قربانی شدن به عنوان علت بزهکاری: نقش افسردگی و جنسیت

Abstract Victimization as a correlate of delinquency has been largely neglected in the criminological literature, despite research on Agnew's general strain theory (GST) suggesting that victimization is a type of strain likely to cause delinquency (Agnew, 2002). This study examined the role of depression and gender as potentially indispensable mechanisms in the victimization-delinquency relationship. Findings indicated that victimization has a positive effect on both delinquency and depression, and consistent with a GST explanation, the connection between victimization and delinquency is most pronounced for males with trait depression. This study added to the GST literature by distinguishing between trait and state emotion and by delineating GST predictions regarding each emotional form.

Introduction While studies have shown a clear empirical link between criminal behavior and victimization (Esbenson and Huizinga, 1991 and Fagan et al., 1987), there is relatively little focus in the criminological literature on victimization as a cause of crime. Much of the victimization research has highlighted the homogeneity of offender and victim characteristics and suggested that victims and offenders may not represent two truly distinct groups (Sampson and Lauritsen, 1990, Singer, 1981 and Wolfgang, 1958). The literature examining a potentially causal relationship between offending and victimization has tended to approach criminal behavior as an antecedent to victimization (see Lauritsen et al., 1991, Mustaine and Tewksbury, 1998, Nofziger and Kurtz, 2005 and Sampson and Lauritsen, 1990). In recent years, however, a small literature has begun to explicitly examine the effect of victimization on offending using Agnew's general strain theory (GST) and found a positive effect (Hay and Evans, 2006, Moon et al., 2008, Ostrowsky and Messner, 2005, Piquero and Sealock, 2004 and Spano et al., 2006). This study added to the GST literature by testing the extent to which victimization increases delinquency among a sample of juveniles. It further explored the role of depression and gender in the GST process by distinguishing between state and trait depression, outlining GST predictions for each, and testing for effects separately by gender.

Results Descriptive statistics Table 1 presents the means, standard deviations, and alphas for the independent variables in the full sample. As seen in this table, the majority of respondents did not experience symptoms of depression. About 75 percent of respondents reported not having a period of two weeks or more in which they experienced depressive symptoms (1,003 out of 1,332). Out of those who did report depressive feelings, the mean on the depression scale was 3.4 with a standard deviation of 2.6. Most respondents also did not report being victimized in the past year. Of the ones that did, the mean on the victimization scale was 2.0 with a standard deviation of 1.3. Table 1. Means, standard deviations, ranges, and alphas for variables used in analyses (N = 1,332) M SD R α Victimization5 .98 1.38 0–11 .60 Depression6 .85 2.0 0–15 .78 Delinquency6 .55 .50 0–1 -- Gender (female = 1)6 .48 .50 0–1 -- Race (African American = 1)6 .15 .36 0–1 -- Race (other = 1)6 .05 .23 0–1 -- Age6 20.79 1.96 18–24 -- Neighborhood disorganization6 1.31 .33 1–3 .89 Marital status (married = 1)6 .25 .43 0–1 -- Attachment to family6 4.14 .71 1–5 .70 Attachment to employment6 3.63 1.51 0–5 .70 Delinquent peers6 1.87 .55 1–5 .82 Prior delinquency5 1.13 .20 1–9 .85 5 = Measured from wave five of the NYS (Time 1). 6 = Measured from wave six of the NYS (Time 2). Table options Analysis of variance revealed expected findings; males and females experienced significantly different amounts of depressive feelings (F(1, 1330) = 29.45, p.001) as well as significantly different amounts of experiences with victimization (F(1, 1330) = 33.19, p.001). While males experienced significantly more victimization (Mmales = 1.18 versus Mfemales = .75), females experienced significantly more depressive symptoms (Mfemales = 1.15 versus Mmales = .57). Table 2 presents zero-order correlations. Though some correlations, such as the correlations between prior delinquency and delinquent peers (r = .43) and victimization and prior delinquency (r = .43), are relatively high, they do not suggest that multicollinearity is a threat to analysis. Table 2. Zero-order correlation coefficients of variables used in analyses (N = 1,332) 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Victimization5 1.00 2. Depression6 .16 1.00 3. Delinquency6 .24 .10 1.00 4. Age6 .04 .05 − .06 1.00 5. Gender (female = 1)6 − .16 .16 − .24 − .03 1.00 6. Race (African American = 1)6 .00 − .06 − .06 − .02 − .04 1.00 7. Race (other = 1)6 − .01 .05 − .03 − .05 .00 − .10 1.00 8. Marital status (married = 1)6 − .04 − .03 − .16 .28 .16 − .08 − .01 1.00 9. Neighborhood disorganization6 .14 .11 .07 .01 − .01 .22 .06 − .01 1.00 10. Attachment to family6 − .12 − .15 − .18 − .08 .07 .09 .02 .02 − .05 1.00 11. Attachment to employment6 − .00 − .06 − .02 .13 − .12 − .10 − .02 .02 − .11 .04 1.00 12. Delinquent peers6 .25 .15 .45 − .03 − .15 − .10 − .05 − .11 .13 − .27 − .02 1.00 13. Prior delinquency5 .43 .14 .30 − .04 − .20 − .02 .01 − .06 .15 − .21 − .03 .43 1.00 5 = Measured from wave five of the NYS (Time 1). 6 = Measured from wave six of the NYS (Time 2). Table options Regression analyses Hypotheses in this study predicted that Time 1 victimization would predict the tendency to experience depression in Time 2 (Hypothesis 1), that the tendency to experience depression in Time 2 would influence delinquent behavior also occurring in that time (Hypothesis 2), that Time 1 victimization would directly predict delinquent behavior in Time 2 (Hypothesis 3), but that the tendency to experience depression in Time 2 would not account for this relationship in full as would a true mediating variable (Hypothesis 4). Rather, it was expected that Time 2 depression would moderate the relationship between Time 1 victimization and Time 2 delinquency, such that trait depression would make a delinquent response to victimization more likely (Hypothesis 5). Lastly, it was expected that that the role of depression in the victimization-delinquency relationship would be particularly strong for males (Hypothesis 6). OLS regression results in Table 3 show full support for Hypothesis 1: every standard deviation increase on the victimization scale increased the score on the depression scale by .12 standard deviations. Expected gender differences were also found as females were more likely than males to have depressive tendencies. Logistic regression results in Table 4 also show full support for Hypothesis 2 and Hypothesis 3: for every one point increase on the depression scale, the odds of being delinquent increased by 10 percent (Model A), and for every one point increase on the victimization scale, the odds of being delinquent increased by 19 percent (Model B). Females, however, were much less likely to be delinquent in both models. Table 4 also shows that trait depression does not completely mediate the victimization-delinquency relationship found: the coefficient for victimization retained significance and was reduced only by .02 (Model C). While GST predicts that state feelings of sadness would have had such a mediating effect, this effect was not anticipated for the current operationalization reflecting trait depression, or the tendency to experience symptoms of depression. Instead, a moderating effect was anticipated for this conceptualization and measurement (Hypothesis 5), and this moderating effect was expected to be stronger for males when compared to females (Hypothesis 6). Table 3. OLS regression of victimization and control variables on depression (N = 1,332) Depression b SE β Victimization .17⁎⁎⁎ .04 .12 Gender (female = 1) .79⁎⁎⁎ .11 .20 Race (African American = 1) − .15 .15 − .03 Race (other = 1) .56⁎ .23 .06 Age .06⁎ .03 .06 Marital status (married = 1) − .33⁎⁎ .13 − .07 Neighborhood disorganization .37⁎ .17 .06 Attachment to family − .31⁎⁎⁎ .08 − .11 Attachment to employment − .03⁎ .04 − .02 Delinquent peers .22⁎ .11 .06 Prior delinquency .01 .01 .05 Adjusted R2 .088 ⁎p ≤ .05. ⁎⁎p ≤ .01. ⁎⁎⁎p ≤ .001. Table options Table 4. Logistic regression of depression, victimization and control variables on delinquency (N = 1332) Delinquency Model A Model B Model C Victimization 1.19⁎⁎ 1.17⁎⁎ (.07) (.07) Depression 1.10⁎⁎⁎ 1.09⁎ (.04) (.04) Gender (female = 1) .47⁎⁎⁎ .51⁎⁎⁎ .48⁎⁎⁎ (.07) (.07) (.07) Race (African American = 1) .80 .78 .79⁎⁎⁎ (.16) (.15) (.16) Race (other = 1) .79 .85 .78 (.23) (.25) (.23) Age .95 .95 .94 (.03) (.03) (.03) Marital status (married = 1) .61⁎⁎ .59⁎⁎⁎ .61 (.10) (.09) (.10) Neighborhood disorganization 1.10 1.09 1.06⁎⁎ (.24) (.24) (.23) Attachment to family .88 .86 .87 (.09) (.08) (.09) Attachment to employment .97 .97 .97 (.04) (.04) (.04) Delinquent peers 5.45⁎⁎⁎ 5.52⁎⁎⁎ 5.44⁎⁎⁎ (.88) (.89) (.89) Prior delinquency 1.14⁎⁎⁎ 1.18⁎⁎⁎ 1.12⁎⁎⁎ (.03) (.03) (.03) Pseudo R2 .243 .243 .247 Note: Odds ratios shown. Standard errors in parentheses. ⁎p ≤ .05. ⁎⁎p ≤ .01. ⁎⁎⁎p ≤ .001. Table options A dichotomous variable assessing whether or not a person experienced depressive symptoms within the past year was created (1 = yes, 0 = no) and multiplied by the victimization variable in order to test these predictions. Table 5 displays the results for males, females, and the full sample. As seen in the full sample, those who experienced depressive symptoms in the past year were 25 percent more likely to respond to victimization with delinquency. This provides support for Hypothesis 5, but in support of Hypothesis 6, this moderating effect was limited to the male sample. Males who experienced depressive symptoms in the past year were 50 percent more likely to respond to victimization with delinquency than males who did not. Females who experienced depressive symptoms in the past year were no more likely than other females to respond to victimization with delinquency. Table 5. Logistic regression of depression x victimization and control variables on delinquency for males (N = 686), females (N = 646), and the full sample (N = 1332) Delinquency Males Females Full Sample Victimization 1.17+ 1.05 1.11 (.11) (.10) (.07) Depression .99⁎⁎ 1.07 1.05 (.10) (.05) (.05) Victimization X Depression 1.50⁎ 1.10 1.25+ (.31) (.19) (.16) Gender (female = 1) .49⁎⁎⁎ (.07) Race (African American = 1) .72 .83 .79 (.19) (.25) (.15) Race (other = 1) .81 .73 .79 (.34) (.31) (.23) Age .91 .98 .95 (.05) (.05) (.03) Marital status (married = 1) .59⁎ .59⁎⁎ .60⁎⁎ (.15) (.12) (.10) Neighborhood disorganization .96 1.13 1.05 (.31) (.34) (.23) Attachment to family .87 .88 .87 (.13) (.12) (.09) Attachment to employment .99 .96 .97 (.07) (.06) (.04) Delinquent peers 8.68⁎⁎⁎ 3.68⁎⁎⁎ 5.47⁎⁎⁎ (2.12) (.81) (.89) Prior delinquency 1.07⁎⁎ 1.22⁎⁎⁎ 1.11⁎⁎⁎ (.03) (.05) (.03) Pseudo R2 .276 .177 .248 Note: Odds ratios shown. Standard errors in parentheses. +p ≤ .10. ⁎p ≤ .05. ⁎⁎p ≤ .01. ⁎⁎⁎p ≤ .001.