روابط عاشقانه و رفتارهای بزهکارانه در دوران نوجوانی: نقش تعدیلی استعداد بزهکاری

Abstract There is some evidence that adolescent romantic involvement is associated with delinquent behaviour. One aim of this longitudinal study was to determine whether this holds for romantic relationships deemed important by the participants. A second aim was to test whether this association was stronger for adolescents with pre-existing delinquent behaviour and personality traits of impulsivity and thrill seeking (delinquency propensity). Sex differences also were examined. Participants were 686 7th and 8th grade students who completed three assessments over three years. The results showed that delinquency was associated with earlier romantic relationships among those who were higher in delinquency propensity one year earlier. This association was stronger among girls than boys. Thus, romantic relationships amplified girls' and boys' existing delinquency propensity, but this was strongest among girls.

Results Descriptives In total, 58 of the 346 girls (16.8%) and 34 of the 340 boys (10.0%) reported entering a romantic relationship at Time 2. A higher proportion of girls than boys entered a romantic relationship (χ2(1, n = 686) = 6.75, p < .01). Of the girls who nominated a romantic partner, 44% had an older partner (1–3 years older), 44% had a same-age partner and 12% had a younger partner (1 year younger). Of the boys reporting a romantic relationship, 20% had an older partner (1 year older), 40% had a same-age partner and 40% had a younger partner (1–2 years younger). Thus, girls were more likely than boys to have older romantic partners, whereas boys were more likely than girls to have younger partners (χ2(2, n = 59) = 7.37, p < .05). The correlations between romantic relationship status (Time 2), delinquency (Time 1–3), and impulsivity-thrill seeking (Time 1–2) are presented in Table 1. Table 1. Correlations between romantic relationship status (Time 2), delinquency (Time 1–3), and impulsivity-thrill seeking (Time 1–2) for boys (below the diagonal) and girls (above the diagonal). 1 2 3 4 5 6 Romantic relationships 1.Time 2 – .14* .04 .14** .13* .15** Delinquency 2.Time 1 .17** – .32*** .37*** .35*** .32*** 3.Time 2 .15** .61*** – .37*** .17** .27*** 4.Time 3 .15** .59*** .60*** – .22*** .29*** Impulsivity/Thrill seeking 5.Time 1 .11 .34*** .20*** .26*** – .67*** 6.Time 2 .10 .33*** .32*** .30*** .60*** – Note. *p < .05, **p < .01, ***p < .001. Table options Predicting the likelihood of becoming romantically involved and subsequent delinquency Path analysis was conducted to examine: (a) whether youths with a delinquency propensity are more likely to enter a romantic relationship, (b) whether entering a romantic relationship contributes to youths' delinquency, over and above their propensity towards delinquent behaviour, and (c) whether romantic relationships promote delinquency most strongly for youths who are prone to delinquency. In order to examine whether youths' propensity towards delinquent behaviour moderates the effect of romantic relationships on subsequent delinquency, an interaction term between youths' delinquency propensity and romantic relationship status was entered into the path analysis. In the first set of analyses (Model 1), delinquent behaviour at both Time 1 and Time 2 were used as indicators of youths' delinquency propensity, and an interaction between delinquency (Time 2) and romantic relationship status (Time 2) in the prediction of delinquency (Time 3) was also tested. In the second set of analyses (Model 2), an interaction between impulsivity-thrill seeking (Time 2) and youths' romantic relationship status (Time 2) was included in the model. Initial models The initial model (Model 1) including an interaction between delinquency (Time 2) and romantic relationship status (Time 2) had an adequate fit (χ2(1) = 0.778, p = .378, CFI = 1.000, RMSEA = 0.000). The results from the path analysis, presented in Fig. 1, showed significant stability in delinquency and impulsivity-thrill seeking over time. The results also showed that delinquency and impulsivity-thrill seeking at Time 1 predicted entering a romantic relationship at Time 2. Thus, youths with higher scores on delinquent behaviour or impulsivity-thrill seeking were more likely to become romantically involved. Furthermore, romantic relationships did not predict subsequent delinquency (Time 3), over and above youths' prior delinquency and impulsivity-thrill seeking. However, there was a significant, positive interaction effect between youths' delinquent behaviour and their romantic relationships status in predicting subsequent delinquency, indicating that the effect of romantic relationships on subsequent delinquency was dependent on youths' delinquency propensity. The explained variances in the endogenous variables in the model were: romantic relationships (Time 2: R2 = .045, p < .05), delinquency (Time 2: R2 = .253, p < .001; Time 3: R2 = .358, p < .001), and impulsivity-thrill seeking (Time 2: R2 = .418, p < .001). Standardised path estimates for the initial unconstrained model (Model 1) ... Fig. 1. Standardised path estimates for the initial unconstrained model (Model 1) including an interaction between delinquency (Time 2) and romantic relationships (Time 2). Cross-sectional correlations are not presented in the figure. Figure options Model 2 (not shown), including an interaction between impulsivity-thrill seeking (Time 2) and youths' romantic relationship status (Time 2) also had an adequate fit (χ2(1) = 0.208, p = .648, CFI = 1.000, RMSEA = 0.000). There was a significant interaction between youth's impulsivity-thrill seeking and romantic relationship status predicting subsequent delinquency (β = .08, p < .001). The other parameter estimates and the explained variances in the endogenous variables were almost identical to those in Model 1 and are, therefore, not presented. Thus, in accordance with our hypotheses, the results showed that adolescents who were delinquent or impulsive and thrill seeking more frequently than others became involved in romantic relationships. The romantic relationships, in turn, appeared to increase delinquency for adolescents who were already prone to this kind of behaviour. Multigroup analysis to examine sex differences Multigroup path analysis was performed to test for sex differences. The fit of the unconstrained Model 1, including an interaction between delinquency (Time 2) and romantic relationships (Time 2), was satisfactory (χ2(2) = 0.659, p = .719, CFI = 1.000, RMSEA = 0.000). The results are presented for girls and boys separately in Fig. 2. The model with all path estimates constrained to be equal for boys and girls had significantly poorer fit than the unconstrained model (p < .001), indicating sex differences in some of the estimates. More detailed difference testing was, therefore, performed. There were significant sex differences with respect to stability of delinquency over time (Time 1–Time 2: p < .001; Time 2–Time 3: p < .001), with boys displaying significantly higher stability of delinquency. Delinquency but not impulsivity-thrill seeking predicted entering a romantic relationship for boys, while the results showed the opposite for girls. Nevertheless, there were no significant sex differences in these parameter estimates. Romantic relationships, in turn, did not significantly predict subsequent delinquency for boys or girls. However, the interaction effect between delinquency and romantic relationships was significant for both sexes and a significant sex difference showed that the interaction effect was greater for girls than for boys (p < .001). The explained variances in the endogenous variables in the model for boys were: romantic relationships (Time 2: R2 = .050, ns), delinquency (Time 2: R2 = .372, p < .001; Time 3: R2 = .448, p < .001), and impulsivity-thrill seeking (Time 2: R2 = .379, p < .001). The explained variances for girls were: romantic relationships (Time 2: R2 = .045, ns), delinquency (Time 2: R2 = .103, p < .001; Time 3: R2 = .289, p < .001), and impulsivity-thrill seeking (Time 2: R2 = .457, p < .001). Standardised path estimates for the unconstrained multigroup model (Model 1) ... Fig. 2. Standardised path estimates for the unconstrained multigroup model (Model 1) including an interaction between delinquency (Time 2) and romantic relationships (Time 2) for boys (a) and girls (b). Cross-sectional correlations are not presented in the figure. Figure options The unconstrained model (Model 2, not shown) including an interaction between impulsivity-thrill seeking (Time 2) and youths' romantic relationship status (Time 2) had satisfactory fit (χ2(2) = 1.120, p = .571, CFI = 1.000, RMSEA = 0.000). The parameter estimates for girls and boys and the explained variances in the endogenous variables were almost identical to those in Model 1 and are, therefore, not presented. There was only one way in which this model differed from Model 1. The interaction effect was significant for girls (β = 0.15, p < .001), but it was non-significant for boys (β = −0.01, ns). The fit of the model with all path estimates constrained to be equal for boys and girls was significantly worse than the fit of the unconstrained model (p < .001), indicating sex differences in some estimates. The demonstrated sex differences were also in accordance with the results for the model including an interaction between delinquency (Time 2) and romantic relationship status (Time 2). Thus, delinquency was more stable over time for boys than for girls (Time 1–Time 2: p < .001; Time 2–Time 3: p < .01). Furthermore, the sex difference in the interaction effect between impulsivity-thrill seeking and romantic relationships was significant (p < .001). In sum, the findings showed that boys who engaged in delinquency and girls who were impulsive and thrill seeking were more likely than other adolescents to become romantically involved. Romantic relationships, in turn, were associated with subsequent delinquency for delinquency-prone adolescents and most strongly for girls. Thus, the results confirmed our hypotheses.