for engaging in criminal behaviors (Dunaway, Cullen, Burton, & Evans, 2000), and that family economic hardship affects parenting practices that in turn impact youths' adjustment including externalizing behaviors (Mistry, Vandewater, Huston, & McLoyd, 2002; Sampson & Laub, 1994). Thus, in Roma who report much lower levels of SES than majority populations (Ginter et al., 2001; Janevic et al., 2012), associations between parenting and deviance might differ from the ones found in higher SES, majority groups. Second, Roma endorse more traditional gender socialization of children than majority populations in Central and Eastern Europe, including in the Czech Republic (Kelly et al., 2004). Again, this might affect hypothesized relationships in Roma versus non-Roma females or Roma versus non-Roma males. Thus, in addition to comparisons of both ethnic groups, we also tested whether the relationships between parenting, low self-control, and deviance differed in Roma versus non-Roma girls and in Roma versus non-Roma boys. Third, members of the majority populations perceive Roma as specific in their lifestyle, behavior, and even psychological characteristics (“Romaniness”; Benedik, Göderle, & Tiefenbacher, 2013), and very different from themselves. This tendency termed essentialist view about the Roma ethnicity constitutes a basis for persistent differential treatment and discrimination that affects Roma youth unlike their non-Roma counterparts (Morton, Hornsey, & Postmes, 2009). In contrast, consistent with predictions by self-control theory (Gottfredson & Hirschi, 1990), previous cross-national and crosscultural comparative research has provided evidence that ethnic group membership or unique cultural context does not moderate the relationships between low self-control and deviance (e.g., Cheung & Cheung, 2008; Vazsonyi & Crosswhite, 2004; Vazsonyi et al., 2001; Vazsonyi et al., 2004; Vera & Moon, 2013) or the relationships between parenting, low self-control, and deviance (Vazsonyi & Belliston, 2007). Thus, competing hypotheses for the relationships between parenting, self-control, and deviance were developed in samples of Roma and non-Roma adolescents. On the one hand, consistent with selfcontrol theory and previous empirical work, we hypothesized that few differences would be found in the examined links across ethnic groups. On the other hand, and consistent with the unique characteristics of the developmental context, we expected that these links would vary as function of being a member of Roma versus non-Roma population.

deviance, and positive direct effects of maternal conflict and low selfcontrol on deviance; and (b) indirect effects of all parenting measures on deviance via low self-control (Fig. 1). We examined the effect of core parenting dimensions, outlined by Steinberg and Silk (2002) as harmony, autonomy, and parent– child conflict (operationalized as support, monitoring, and conflict), and the level of self-control on deviance among Roma and non-Roma adolescents in the Czech Republic. Three measures of deviance were examined that seemed the most developmentally appropriate with a range of severity, namely alcohol use, school misconduct, and theft. Given the scarcity of studies on Roma youth, the current study is one of the first that examines family processes and rates of deviance among Roma adolescents with direct comparison to a sample from a majority population, in our case non-Roma Czech adolescents. Participants Anonymous, self-report data were collected from Roma (n = 239, 47.5% female, mean age = 14.02) and non-Roma (n = 130, 47.7% female, mean age = 14.71) adolescents across three urban areas in the Czech Republic. Participants were recruited from the 2nd stage elementary schools1 and completed surveys during their regular school hours by either paper and pencil or online.2 The study was approved by a university IRB as well as an ethics committee in the Czech Republic. Data were collected in 2010 and 2011. Forty-nine schools were contacted with a request for cooperation in data collection. Of these schools, 9 were successfully recruited and took part in the survey. Data were collected in 19 classrooms, supervised by teachers, following a standardized written data collection protocol (89.3% participation rate). Of the 475 returned surveys, 48 were unusable, due to incomplete or missing data, and thus dropped. The hypothesized models were compared by ethnicity and sex; thus, participants who did not identify their sex or ethnicity (46 participants) were excluded. Lastly, because of a focus on early to middle adolescents attending elementary schools, in the context of the Czech Republic, this included youth who were 16 years old or younger. Thus, adolescents older than 16 years in age were removed from the sample. This procedure led to final sample size of 369 participants. Descriptive statistics of the sample (sex, age, family structure, parent's education level, and family primary wage earner's type of work) are included in Table 1, separately for Roma and non-Roma adolescents. To evaluate possible impact of sample reduction on results, we performed all analyses twice, using both full and final analytical sample. The results based on both samples were not different in terms of statistical significance of the associations, adequacy of model fit, or conclusion about fit invariance between groups. Measures Parenting processes

Research rationale and study hypotheses The aim of this study was to investigate both mean levels as well as predictors of deviance in adolescents residing in the Czech Republic and to test whether these varied between Roma versus non-Roma youth. Thus, the study juxtaposes the theoretical prediction of similarities across different ethnic, racial, and cultural groups with the one of idiosyncratic experiences in unique developmental contexts. Consistent with self-control theory (Gottfredson & Hirschi, 1990), previous research has shown that both parenting processes as well as levels of self-control are key in understanding variability in deviance (e.g., Vazsonyi & Belliston, 2007). Additionally, previous scholarship found evidence that self-control may serve as a mediator between parenting processes and externalizing behavior (e.g., Gibbs et al., 2003; Vazsonyi & Belliston, 2007). Specifically, we hypothesized: (a) negative direct effects of maternal monitoring and support on

Parenting processes were measured by three subscales from the Adolescent Family Process measure (AFP; Vazsonyi, Hibbert, & Snider, 2003). The current study focused on maternal parenting as mothers tend to be the primary caregiver in most families. In addition to Cronbach's alpha, Omega reliability coefficients are also reported as the reliance on alpha as the sole estimator of reliability has been recently criticized (Crutzen, 2014; Dunn, Baguley, & Brunsden, 2014; Peters, 2014; Sijtsma, 2009). Omega coefficients are based on a more realistic assumption than Cronbach's alpha, namely there is no assumption that all items have equal item covariances as well as equal associations with the latent construct; thus, they yield a more accurate estimate of reliability according to recent evidence (Dunn et al., 2014). Scale scores for maternal support (4-item scale, α = .71, ω = .72 in Roma sample, α = .79, ω = .79 in non-Roma sample), maternal monitoring (4-item scale, α = .68, ω = .70 in Roma sample, α = .74, ω =

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

69

Fig. 1. The hypothesized relationships between parenting, self-control, and deviance.

.76 in non-Roma sample), and maternal conflict (3-item scale, α = .77, ω = .77 in Roma sample, α = .83, ω = .83 in non-Roma sample) were computed by averaging the items part of subscales. All items were rated on 5-point Likert-type scale ranging from strongly disagree (1) to strongly agree (5). For specific items part of the measure please see Appendix A.

were instructed to express either disagreement or agreement with items pertaining to their level of self-regulation and risk-seeking (e.g. “I am more concerned with what happens to me in the short run than in the long run.”).

Self-control

Lifetime deviance was measured by three subscales from the Normative Deviance Scale (NDS; Vazsonyi et al., 2001) that assess lifetime frequency of engaging in various deviant behaviors. We focused on lifetime prevalence because we were interested in predicting deviance rather than in presenting prevalence rates of these behaviors. Additionally, the reported number of incidents in early to middle adolescent sample is expectedly low, thus, a lifetime prevalence measure is best suited to capture deviant behaviors (Vazsonyi et al., 2001). Lastly, substantial empirical evidence has shown that lifetime versus recency deviance measures seem to be highly associated and thus similar, both in terms of reliability and validity (e.g. Elliott & Huizinga, 1988, Gottfredson & Hirschi, 1996, Thornberry & Krohn, 2000). In the current study, and related to the age of the sample, three subscales were used, namely alcohol use (7-item scale, α = .88, ω = .88 in Roma sample, α = .88, ω = .89 in non-Roma sample), school misconduct (6-item scale, α = .67, ω = .69 in Roma sample, α = .70, ω3 = .57 in non-Roma sample), and theft (7-item scale, α = .76, ω = .82 in Roma sample, α = .87, ω = .86 in non-Roma sample). Scale scores were computed by averaging subscale items. Items were rated on a 5-point scale ranging from never (1) to 6 or more times (5). Adolescents were instructed to report the frequency of engaging in various problem behaviors (e.g. “Have you ever got drunk intentionally just for the fun of it?”).

Low self-control was assessed by 24-item Low Self-control Scale (LSC; Grasmick et al., 1993; α = .83, ω = .82 in Roma sample, α = .82, ω = .80 in non-Roma sample) by averaging the items constituting this scale. All items were rated on a 5-point Likert-type scale rather than a 4-point scale as originally developed (see Vazsonyi et al., 2001), ranging from strongly disagree (1) to strongly agree (5). The adolescents

Table 1 Descriptive statistics of demographic variables by sample. Roma Mean age (SD) Sex Males Females Family structure Two biological parents Other Primary wage earner Laborer Semiskilled Clerical Semiprofessional Professional Executive Father's education Elementary school High school Some college Undergraduate degree Graduate degree Mother's education Elementary school High school Some college Undergraduate degree Graduate degree

Non-Roma

14.02

1.04

14.71

0.87

124 112

52.5% 47.5%

68 62

52.3% 47.7%

169 69

71.0% 29.0%

104 26

80.0% 20.0%

47 90 51 22 9 0

21.5% 41.1% 23.3% 10.0% 4.1% 0%

7 16 45 20 33 6

5.5% 12.6% 35.4% 15.7% 26.0% 4.7%

23 164 2 1 7

11.7% 83.2% 1.0% 0.5% 3.6%

3 88 4 4 25

2.4% 71.0% 3.2% 3.2% 20.2%

39 158 3 2 9

18.5% 74.9% 1.4% 0.9% 4.3%

7 85 10 5 19

5.6% 67.5% 7.9% 4.0% 15.1%

Note. All percentages are valid percents.

Deviance

Control variables In the main analyses, we controlled for background variables, namely age, sex, family structure, and socioeconomic status (SES). Family structure was recoded into a dichotomous variable indicating whether adolescent was from a two-parent family or other type of family. SES was computed from combining standardized measures of maternal and paternal education and primary wage earner's type of work, an ordered categorical variable answered on 6-point scale ranging from laborer or service worker (1) to position requiring advanced degree, government official, executive professional, or similar (6).

70

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

Analytic procedure The main data analyses included four steps. First, correlations were computed for the entire sample to examine strength and direction of the associations between variables of interest. Second, mean-level differences on key study constructs between Roma and non-Roma were compared. More specifically, a series of eight t-tests were computed to examine similarities and differences on measures of SES, parenting processes (monitoring, support, and conflict), low self-control, and measures of deviance (alcohol use, theft, and school misconduct). A Bonferroni correction was implemented to adjust for inflated Type I error rates due to multiple comparisons. Additionally, similarities or differences in family structure between Roma and non-Roma adolescents were tested by χ2 statistics. Third, and in preparation for the main analyses, a series of Confirmatory Factor Analyses (CFA) were conducted on each of the three subscales part of the Adolescent Family Process measure (AFP; Vazsonyi et al., 2003) as well as the Low Self-control Scale (LSC; Grasmick et al., 1993). This was a necessary first step to examine the extent to which measurement of key constructs replicated across the two samples. Fourth, Structural Equation Modeling (SEM) was used to test both direct effects of parenting on deviance and indirect effects of parenting on deviance via low self-control. Inferential statistics were computed in SPSS 21; SEM analyses were completed in Amos 21 (Arbuckle, 2012). To test the hypothesized SEM model, latent constructs for the variables of interest (support, monitoring, conflict, low self-control, and deviance) were developed with the use of two item parcels as indicators. Parcels were developed by evaluating relative item loadings within each scale, based on an exploratory factor analysis, and then assigning scale items to parcels, where the highest loading item was assigned to the first parcel, the second highest to the second, the third to the first, and so forth (Little, Cunningham, Shahar, & Widaman, 2002). For the deviance latent construct, three subscales of the original NDS, namely alcohol use, school misconduct, and theft were specified as indicators. An initial model test based on the total sample examined model fit, direct effects of parenting and self-control on deviance, as well as indirect effects of parenting on deviance via low self-control. Control variables (age, sex, family structure, SES, and ethnicity) were not entered into the model; instead, residualized scores of indicators were used to partial out their effects. This approach is preferred over entering control variables into the models due to increased model complexity and resulting problems with convergence. To handle missing data, the full information maximum likelihood (FIML) feature in AMOS was implemented (9.3% of missing data). To estimate the statistical significance of indirect effects, however, multiple imputation was used instead of FIML because the FIML procedure does not allow for bootstrapping. Bootstrapping is needed to correctly estimate the significance of indirect effects. Thus, in SPSS, we imputed 10 data sets; next, in AMOS, a bootstrapping technique with 5000 re-samples was completed, separately for each imputed data set. Results (standardized estimates and errors) were pooled from the 10 iterations based on the Rubin/Schafer formula (Rubin, 1987; Schafer, 1997). Lastly, a series of multi-group analyses were completed to test for invariance by ethnic group (a) in the full sample (Roma versus nonRoma) and (b) separately for female and male youth (Roma versusnon-Roma). This was accomplished by imposing constraints on the structural paths between the latent constructs and evaluating changes in χ2, CFI, and RMSEA between the freely estimated baseline model and a constrained model. The χ2 difference test is the most common index for evaluating change in model fit; a statistically non-significant Δχ2 between baseline and constrained model is interpreted as no difference between models, in other words, as a support for invariance (e.g., Kline, 2010). However, there is an evidence that χ2 difference test may not perform well in large samples and may yield statistically significant results even though difference between models is trivial (Cheung & Rensvold, 2002). Thus, Cheung and Rensvold (2002) suggest

evaluating differences in alternative fit indices, including the CFI (ΔCFI less than .01 for evidence of invariance). Additionally, Chen (2007) suggests that a change in RMSEA larger than .010 should be considered as evidence of non-invariance. Preliminary Analyses Pearson's bivariate correlations Table 2 includes bivariate correlations among the main study variables for each ethnic group. Results indicated that maternal support was significantly and negatively correlated with low self-control as well as with deviance measures in both samples. Additionally, maternal monitoring was found to be significantly and negatively related to alcohol use in both samples, and low self-control, school misconduct, as well as theft in the Roma sample only. Furthermore, maternal conflict was significantly and positively related to low self-control as well as all deviance subscales in both samples. Additionally, low self-control was positively related to all deviance subscales in both samples; lastly, as expected, all deviance subscales were significantly associated. Mean level differences between Roma and non-Roma Based on t-tests and χ2 results, a significant difference was found between Roma (M = 3.90, SD = 0.98) and non-Roma (M = 3.57, SD = 1.00) youth in maternal monitoring, t(356) = −3.10, p b .01. Furthermore, Roma (M = − 0.35, SD = 0.74) and non-Roma (M = 0.55, SD = 0.92) adolescents significantly differed in SES, t(365) = 10.21, p b .001 (Table 3). These differences remained significant even after a Bonferroni correction for inflated Type I error. Confirmatory factor analysis: measurement A CFA on the three AFP subscales was tested as a model with three correlated latent parenting constructs (i.e., support, monitoring, and conflict), measured by the items part of each subscale, yielded very good fit based on the total sample (χ2 = 156.287 df = 82, χ2/df = 1.906, p b .001, CFI = .930, RMSEA = .050 [90% CI = .038, .061]). Additionally, we found support for metric invariance of the measure in the Roma and non-Roma samples. A test of the free versus fixed model provided evidence of invariance across the two groups (Δχ2 = 13.039, Δdf = 14, p = .523, ΔCFI = .001, ΔRMSEA = .004). Similarly, a CFA conducted on the LSC scale indicated by the six selfcontrol subscales (risk-seeking, impulsivity, self-centeredness, physical activity, preference for easy tasks, and temper) showed a good fit based on the total sample (χ2 = 38.169 df = 18 χ2/df = 2.120, p = .004, CFI = .936, RMSEA = .055 [90% CI = .030, .080]). Additionally, metric invariance for the measure was found across the two ethnic groups (Δχ2 = 11.490, Δdf = 6, p = .074 ΔCFI = .017 ΔRMSEA = .001). Structural equation modeling: model tests Full model To test model fit of the hypothesized relationships as well as to obtain direct and indirect effects estimates, the hypothesized model was first tested in the full sample which included both Roma and nonRoma adolescents; all control variables were included as described (i.e. sex, ethnicity, age, SES, and family structure). The model showed excellent fit to the data: χ2 = 54.000, df = 34, χ2/df = 1.588, p = .016, CFI = .983, RMSEA = .040 [90% CI = .018, .059]. Results revealed statistically significant negative direct effect of maternal support (β = −.43, p b .001) on low self-control. Additionally, significant positive direct effects of maternal conflict (β = .30, p b .01) and of low self-control (β = .36, p b .001) on deviance were found. Maternal conflict was significantly negatively correlated with maternal support

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

71

Table 2 Pearson's bivariate correlations between main study variables. α

ω

1

0.79 0.71 0.74 0.68 0.83 0.77 0.82 0.83 0.88 0.88 0.70 0.67 0.87 0.76

0.79 0.72 0.76 0.70 0.83 0.77 0.80 0.82 0.88 0.89 0.57 0.69 0.86 0.82

0.18 0.14 −0.65⁎⁎⁎ −0.46⁎⁎⁎ −0.42⁎⁎⁎ −0.35⁎⁎⁎ −0.33⁎⁎⁎ −0.26⁎⁎ −0.30⁎⁎ −0.21⁎⁎ −0.26⁎⁎ −0.19⁎

Variable 1. Support 2. Monitoring 3. Conflict 4. Low self-control 5. Alcohol use 6. School misconduct 7. Theft

2

−0.07 −0.10 0.15 −0.22⁎⁎ −0.30⁎⁎ −0.20⁎⁎ −0.09 −0.19⁎⁎ −0.18 −0.20⁎⁎

3

4

5

6

0.37⁎⁎⁎ 0.25⁎⁎ 0.25⁎⁎ 0.33⁎⁎⁎ 0.25⁎⁎ 0.29⁎⁎⁎ 0.23⁎ 0.26⁎⁎⁎

0.40⁎⁎⁎ 0.31⁎⁎⁎ 0.45⁎⁎⁎ 0.34⁎⁎⁎ 0.24⁎⁎ 0.18⁎

0.54⁎⁎⁎ 0.68⁎⁎⁎ 0.50⁎⁎⁎ 0.55⁎⁎⁎

0.61⁎⁎⁎ 0.50⁎⁎⁎

Note. First row shows results non-Roma, second row shows results for Roma. ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001.

(r = −.67, p b .001) and monitoring (r = −.20, p b .05), and maternal support was positively correlated with maternal monitoring (r = .18, p b .05). The tested model explained 24% of variance in low selfcontrol and 32% of variance in deviance, net any effects by background variables (see Fig. 2 for summary of direct effect estimates). Lastly, a significant indirect effect of maternal support on deviance was found, via low self-control (β = −.12, SE = .05, p b .05). However, no significant indirect effects of maternal monitoring (β = −.01, SE = .02, p = .578) or maternal conflict (β = .03, SE = .04, p = .431) on deviance were observed. Multigroup model comparisons Model invariance tests were conducted using multigroup tests. First, invariance between ethnicities was examined. For this, a baseline, unconstrained model, controlling for age, sex, SES, and family structure, but not for ethnicity, was tested. Next, a constrained model was specified where all paths between latent constructs were constrained to equality between the Roma and non-Roma subsamples; change in the model fit between this constrained model and the free model was compared. The baseline, unconstrained model showed excellent fit to the data (χ2 = 100.288, df = 68, χ2/df = 1.475, p = .007, CFI = .974, RMSEA = .036 [90% CI = .019, .050]). After constraining structural paths to equality the model fit remained excellent (χ2 = 112.057, df = 75, χ2/df = 1.494, p = .004, CFI = .970, RMSEA = .037 [90% CI = .021, .050]), and importantly, inspection of difference statistics indicated that fit did not significantly worsen compared to the baseline

model (Δχ2 = 11.768, Δdf = 7, p = .108 ΔCFI = .004 ΔRMSEA = .001). Thus, no significant differences were found and the model was deemed invariant in Roma versus non-Roma youth. Second, to be conservative and to address potential gender socialization differences, we also tested for the invariance between male and female youth across ethnic groups (i.e., Roma versus non-Roma girls, as well as Roma versus non-Roma boys). The same analytic procedures were repeated as previously described. For female youth, the baseline model with controls provided good fit: χ2 = 94.357, df = 68 χ2/df = 1.388, p = .019, CFI = .959, RMSEA = .047 [90% CI = .020, .069]. Next, all structural paths were constrained to equality to assess whether there existed differences between Roma and non-Roma female adolescents. The following fit indices were found for this model: χ2 = 104.654, df = 75 χ2/df = 1.395, p = .013, CFI = .953, RMSEA = .048 [90% CI = .023, .069]. Again, the difference between the unconstrained and constrained models was small and non-significant, Δχ2 = 10.297, Δdf = 7, p = .172 ΔCFI = .006 ΔRMSEA = .001, thus providing evidence of invariance between Roma and non-Roma girls. For male adolescents, the free model also yielded excellent fit: χ 2 = 85.908, df = 68 χ2 /df = 1.263, p = .070, CFI = .974, RMSEA = .037 [90% CI = .000, .060]. This was also true of the constrained model: χ2 = 97.913, df = 75 χ2/df = 1.307, p = .039, CFI = .966, RMSEA = .040 [90% CI = .010, .061]. Again, the difference between the unconstrained and constrained models was small and non-significant, Δχ 2 = 12.005, Δdf = 7, p = .100 ΔCFI = .008 ΔRMSEA = .003, thus providing evidence that model fit was invariant for Roma versus non-Roma male adolescents.

Table 3 Mean level differences in independent, dependent, and control variables by ethnic group. Variable

Support Monitoring Conflict LSC Alcohol use School misconduct Theft SES

Family structure

Roma

Non-Roma

M

SD

n

min–max

M

SD

n

min–max

3.60 3.90 2.38 2.97 2.03 1.97 1.25 -0.35

1.03 0.98 0.97 0.60 1.02 0.75 0.43 0.74

231 231 232 192 232 232 228 237

1–5 1–5 1–5 1–4.32 1–5 1–4.43 1–4.5 -2.06–3.05

3.61 3.57 2.45 2.88 2.09 1.84 1.21 0.55

1.04 1.00 0.97 0.58 1.01 0.70 0.42 0.92

128 127 129 125 128 126 127 130

1–5 1–5 1–5 1–4.13 1–5 1–4.14 1–3.43 -1.40–2.77

n two-parent

n other

n two-parent

n other

169 (71%)

69 (29%)

104 (80%)

26 (20%)

238

1–2

Note. SES = socioeconomic status computed from parental education and primary wage earner work type (in z-scores).

130

1–2

t

p

0.06 -3.10 0.67 -1.31 0.51 -1.62 -0.78 10.21

.955 .002 .504 .192 .609 .107 .436 b.001

χ2

p

3.55

.60

72

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

Fig. 2. Results from SEM model test: standardized effects (total sample).

Discussion The current study tested mean level differences as well as the associations between parenting, low self-control and several indicators of deviant behaviors in samples of Roma and non-Roma adolescents residing in the Czech Republic. By doing so, it addressed several gaps in the literature. First, it examined prevalence and etiology of problem behaviors in Roma adolescents, an understudied ethnic minority from Europe. Second, it is one of the first studies that tested validity of self-control theory in ethnic minority sample outside the United States, and third, it tested whether theoretical predictions of self-control theory held in both Roma and nonRoma Czech adolescents, or alternatively, whether Roma differed from non-Roma in the hypothesized relationships. In regards to mean level differences in the variables of interest, only a few statistically significant differences were found between Roma and non-Roma adolescents. Roma reported a lower level of SES and a higher level of maternal monitoring as compared to non-Roma youth. Both of these findings are in fact consistent with previous work. Low SES remains a persistent challenge for members of Roma communities (Janevic et al., 2012; Pogány, 2006; Simko & Ginter, 2009). Additionally, some studies have shown that higher levels of parental monitoring in disadvantaged neighborhoods seems to exist in order to compensate for the risk of residing in such an environment (Rankin & Quane, 2002). Roma and Roma youth today are highly stigmatized across Europe, often times implicated as being responsible for norm violations (Eurobarometer survey, 2012; Gatenio Gabel, 2009). In contrast to these prevailing views, Roma adolescents did not report higher levels of deviant behaviors than non-Roma youth. Consistent with the current evidence, Kolarcik et al. (2010)found that Roma youth from Slovakia between the ages of 12 and 17 reported lower rates of smoking, alcohol use, and drug use than their non-Roma peers. Additionally, we examined direct effects of parenting dimensions on deviant behaviors, and on low self-control, and indirect effects of parenting in deviance via low self-control. These analyses sought to clarify whether the etiology of deviant behaviors is similar between Roma and non-Roma adolescents, as predicted by self-control theory, or whether it is different due to the unique and idiosyncratic developmental contexts in which these adolescents reside. Findings from the current

study are generally consistent with previous work on the effects of parental behaviors on measures of deviance and externalizing behaviors (Gibbs et al., 2003; Hofer et al., 2013; McCarty, Zimmerman, Digiuseppe & Christakis, 2005; Metzler, Biglan, Ary, & Li, 1998; Vazsonyi & Belliston, 2007; Yeh, 2011). They provide empirical support for the co-existence of direct effects of parenting as well as indirect effects via low self-control. Maternal support indirectly affected deviance, while maternal conflict did so only directly. No significant association was found for maternal monitoring. Additionally, low self-control was strongly associated with deviance. The full model provided excellent fit to the data and explained 32% of the variance in deviance, net any effects by background variables. The link between low self-control and deviant behaviors has been well established (Perrone et al., 2004; Pratt & Cullen, 2000; Pratt, Turner, & Piquero, 2004), thus the fairly strong association between low self-control and deviance only further supported the hypothesized relationship. Additionally, the effect of maternal support on deviance has been previously found with some studies reporting direct effect on deviance (Barnes, Hoffman, Welte, Farrell, & Dintcheff, 2006; McCarty, Zimmerman, Digiuseppe and Christakis, 2005), while others both direct and indirect effect via low self-control (Vazsonyi & Belliston, 2007). Non-significant associations between maternal monitoring, low self-control, and problem behaviors were unexpected as most of previous work did find an effect of monitoring on low selfcontrol and problem behaviors (e.g. Stevens, Vollebergh, Pels, & Crijnen, 2005). Our results are more in line with the work of Coohey, Renner, and Sabri (2013), for instance, who did not find relationship between monitoring and problem behaviors. These results might also be related to the specific age of the youth studied, when, consistent with theory, monitoring becomes less salient developmentally. Finally, although only indirectly related to theory as an indicator of poor parent–adolescent relationships, parental conflict has been found to be important in predicting adolescent problem behaviors (Burt, McGue, Iacono, & Krueger, 2006; Crean, 2008). Findings supported the direct link between conflict and deviance; however, no indirect link between these variables via low self-control was found. In sum, based on the model fit, variance explained in outcome variables, and strength of the links between the variables of interest,

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

consistent support for the hypothesized model was found, with the exception of the non-significant associations between maternal monitoring, low self-control, and deviance. Model invariance testing showed no differences in links between tested variables between the two ethnic groups. In other words, there were no differences in the relations of parenting behaviors and problem behaviors between Roma and non-Roma youth when background variables were held constant. This suggests that the way parenting processes influence adolescent deviant behaviors is generally the same for both ethnic groups. This conclusion holds for both girls and boys — our comparisons showed no differences between Roma girls and non-Roma girls as well as Roma boys and non-Roma boys. These findings parallel ones from previous work conducted across cultural groups (Vazsonyi et al., 2001) and different ethnic groups in the United States (Miller et al., 2009; Vazsonyi & Crosswhite, 2004). The no difference findings between Roma and non-Roma Czech adolescents, in both mean levels of deviance and in the models predicting deviance is of critical importance. First, it provides additional evidence of generalizability of self-control theory to adolescent populations outside the United States and diverse minority groups. Second, it supports the argument that basic developmental processes as influence of parenting and self-control on problem behaviors operate similarly in youth of diverse cultural and socioeconomic background. We found socioeconomic status (SES) to correlate only weakly with family structure; adolescents who did not live with both biological parents reported lower SES (r = −.18, p = .001). Additionally, school misconduct that was negatively associated with SES (r = −.11, p = .04). No other statistically significant relationships between SES and the remaining variables were found; thus, it does not seem that lower SES serves as a risk factor for deviance in Roma adolescents. Third, the results of this study speak strongly against essentialist views of Roma that hold members of majority populations. Perceptions of Roma adolescents as fundamentally different than their non-Roma peers justify differential and discriminatory treatment of Roma youth. However, based on our results, Roma youth seem quite similar to their non-Roma counterparts in both rates and etiology of deviance. Thus, the data do not support the stereotype that Roma are prevalent perpetrators of deviant behaviors. Roma adolescents in the current study sample attended standard, ethnically mixed schools. This sample characteristic may be interpreted both as an advantage and as a limitation. Research on Roma minority has traditionally focused more on the most at-risk samples, such as Roma living in socially excluded communities or Roma settlements (e.g. Gerevich et al., 2010). According to Prokop (2015, April 18) about one third of Czech Roma currently live in socially excluded areas. Additionally, a similar number of Roma children (approximately 20%) attend “special schools” as they are called in the Czech Republic (schools designed for children with intellectual disabilities who are not able to attend regular elementary schools; The Ministry of Education, Youth and Sport of the Czech Republic, 2012), although they would be able to complete standard education with appropriate supports. These are not negligible numbers, however, they also indicate that majority of Roma youth come from fairly “integrated” environments and attend regular elementary schools. This “invisible majority” is not commonly represented in research and furthermore, does not correspond to the prototypical portrait of Roma as problematic and deviant. Thus, in this sense, the current study draws a probably more accurate picture of Roma adolescents' developmental characteristics in comparison to work that focuses solely on the most at-risk samples. Limitations The characteristics of our sample discussed constitute a limitation as well. The majority of Roma do not live in socially excluded communities; however, they are still overrepresented among the poorest

73

and most disadvantaged youth in the Czech Republic (Nesvadbová et al., 2000; Pogány, 2006). Roma children and adolescents coming from these families are very likely systematically different from Roma students in our sample with regards to SES, educational aspirations, access to education, and neighborhood quality. Thus, the results of present study may not be readily applicable to these groups. To overcome this limitation, data from Roma youth attending not only standard elementary schools but also students enrolled in special schools should be collected, to enable direct comparisons to Roma attending regular schools. This would be one of many important next steps in this line of research. Second and an equally important limitation was the crosssectional study design. The hypothesized links in the tested model are supported by results of longitudinal studies (e.g., Vazsonyi & Huang, 2010), suggesting that parenting processes precede the development of self-control and deviant behaviors; however, the proposed model in the current investigation only captures concurrent associations between parenting and deviance and thus is unable to address causality or directionality of the effects. Clearly, a longitudinal study with several time points would be able to partially overcome these limitations. Future directions and conclusion The current study uniquely focuses on Roma adolescents, an effectively unknown population in terms of empirical, scholarly work. Interestingly, even though the situation of Roma in Europe, and in the Czech Republic in particular, deserves more attention, it remains a largely “under-researched” topic. Yet, a deeper understanding of cultural and social characteristics of Roma is necessary in order to guide interventions, for instance, which might also include how to improve access to education. We find this work an important first step in an area largely void of empirical scholarship. The evidence suggests that the Roma adolescents included in this study do not differ from ethnic Czech adolescents in terms of rates or developmental processes focused on age-appropriate indicators of deviance. This information by itself contradicts widespread opinions about Roma youth held by the ethnic Czechs and should be communicated to the Czech public in order to provide scientific basis for discussions on inclusion of Roma minority. It also extends the reach of Gottfredson and Hirschi's (1990) work by showing that its theoretical predictions were largely supported in one of the largest ethnic minorities in Europe, and importantly, the study provides evidence for cultural invariance of the relationships between parenting, low self-control, and deviant behaviors. In addition to longitudinal studies that are best equipped to capture causal links between developmental processes, additional work seems essential to further examine Roma populations residing diverse areas, ranging from the most socially excluded areas to fully integrated ones. Understanding how and whether prevalence estimates and the etiology of problem behaviors and adjustment difficulties vary among these groups will help shape policy, and thus interventions, aimed to mitigate the impact of social exclusion on Roma children and adolescents. We look forward, with a sense of optimism, to more rigorous scholarship on Roma youth, to longitudinal work, matching the urgency with which the EU and member states are seeking to address known disparities, and perhaps, the numerous misunderstandings. Acknowledgments We are grateful for the cooperation of all study schools, its administrators, and the students who completed the surveys. We would particularly like to thank Drs. Terezie Pilátová Osecká and Veronika Sobotková for their assistance in collecting the data. Data collection was supported in part by a Fulbright-Masaryk University Distinguished Chair in Social Studies fellowship to the first author.

74

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

Appendix A. Adolescent family process measure (maternal)

1

2

3

4

5

Strongly disagree

Disagree

Neither disagree nor agree

Agree

Strongly agree

Support (reverse coded) 1. My mother sometimes puts me down in front of other people. 2. Sometimes my mother won't listen to me or my opinions. 3. My mother sometimes gives me the feeling that I'm not living up to her expectations. 4. My mother seems to wish I were a different type of person. Monitoring 5. My mother wants to know who I am with when I go out with friends or on a date. 6. In my free time away from home, my mother knows who I'm with and where I am. 7. My mother wants me to tell her where I am if I don't come home right after school. 8. When I am not home, my mother knows my whereabouts.

1

2

3

4

5

Never

Occasionally

Sometimes

Often

Very often

Conflict 9. How often do you have disagreements or arguments with your mother? 10. How often do you purposely not talk to your mother because you are mad at her? 11. How often do you get angry at your mother? Notes 1 The elementary schools in the Czech Republic are divided into a primary or 1st “stage” elementary school, grades 1 through 5, and a secondary or 2nd stage elementary school, grades 6 through 9 (The Ministry of Education, Youth and Sport of the Czech Republic, 2011). Thus, 2nd stage elementary school is equivalent to middle schools in the United States in terms of enrolled students' age. 2 To test whether the method of data collection (paper and pencil versus online) had any undue influence on examined variables, we tested a series of four hierarchical linear regressions predicting each of endogenous variables in our model, that is low selfcontrol and deviance subscales (alcohol use, theft, and school misconduct). In a first step, we added the predictors of each endogenous variable (parenting variables for low selfcontrol, and parenting variables and low self-control for deviance subscales) along with all control variables (sex, age, ethnicity, family structure, and SES). In a second step, we added a dummy coded variable “method of data collection” (paper and pencil versus online) to test whether data collection method predicted unique variance in the dependent measures above and beyond predictors entered into the analysis in the first step. Findings indicated that this term was not significant, and thus, the method of data collection did not influence dependent variables or study results. 3 The low omega coefficient estimate of school misconduct measure in non-Roma sample required some additional attention related to the items part of the scale. Further investigation revealed that factor loading of one item “Have you ever cheated on school/ college/university tests (e.g., cheat sheet, copy from neighbor)?” was very low (λ = .001). As there was no conceptual reason why this would be the case, we tested the reliability of the school misconduct measure in 14 ethnically and nationally diverse samples of adolescents (N N 16,000). The school misconduct scale showed good to excellent reliability across all 14 samples, ranging from ω = .71 in a sample of Slovenian college students to ω = .89 in a sample of Chinese adolescents. Thus, we concluded that the low omega estimate in the sample of non-Roma Czech adolescents was simply idiosyncratic to this sample and retained all items as originally proposed.

References Arbuckle, J. L. (2012). Amos 21 user's guide. Chicago, IL: SPSS. Barnes, G. M., Hoffman, J. H., Welte, J. W., Farrell, M. P., & Dintcheff, B. A. (2006). Effects of parental monitoring and peer deviance on substance use and delinquency. Journal of

Marriage and Family, 68(4), 1084–1104. http://dx.doi.org/10.1111/j.1741-3737.2006. 00315.x. Baumeister, R. F., Heatherton, T. F., & Tice, D. M. (1994). Losing control: How and why people fail at self-regulation. San Diego, CA: Academic Press. Beaver, K. M., Connolly, E. J., Schwartz, J. A., Al-Ghamdi, M. S., & Kobeisy, A. N. (2013). Genetic and environmental contributions to stability and change in levels of self-control. Journal of Criminal Justice, 31, 300–308. http://dx.doi.org/10.1016/j.jcrimjus.2013.07.003. Beaver, K. M., Wright, J. P., DeLisi, M., & Vaughn, M. G. (2008). Genetic influences on the stability of low self-control: Results from a longitudinal sample of twins. Journal of Criminal Justice, 36, 478–485. http://dx.doi.org/10.1016/j.jcrimjus.2008.09.006. Benedik, S., Göderle, W., & Tiefenbacher, B. (2013). Stable concepts in shifting contexts? Promising theoretical approaches to discussions of Romani belonging in Central Europe. In B. Schrammel-Leber, & Tiefenbacher (Eds.), Romani V.: Papers from the annual meeting of the Gypsy Lore Society, Graz 2011 (pp. 135–148). Graz: Grazer Romani Publikationen. Burt, S. A., McGue, M., Iacono, W. G., & Krueger, R. F. (2006). Differential parent-child relationships and adolescent externalizing symptoms: Cross-lagged analyses within a monozygotic twin differences design. Developmental Psychology, 42(6), 1289–1298. http://dx.doi.org/10.1037/0012-1649.42.6.1289. Casey, B. J., Somerville, L. H., Gotlib, I. H., Ayduk, O., Franklin, N. T., Askren, M. K., Jonides, J., ... Shoda, Y. (2001). Behavioral and neural correlates of delay of gratification 40 years later. Proceedings of the National Academy of Sciences of the United States of America, 108, 14998–15003. http://dx.doi.org/10.1073/pnas.1108561108. Chan, H. C., & Chui, W. H. (2015). The influence of low self-control on violent and nonviolent delinquencies: A study of male adolescents from two Chinese societies. The Journal of Forensic Psychiatry and Psychology, 1–21. http://dx.doi.org/10.1080/ 14789949.2015.1012534 (ahead-of-print). Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modelling, 14, 464–504. http://dx.doi.org/10.1080/ 10705510701301834. Cheung, N. W., & Cheung, Y. W. (2008). Self-control, social factors, and delinquency: A test of the general theory of crime among adolescents in Hong Kong. Journal of Youth and Adolescence, 37, 412–430. http://dx.doi.org/10.1007/s10964-007-9218-y. Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233–255. http://dx.doi. org/10.1207/S15328007SEM0902_5. Chui, W. H., & Chan, H. C. O. (2013). Association between self-control and school bullying behaviors among Macanese adolescents. Child Abuse and Neglect, 37, 237–242. http:// dx.doi.org/10.1016/j.chiabu.2012.12.003. Coohey, C., Renner, L. M., & Sabri, B. (2013). Victimization, parenting, and externalizing behavior among Latino and White adolescents. Journal of Family Violence, 28, 359–368. http://dx.doi.org/10.1007/s10896-013-9503-3. de Cortazar, A. R. G., Leon, A. C., Garcia, M. H., & Nunez, J. M. J. (2009). Attitudes of adolescent Spanish Roma toward noninjection drug use and risky sexual behavior. Qualitative Health Research, 19, 605–620. http://dx.doi.org/10.1177/1049732309333813. Coyne, M. A., & Wright, J. P. (2014). The stability of self-control across childhood. Personality and Individual Differences, 69, 144–149. http://dx.doi.org/10.1016/j.paid. 2014.05.026. Crean, H. F. (2008). Conflict in the Latino parent-youth dyad: The role of emotional support from the opposite parent. Journal Of Family Psychology, 22(3), 484–493. http:// dx.doi.org/10.1037/0893-3200.22.3.484. Crutzen, R. (2014). Time is a jailer: What do alpha and its alternatives tell us about reliability? The European Health Psychologist, 16, 70–74. Duckworth, A. L., & Seligman, M. E. P. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological Science, 16, 939–944. http://dx.doi. org/10.1111/j.1467-9280.2005.01641.x. Dunaway, R. G., Cullen, F. T., Burton, V. S., & Evans, T. D. (2000). The myth of social class and crime revisited: An examination of class and adult criminality. Criminology, 38(2), 589–632. Dunbar, E., & Simonova, L. (2003). Individual difference and social status predictors of anti-Semitism and racism: US and Czech findings with the prejudice/tolerance and right wing authoritarianism scales. International Journal of Intercultural Relations, 27, 507–523. http://dx.doi.org/10.1016/S0147-1767(03)00051-8. Dunn, T., Baguley, T., & Brunsden, V. (2014). From alpha to omega: A practical solution to the pervasive problem of internal consistency estimation. British Journal of Psychology, 105, 399–412. http://dx.doi.org/10.1111/bjop.12046. Elliott, D., & Huizinga, D. (1988). Improving self-reported measures of delinquency. In M. W. Klein (Ed.), Cross-national research in self-reported crime and delinquency (pp. 155–186). Dordrecht: Kluwer Academic Publishers. Eurobarometer (2012). Special Eurobarometer 393: Discrimination in the EU in 2012. Retrieved (12/13/2014) from http://ec.europa.eu/public_opinion/archives/ebs/ebs_ 393_en.pdf Fawn, R. (2001). Czech attitudes towards the Roma: “Expecting more of Havel's country?”. Europe Asia Studies, 53, 1193–1219. http://dx.doi.org/10.1080/09668130120093192. Gatenio Gabel, S. (2009). The growing divide: The marginalisation of young Roma children in Bulgaria. International Journal of Social Welfare, 18, 65–75. http://dx.doi.org/ 10.1111/j.1468-2397.2008.00562.x. Gerevich, J., Bacskai, E., Czobor, P., & Szabo, J. (2010). Substance use in Roma and nonRoma adolescents. The Journal of Nervous and Mental Disease, 198, 432–436. http:// dx.doi.org/10.1097/NMD.0b013e3181e07d51. Gibbs, J. J., Giever, D., & Higgins, G. E. (2003). A test of Gottfredson and Hirschi's general theory using structural equation modeling. Criminal Justice and Behavior, 30, 441–458. http://dx.doi.org/10.1177/0093854803253135. Ginter, E., Krajcovicova-Kudlackova, M., Kacala, O., Kovacic, V., & Valachovicova, M. (2001). Health status of Romanies (gypsies) in the Slovak Republic and in the neighbouring countries. Bratislavské Lekárske Listy, 102, 479–484.

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76 Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford: Stanford University Press. Gottfredson, M., & Hirschi, T. (1996). The true value of lambda would appear to be zero: An essay on career criminals, criminal careers, selective incapacitation, cohort studies, and related topics. In D. F. Greenberg, & D. F. Greenberg (Eds.), Criminal careers (pp. 335–356). Brookfield, VT: Dartmouth Publishing Company. 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. http://dx.doi.org/10.1177/0022427893030001002. Griffin, K. W., Scheier, L. M., Acevedo, B., Grenard, J. L., & Botvin, G. J. (2011). Long-term effects of self-control on alcohol use and sexual behavior among urban minority young women. International Journal of Environmental Research and Public Health, 9, 1–23. http://dx.doi.org/10.3390/ijerph9010001. Hay, C. (2001). Parenting, self-control, and delinquency: A test of self-control theory. Criminology, 39, 707–736. http://dx.doi.org/10.1111/j.1745-9125.2001.tb00938.x. Hay, C., & Forrest, W. (2006). The development of self-control: Examining self-control theory's stability thesis. Criminology, 44, 739–774. http://dx.doi.org/10.1111/j.17459125.2006.00062.x. Historie a puvod Romu (2015). Romove v Ceske Republice. Retrieved November 16, 2013, from http://romove.radio.cz/cz/clanek/18530 Hofer, C., Eisenberg, N., Spinrad, T. L., Morris, A. S., Gershoff, E., Valiente, C., ... Eggum, N. D. (2013). Mother–adolescent conflict: Stability, change, and relations with externalizing and internalizing behavior problems. Social Development, 22, 259–279. http:// dx.doi.org/10.1111/sode.12012. Janevic, T., Jankovic, J., & Bradley, E. (2012). Socioeconomic position, gender, and inequalities in self-rated health between Roma and non-Roma in Serbia. International Journal of Public Health, 57, 49–55. http://dx.doi.org/10.1007/s00038-011-0277-1. Kabakchieva, E., Vassileva, S., Kelly, J. A., Amirkhanian, Y. A., Difranceisco, W. J., Mcauliffe, T. L., ... Petrova, E. (2006). HIV risk behavior patterns, predictors, and sexually transmitted disease prevalence in the social networks of young Roma (gypsy) men in Sofia, Bulgaria. Sexually Transmitted Diseases, 33, 485–490. Kelly, J., Amirkhanian, Y., Kabakchieva, E., Csepe, P., Seal, D., Antonova, R., ... Gyukits, G. (2004). Gender roles and HIV sexual risk vulnerability of Roma (gypsies) men and women in Bulgaria and Hungary: An ethnographic study. AIDS Care, 16(2), 231–245. http://dx.doi.org/10.1080/09540120410001641075. Kline, R. B. (2010). Principles and practice of structural equation modelling. New York: The Guilford Press. Kolarcik, P., Geckova, A. M., Orosova, O., van Dijk, J. P., & Reijneveld, S. A. (2010). To what extent does socioeconomic status explain differences in health between Roma and non-Roma adolescents in Slovakia? Social Science & Medicine, 68, 1279–1284. http://dx.doi.org/10.1016/j.socscimed.2008.12.044. Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to parcel: Exploring the question, weighing the merits. Structural Equation Modeling, 9, 151–173. http://dx.doi.org/10.1207/S15328007SEM0902_1. McCarty, C. A., Zimmerman, F. J., Digiuseppe, D. L., & Christakis, D. A. (2005b). Parental emotional support and subsequent internalizing and externalizing problems among children. Journal of Developmental and Behavioral Pediatrics, 26, 267–275. http://dx. doi.org/10.1097/00004703-200508000-00002. Metzler, C. W., Biglan, A., Ary, D. V., & Li, F. (1998). The stability and validity of early adolescents' reports of parenting constructs. Journal of Family Psychology, 12, 600–619. http://dx.doi.org/10.1037/0893-3200.12.4.600. Miller, H. V., Barnes, J. C., & Beaver, K. M. (2011). Self-control and health outcomes in a nationally representative sample. American Journal of Health Behavior, 35, 15–27. http:// dx.doi.org/10.5993/AJHB.35.1.2. Miller, H. V., Jennings, W. G., Alvarez-Rivera, L. L., & Lanza-Kaduce, L. (2009). Self-control, attachment, and deviance among Hispanic adolescents. Journal of Criminal Justice, 37, 77–84. http://dx.doi.org/10.1016/j.jcrimjus.2008.12.003. Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. Science, 244, 933–938. Mistry, R. S., Vandewater, E. A., Huston, A. C., & McLoyd, V. C. (2002). Economic well-being and children's social adjustment: The role of family process in an ethnically diverse low-income sample. Child Development, 73(3), 935–951. Moffitt, T. E. (2005). The new look of behavioral genetics in developmental psychopathology: Gene–environment interplay in antisocial behaviors. Psychological Bulletin, 131, 533–554. http://dx.doi.org/10.1037/0033-2909.131.4.533. Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J., Harrington, H., Houts, R., ... Caspi, A. (2011). A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences of the United States of America, 108, 2693–2698. http://dx.doi.org/10.1073/pnas.1010076108. Morris, G. D., Wood, P. B., & Dunaway, R. G. (2007). Testing the cultural invariance of parenting and self-control as predictors of American Indian delinquency. Western Criminology Review, 8, 32–47. Morton, T. A., Hornsey, M. J., & Postmes, T. (2009). Shifting ground: The variable use of essentialism in contexts of inclusion and exclusion. British Journal of Social Psychology, 48, 35–59. http://dx.doi.org/10.1348/014466607X270287. Nesvadbová, L., Rutsch, J., Kroupa, A., & Sojka, S. (2000). The state of health of the Romany population in the Czech Republic. Central European Journal of Public Health, 8, 141–149. Parekh, N., & Rose, T. (2011). Health inequalities of the Roma in Europe: A literature review. Central European Journal of Public Health, 19, 139–142. Perrone, D., Sullivan, C. J., Pratt, T. C., & Margaryan, S. (2004). Parental efficacy, selfcontrol, and delinquency: A test of a general theory of crime on a nationally representative sample of youth. International Journal of Offender Therapy and Comparative Criminology, 48, 298–312. http://dx.doi.org/10.1177/0306624X03262513. Peters, G. -J. Y. (2014). The alpha and the omega of scale reliability and validity: Why and how to abandon Cronbach's alpha and the route towards more comprehensive assessment of scale quality. The European Health Psychologist, 16, 56–69.

75

Piquero, A. R., Farrington, D. P., Welsh, B. C., Tremblay, R., & Jennings, W. G. (2009). Effects of early family/parent training programs on antisocial behavior and delinquency. Journal of Experimental Criminology, 5, 83–120. http://dx.doi.org/10.1007/s11292009-9072-x. Piquero, A. R., Jennings, W. G., & Farrington, D. P. (2010). On the malleability of selfcontrol: Theoretical and policy implications regarding a general theory of crime. Justice Quarterly, 27, 803–834. http://dx.doi.org/10.1080/07418820903379628. Pogány, I. (2006). Post-communist legal orders and the Roma: Some implications for EU enlargement. In W. Sadurski, A. Czarnota, & M. Krygier (Eds.), Spreading democracy and the rule of law? (pp. 335–356). Dordrecht, Netherlands: Springer. http://dx.doi. org/10.1007/1-4020-3842-9_16. Pratt, T. C., & Cullen, F. T. (2000). The empirical status of Gottfredson and Hirschi's general theory of crime: A meta-analysis. Criminology, 38, 931–964. http://dx.doi.org/10. 1111/j.1745-9125.2000.tb00911.x. Pratt, T. C., Turner, M. G., & Piquero, A. R. (2004). Parental socialization and community context: A longitudinal analysis of the structural sources of low self-control. Journal of Research in Crime and Delinquency, 41, 219–243. Prokop, D. (2015, April 18). Sociolog: Systém produkuje nerovnosti. Školka se krátkodobě nevyplatí. Mladá Fronta Dnes. Retrieved from http://zpravy.idnes.cz/rozhovor-sesociologem-prokopem-d4e-/domaci.aspx?c=A150410_142519_domaci_zt Rankin, B. H., & Quane, J. M. (2002). Social contexts and urban adolescent outcomes: The interrelated effects of neighborhoods, families, and peers on African-American youth. Social Problems, 49, 79–100. http://dx.doi.org/10.1525/sp.2002.49.1.79. Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York, NY: John Wiley and Sons. Sampson, R. J., & Laub, J. H. (1994). Urban poverty and the family context of delinquency: A new look at structure and process in a classic study. Child Development, 65(2), 523–540. Schafer, J. L. (1997). Analysis of incomplete multivariate data. London, UK: Chapman and Hall. Shekarkhar, Z., & Gibson, C. L. (2011). Gender, self-control, and offending behaviors among Latino youth. Journal of Contemporary Criminal Justice, 27, 163–180. http:// dx.doi.org/10.1177/1043986211402224. Sigona, N., & Trehan, N. (2011). Neoliberalism, anti-gypsyism, and the EU's dream deferred (in French). Lignes, 34. Sijtsma, K. (2009). On the use, the misuse, and the very limited usefulness of Cronbach's alpha. Psychometrika, 74, 107–120. http://dx.doi.org/10.1007/s11336-008-9101-0. Simko, V., & Ginter, E. (2009). Short life expectancy and metabolic syndrome in Romanies (gypsies) in Slovakia. Central European Journal of Public Health, 18, 16–18. Sirovátka, T., & Mareš, P. (2006). Poverty, social exclusion and social policy in the Czech Republic. Social Policy and Administration, 40, 288–303. http://dx.doi.org/10.1111/j. 1467-9515.2006.00490.x. Slavíčková, T., & Zvagulis, P. (2014). Monitoring anti-minority rhetoric in the Czech print media: A critical discourse analysis. Journal of Language and Politics, 13, 152–170. http://dx.doi.org/10.1075/jlp.13.1.07sla. Steinberg, L., & Silk, J. S. (2002). Parenting adolescents. In M. H. Bornstein (Ed.), Handbook of parenting. Vol. 1. (pp. 103–133). Mahwah, NJ: Erlbaum. Stejskalova, M. (2013). Can we speak of ghettos in Czech cities? Slovo, 25, 3–17. Stevens, G. W. J. M., Vollebergh, W. A. M., Pels, T. V. M., & Crijnen, A. A. M. (2005). Predicting externalizing problems in Moroccan immigrant adolescents in The Netherlands. Social Psychiatry and Psychiatric Epidemiology, 40, 571–579. http://dx. doi.org/10.1007/s00127-005-0926-x. Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. Journal of Personality, 72, 271–324. http://dx.doi.org/10.1111/j.0022-3506.2004.00263.x. The Ministry of Education, Youth and Sport of the Czech Republic (2011p). The education system in the Czech Republic. Retrieved from: http://www.msmt.cz/mezinarodnivztahy/the-education-system-in-the-czech-republic The Ministry of Education, Youth and Sport of the Czech Republic (2012p). Plán opatření pro výkon rozsudku evropského soudu pro lidská práva v případu D.H. a ostatní proti České Republice. Retrieved from: http://www.msmt.cz/ministerstvo/novinar/msmtpredstavilo-plan-opatreni-k-vykonu-rozsudku-d-h-proti?lang=1 Thornberry, T. P., & Krohn, M. D. (2000). The self-report method for measuring delinquency and crime. Criminal Justice, 4, 33–83. Vaske, J., Ward, J. T., Boisvert, D., & Wright, J. P. (2012). The stability of risk-seeking from adolescence to emerging adulthood. Journal of Criminal Justice, 40, 313–322. http://dx. doi.org/10.1016/j.jcrimjus.2012.06.005. Vazsonyi, A. T., & Belliston, L. M. (2007). THE FAMILY → LOW SELF-CONTROL → DEVIANCE: A cross-cultural and cross-national test of self-control theory. Criminal Justice and Behavior, 34, 505–530. http://dx.doi.org/10.1177/0093854806292299. Vazsonyi, A. T., & Crosswhite, J. M. (2004). A test of Gottfredson and Hirschi's general theory of crime in African American adolescents. Journal of Research in Crime and Delinquency, 41, 407–432. http://dx.doi.org/10.1177/0022427803262060. Vazsonyi, A. T., & Huang, L. (2010). Where self-control comes from: On the development of self-control and its relationship to deviance over time. Developmental Psychology, 46, 245–257. http://dx.doi.org/10.1037/a0016538. Vazsonyi, A. T., Hibbert, J. R., & Snider, J. (2003). Exotic enterprise no more? Adolescent reports of family and parenting processes from youth in four countries. Journal of Research on Adolescence, 13, 129–160. http://dx.doi.org/10.1111/ 1532-7795.1302001. Vazsonyi, A. T., Machackova, H., Sevcikova, A., Smahel, D., & Cerna, A. (2012). Cyberbullying in context: Direct and indirect effects by low self-control across 25 European countries. The European Journal of Developmental Psychology, 9, 210–227. http://dx.doi.org/10.1080/17405629.2011.644919. Vazsonyi, A. T., Pickering, L. E., Junger, M., & Hessing, D. (2001). An empirical test of a general theory of crime: A four-nation comparative study of self-control and the

76

A.T. Vazsonyi et al. / Journal of Criminal Justice 44 (2016) 66–76

prediction of deviance. Journal of Research in Crime and Delinquency, 38, 91–131. http://dx.doi.org/10.1177/0022427801038002001. Vazsonyi, A. T., Roberts, J. W., Huang, L., & Vaughn, M. G. (2015). Why focusing on nurture made and still makes sense: The biosocial development of self-control. In M. DeLisi (Ed.), The Routledge international handbook of biosocial criminology (pp. 350–375). Vazsonyi, A. T., Wittekind, J. E. C., Belliston, L. M., & Loh, T. D. V. (2004). Extending the general theory of crime to “the east:” Low self-control in Japanese late adolescents. Journal of Quantitative Criminology, 20, 189–216. http://dx.doi.org/10.1023/B:JOQC. 0000037731.28786.e3. Vera, E. P., & Moon, B. (2013). An empirical test of low self-control theory among Hispanic youth. Youth Violence and Juvenile Justice, 11(1), 79–93. http://dx.doi.org/10.1177/ 1541204012441628.

Vomastková, K., Souralová, A., & Nekorjak, M. (2011). Stuck in marginality: The education market, ‘Roma schools’ and the reproduction of social and spatial inequalities. Sociologický časopis/Czech Sociological Review, 4, 657–680. Wright, J. P., & Beaver, K. (2005). Do parents matter in creating self-control in their children? A genetically informed test of Gottfredson and Hirschi's theory of low self-control. Criminology, 43, 1169–1202. Yeh, K. -H. (2011). Mediating effects of negative emotions in parent–child conflict on adolescent problem behavior. Asian Journal of Social Psychology, 14, 236–245. http:// dx.doi.org/10.1111/j.1467-839X.2011.01350.