دانلود مقاله ISI انگلیسی شماره 38556
ترجمه فارسی عنوان مقاله

قربانی خشونت و درگیری در بزهکاری:بررسی پیش بینی از نظریه فشار عمومی

عنوان انگلیسی
Violent victimization and involvement in delinquency: Examining predictions from general strain theory
کد مقاله سال انتشار تعداد صفحات مقاله انگلیسی
38556 2006 14 صفحه PDF
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Journal of Criminal Justice, Volume 34, Issue 3, May–June 2006, Pages 261–274

ترجمه کلمات کلیدی
- قربانی خشونت - درگیری - بزهکاری
کلمات کلیدی انگلیسی
Violent victimization .involvement .delinquency.
پیش نمایش مقاله
پیش نمایش مقاله  قربانی خشونت و درگیری در بزهکاری:بررسی پیش بینی از نظریه فشار عمومی

چکیده انگلیسی

Abstract In a recent critique and elaboration of general strain theory, Agnew (2001) argued that criminal victimization might be among the most consequential strains experienced by adolescents, and therefore might be an important cause of delinquency. Few studies to date, however, had examined victimization as a potential cause—rather than outcome—of delinquency. This article addresses this void by examining predictions from general strain theory about the effects of victimization on later involvement in delinquency. The analyses indicated that violent victimization significantly predicted later involvement in delinquency, even when controlling for the individual's earlier involvement in delinquency. Moreover, general strain theory appears to be a useful theoretical framework for examining this relationship. The effects of victimization on delinquency were explained in part by its effects on anger (the key intervening variable specified by the theory). Partial support also emerged for the theory's hypothesis that the effects of strain should be conditional upon other factors. Specifically, the effects of victimization were marginally greater for juveniles with weak emotional attachment to their parents and significantly greater for those low in self-control.

مقدمه انگلیسی

Introduction One of the more significant social problems faced by U.S. juveniles is their high rate of violent victimization. National Crime Victimization Survey (NCVS) data indicate that around 1.7 million violent victimizations occur each year to individuals aged nineteen or younger (Maguire & Pastore, 2004). This translates to a rate of approximately fifty victimizations per 1,000 persons—a rate that is double what is seen for adults aged twenty-five to thirty-four. While simple assaults are the most common type of violent victimization for juveniles, more serious incidents also occur at high levels, with NCVS data revealing 500,000 aggravated assaults, 160,000 robberies, and 100,000 rapes in a typical year. The frequency of juvenile victimization also was confirmed in the limited national self-report data that were available. Menard's (2000) analysis of National Youth Survey data, for example, indicated that adolescents and young adults experience on average more than two violent or property victimizations each year. These patterns direct attention to the possible consequences of victimization for juveniles who experience it. Consequences might take many forms, including psychological distress, poor school performance, or financial costs that arise from medical care or replacing stolen property (Miller, Cohen, & Wiersema, 1996). Another consequence of interest to criminologists—and the one that was the focus of this study—involves the possibility that victimization increases one's later involvement in delinquency. In short, might being a victim of a physical attack make someone more likely to commit his/her own later acts of crime and deviance? Research seeking to answer this question was notably lacking in the criminological literature. Aside from research focusing on child abuse in particular (Widom, 1989), the few criminological studies that considered the victimization-delinquency link had done so more with the opposite causal order in mind (Esbensen & Huizinga, 1991, Jensen & Brownfield, 1986 and Lauritsen et al., 1991). Working from an opportunity perspective, those studies considered that delinquency increases one's chances of victimization by increasing proximity to other delinquents. While such an argument is reasonable and may be partially true, it does not itself indicate that victimization has no implications for later delinquency. In short, victimization may increase later delinquency, even after controlling for earlier involvement in delinquency. In a significant elaboration of general strain theory (GST), Agnew, 2001 and Agnew, 2002 brought attention to this possibility. Like other strain theories, GST argues that strainful circumstances pressure individuals into committing delinquent acts. Agnew's (2001) recent elaboration was directed in large part to identifying which strains should be most consequential and why. As part of this discussion, Agnew (2002, p. 306) identified criminal victimization as “among the types of strain that are most likely to lead to delinquency,” pointing out, however, that victimization had been to date “largely neglected as a cause of delinquency.” Agnew, 2001 and Agnew, 2002 argued that victimization should be important because it often will be perceived as unjust and traumatic, and therefore will evoke negative emotions like anger or resentment. These emotions, in turn, increase the chances that a juvenile will resort to crime or deviance in an attempt to cope with their strain. This might involve substance use meant to soothe psychological distress, but also could involve predatory acts directed at the perceived offender or displaced upon others. These arguments call for research that examines this potentially important cause of delinquency and does so in a way that provides insight into the accuracy of GST. The purpose of this article is to provide such research. Specifically, this article examines three hypotheses that are central to the victimization-delinquency relationship and that shed light on GST's ability to explain it. The first hypothesis is that violent victimization should be positively related with later involvement in delinquency, even after including the necessary statistical controls (including a control for involvement in prior delinquency). Second, if GST's view of this relationship is correct, the effects of victimization on delinquency should be at least partly explained by the juvenile's level of anger, the key intervening variable specified by GST. Third, if GST is to be further supported, the effects of victimization should depend in part on variables it identifies as factors that condition the effects of strain. The original statement of GST (Agnew, 1992, pp. 70–74) cited a number of such variables. This article examines this hypothesis by considering two in particular: juveniles' emotional attachment to parents (an indicator of access to conventional social support) and their level of self-control (an indicator of individual coping resources). These hypotheses are examined with data from the National Survey of Children, a panel study of U.S. adolescents and their families (Zill, Furstenburg, Peterson, & Moore, 1990). These data have been used in other investigations of adolescent outcomes (Agnew et al., 2002 and Baumer & South, 2001) and are well-suited to testing these hypotheses in particular. Most importantly, they contain not only strong measures of victimization and delinquency, but also valid and reliable measures of anger and two key conditioning factors specified by GST (parent–child attachment and self-control). Also, the project's panel design enables a longitudinal test that matches the causal order specified by GST. Before further discussing the data, GST is first described in greater detail by summarizing its key arguments and the results from empirical tests. This is followed by a discussion of Agnew's (2001) recent elaboration of the theory that guides new research and that calls attention to victimization as a potentially consequential form of strain.

نتیجه گیری انگلیسی

Results The effects of victimization on delinquency The first step in the analysis was to consider whether there was a significant relationship between victimization and delinquency. To consider this, four equations were estimated. The first of these considered this relationship cross-sectionally—the wave one measure of general delinquency was regressed on wave one victimization and controls for age, sex, race-ethnicity, and family income. The results for this equation are shown in the first column of Table 1. Consistent with GST, there was a significant relationship between victimization and delinquency, with a standardized effect of .153 (p < .001). While suggestive of an effect of victimization on delinquency, the cross-sectional nature of the analysis precluded any establishment of causal order—victimization might increase delinquency (as GST predicts), or this result might simply reflect the greater chances of victimization among those who were already delinquent. Table 1. The effects of violent victimization on delinquency Cross-sectional (wave one delinquency) Longitudinal (wave two delinquency) General delinquency Substance use Violent-property General delinquency Victimization .153⁎⁎ .051⁎ .062⁎ .059⁎ .16 (.03) .05 (.03) .06 (.03) .06 (.03) Age .00 .38⁎⁎ − .01 .24⁎⁎ .00 (.02) .24 (.02) − .01 (.02) .15 (.02) Sex − .21⁎⁎ .01 − .15⁎⁎ − .09⁎⁎ − .43 (.05) .03 (.05) − .31 (.05) − .19 (.05) Race .02 − .09⁎⁎ .03 − .05 .04 (.07) − .20 (.06) .07 (.07) − .10 (.07) Family income − .10⁎⁎ .07⁎ − .05 − .02 − .04 (.01) .02 (.01) − .02 (.01) − .01 (.01) Wave one delinquency – .13⁎⁎ .21⁎⁎ .26⁎⁎ .13 (.03) .21 (.03) .26 (.03) R-squared .08 .18 .09 .16 Note: For each variable, standardized effects appear in the top row, with unstandardized coefficients and standard errors (in parentheses) shown in the bottom row. Also, to provide a more precise view of the effects of victimization, its coefficients were carried out to a third decimal point. ⁎ p < .05. ⁎⁎ p < .01. Table options To consider this issue, the next three equations estimated the effects of wave one victimization on wave two measures of delinquency while controlling for the measure of wave one general delinquency (in addition to controlling for age, sex, race-ethnicity, and family income).5 The results of these equations are shown in the final three columns of Table 1. As expected, wave one general delinquency significantly predicted all three measures of wave two delinquency, with standardized effects of .13 on substance use, .21 on violent-property offending, and .26 on general delinquency. The longitudinal results also revealed, however, persistent significant effects of victimization on delinquency, with effects of .051 on substance use, .062 on violent-property offending, and .059 on general delinquency. Thus, consistent with GST, these results suggested that victimization did increase delinquency (net of its correlation with prior delinquency), and that these effects existed for a range of different delinquent behaviors.6 Considering anger as a mediating variable If GST is to be supported, anger should at least partially mediate the victimization-delinquency relationship. To consider this, four equations were estimated. The first considered whether victimization was in fact associated with an increase in anger. This equation regressed wave two anger on wave one victimization and the controls for age, sex, race-ethnicity, family income, and wave one general delinquency. The results for this equation are shown in the first column of Table 2, and they revealed a significant effect of victimization on anger (B = .055, p = .042). The first half of GST's predicted causal chain therefore was supported—victimization produced a small but statistically significant increase in feelings of anger. Table 2. The effects of violent victimization on anger and delinquency Wave two outcomes Anger Substance use Violent-property General delinquency Wave one victimization .055⁎ .043 .055⁎ .048 .06 (.03) .04 (.03) .06 (.03) .05 (.03) Age − .03 .38⁎⁎ − .01 .25⁎⁎ − .02 (.02) .24 (.02) − .01 (.02) .16 (.02) Sex .01 .01 − .15⁎⁎ − .10⁎⁎ .02 (.05) .03 (.05) − .31 (.05) − .19 (.05) Race − .02 − .09⁎⁎ .03 − .05 .04 (.07) − .20 (.06) .06 (.07) − .10 (.06) Family income − .12⁎⁎ .08⁎⁎ − .04 .00 − .04 (.01) .03 (.01) − .02 (.01) .00 (.01) Wave one delinquency .19⁎⁎ .11⁎⁎ .19⁎⁎ .23⁎⁎ .19 (.03) .11 (.03) .19 (.03) .23 (.03) Wave two anger – .13⁎⁎ .12⁎⁎ .19⁎⁎ .13 (.03) .12 (.03) .19 (.03) R-squared .06 .20 .11 .19 Note: For each variable, standardized effects appear in the top row, with unstandardized coefficients and standard errors (in parentheses) shown in the bottom row. Also, to provide a more precise view of the effects of victimization, its coefficients were carried out to a third decimal point. ⁎ p < .05. ⁎⁎ p < .01. Table options The next three equations considered whether anger in turn was associated with an increase in delinquency. These results are shown in the final three columns of Table 2 and they revealed support for this second half of GST's causal chain—anger had significant and notable effects on delinquency, with betas of .13 for substance use, .12 for violent-property delinquency, and .19 for general delinquency. Moreover, the direct effect of victimization on delinquency ceased to be significant for two of the three measures of delinquency once anger is controlled. Overall, these results were consistent with GST's predictions—victimization increased delinquency in part by affecting anger. The evidence was not overwhelming, given that the coefficients were rather modest in magnitude. This was consistent, however, with what was observed in other longitudinal tests of criminological theory (including tests of GST)—lagged effects were almost always found to be smaller than contemporaneous effects (Agnew, 1991, Agnew & White, 1992, Brezina, 1996 and Thornberry et al., 1994). Indeed, finding an effect of victimization on anger and delinquency reported five years later was notable, given the use of an extensive set of controls that included a control for prior involvement in delinquency. Considering attachment and self-control as conditioning variables The final part of the analysis involved GST's prediction that the effect of strain on delinquency depends on the level of parent-child attachment and the child's level of self-control. This was assessed with product-term analysis that used interaction terms that were the product of the predictor variable (victimization) and the hypothesized conditioning variable (attachment or self-control) (Aiken & West, 1991 and Jaccard et al., 1990). The interaction terms were standardized to enable interpretation of the interactions in terms of standard deviation unit impacts on the victimization-delinquency relationship (see Agnew et al., 2002). Entering the interaction term into an equation that includes the main effects indicates whether the effects of victimization vary across values of the conditioning variable. Specifically, the interaction coefficient indicates how much the effect of victimization increases or decreases in response to a one-unit change in the conditioning variable (Cohen & Cohen, 1983 and Jaccard et al., 1990). GST predicts that the interaction coefficients will be negative—the effect of victimization should be greatest for those with low values of attachment and self-control. Two equations were estimated to assess each of the two conditioning variables. The first included the control variables and main effects for both victimization and the conditioning variable. The second equation then added the interaction term. Table 3 provides the results for the interaction between victimization and attachment for the three delinquency measures. These results suggested that no interaction existed. While the three Equation 1s revealed that attachment was indeed significantly associated with each measure of delinquency, the three Equation 2s revealed that the interactions between victimization and attachment were insignificant. For two of the three (substance use and general delinquency), the interaction was negative (suggesting that victimization leads to greater delinquency among those low in attachment), but it missed statistical significance, with betas of .04 that had p-values of .08 and .15. Table 3. The interaction between victimization and parent-child attachment Substance use Violent-property General delinquency 1 2 1 2 1 2 Age .35⁎⁎ .35⁎⁎ − .04 − .04 .21⁎⁎ .21⁎⁎ .22 (.02) .22 (.02) − .02 (.02) − .02 (.02) .13 (.02) .13 (.02) Sex − .01 − .01 − .17⁎⁎ − .17⁎⁎ − .12⁎⁎ − .12⁎⁎ − .02 (.05) − .02 (.05) − .35 (.05) − .35 (.05) − .25 (.05) − .25 (.05) Race − .08⁎⁎ − .08⁎⁎ .04 .04 − .03 − .04 − .18 (.06) − .18 (.06) .08 (.07) .08 (.07) − .08 (.06) − .08 (.06) Family income .07⁎⁎ .07⁎ − .05 − .05 − .02 − .02 .02 (.01) .02 (.01) − .02 (.01) − .02 (.01) − .01 (.01) − .01 (.01) Wave one delinquency .11⁎⁎ .11⁎⁎ .19⁎⁎ .19⁎⁎ .23⁎⁎ .23⁎⁎ .11 (.03) .10 (.03) .19 (.03) .19 (.03) .23 (.03) .23 (.03) Victimization .04 .04 .06⁎ .06⁎ .05⁎ .05⁎ .04 (.03) .04 (.03) .06 (.03) .06 (.03) .05 (.03) .05 (.03) Attachment − .21⁎⁎ − .21⁎⁎ − .15⁎⁎ − .15⁎⁎ − .25⁎⁎ − .24⁎⁎ − .21 (.03) − .21 (.03) − .16 (.03) − .16 (.03) − .25 (.03) − .25 (.03) Victimization x attachment – − .04 – .00 – − .04 − .04 (.03) .00 (.03) − .04 (.03) R-squared .221 .223 .116 .116 .212 .213 Note: For each variable, standardized effects appear in the top row, with unstandardized coefficients and standard errors (in parentheses) shown in the bottom row. ⁎ p < .05. ⁎⁎ p < .01. Table options Table 4 provides results for equations in which self-control was the conditioning variable. Contrary to what was just observed, these results suggested that an interaction was occurring. Self-control was significantly and negatively related with each measure of delinquency, and two of the three interactions terms (for substance use and general delinquency) were significant as well. The third (for violent-property offending) had a p-value of .096 that marginally missed the traditional level of significance of p = .05. 7 Table 4. The interaction between victimization and self-control Substance use Violent-property General delinquency 1 2 1 2 1 2 Age .39⁎⁎ .39⁎⁎ − .01 − .01 .25⁎⁎ .25⁎⁎ .24 (.02) .24 (.02) − .01 (.02) − .01 (.02) .16 (.02) .16 (.02) Sex .03 .03 − .14⁎⁎ − .14⁎⁎ − .07⁎⁎ − .07⁎⁎ .07 (.05) .07 (.05) − .28 (.05) − .27 (.05) − .14 (.05) − .13 (.05) Race − .08⁎⁎ − .08⁎⁎ .03 .03 − .03 − .03 − .18 (.06) − .18 (.06) .08 (.07) .07 (.07) − .07 (.06) − .08 (.06) Family income .09⁎⁎ .09⁎⁎ − .03 − .04 .01 .01 .03 (.01) .03 (.01) − .01 (.01) − .01 (.01) .00 (.01) .00 (.01) Wave one delinquency .11⁎⁎ .11⁎⁎ .19⁎⁎ .19⁎⁎ .23⁎⁎ .23⁎⁎ .11 (.03) .11 (.03) .19 (.03) .19 (.03) .23 (.03) .23 (.03) Victimization .04 .04 .05⁎ .05 .04 .04 .04 (.03) .04 (.03) .05 (.03) .05 (.03) .04 (.03) .04 (.03) Self-control − .13⁎⁎ − .13⁎⁎ − .12⁎⁎ − .12⁎⁎ − .19⁎⁎ − .19⁎⁎ − .13 (.03) − .13 (.03) − .12 (.03) − .12 (.03) − .19 (.03) − .19 (.03) Victimization x self-control – − .07⁎⁎ – − .04 – − .05⁎ − .07 (.03) − .04 (.03) − .05 (.03) R-squared .196 .200 .107 .109 .188 .191 Note: For each variable, standardized effects appear in the top row, with unstandardized coefficients and standard errors (in parentheses) shown in the bottom row. ⁎ p < .05. ⁎⁎ p < .01. Table options In substantive terms, these significant, negative interactions supported GST—they indicated that victimization produced a greater increase in delinquency for those low in self-control. With all variables having been standardized, this pattern can be described in terms of standard deviation unit changes in the victimization-delinquency relationship. For general delinquency, for example, a standardized effect of victimization of .04 when self-control is estimated at its mean increases to .09 when self-control is one standard deviation below its mean; this effect increases to .14 when self-control is two standard deviation units below its mean. For substance use, the impact of the interaction is greater. An effect of victimization of .04 when self-control is at its mean increases to .11 when self-control is one standard deviation unit below its mean and to .18 when it is two standard deviation units below its mean.