فعالیت های معمول به عنوان عوامل موثر بر تفاوت های جنسیتی در بزهکاری
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|38580||2010||8 صفحه PDF||سفارش دهید||7025 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Criminal Justice, Volume 38, Issue 5, September–October 2010, Pages 913–920
Abstract This study examined the extent to which gender differences in delinquency can be explained by gender differences in participation in, or response to, various routine activity patterns (RAPs) using data from the second and third waves of the National Education Longitudinal Survey of 1988. While differential participation in routine activities by gender failed to explain males’ high levels of deviance relative to females, two early RAPs moderated the effect of gender on subsequent deviant behavior. Participation in religious and community activities during the sophomore year in high school decreased, while unstructured and unsupervised peer interaction increased, levels of delinquency two years later substantially more for males than for females, suggesting there are gender differences in reactivity to contextual opportunities for deviance during early high school with effects that persist over time.
Introduction Although the gender gap in property and violent offenses as well as in more general deviance has declined, especially among youth, adolescent males consistently exhibit higher levels of delinquency than their female counterparts (Regoli, Hewitt, & DeLisi, 2010). Working within the routine activities framework (Cohen and Felson, 1979 and Hawdon, 1996), the extent to which this difference might be explained by gender differences in participation in, or response to, common sets of behavioral patterns is assessed. Routine activities as social control Control theory is one of the most widely tested models of juvenile delinquency (see Kempf, 1993 for a review of these studies). Key to this theoretical perspective is the assumption that internal bonds to society prevent deviance through attachment to individuals or institutions that uphold the normative order, commitment to or investment in institutions that promote conformity, belief in the validity of societal mores, and involvement in conventional activities that limit opportunities for delinquency (Hirschi, 1969). Arguing that conventional activities can themselves provide opportunities for deviance insofar as they lack purpose and are invisible to agents of social control, Hawdon, 1996 and Hawdon, 1999 redefined Hirshi's (1969) concept of involvement as participation in routine activities, a construct initially used to explain crime victimization (Cohen and Felson, 1979 and Felson, 1994). In an extension of this literature, Hawdon, 1996 and Hawdon, 1999 examined the relationship between what he termed routine activity patterns (RAPs) and delinquency. RAPs are relatively stable clusters of related behaviors that characterize individuals’ daily routines that vary in structure and purpose (instrumentality) and visibility to agents of social control. Hawdon (1996) suggested that RAPs that are high in both instrumentality and visibility should reduce the frequency of delinquent behaviors, while RAPs centered on activities that lack these characteristics should increase delinquency by enhancing adolescents’ opportunities for deviance. Thus, unlike Hirshi's (1969) control theory, this model emphasizes external, rather than internal, social controls. Consistent with Hawdon's predictions, measures of delinquency were inversely associated with participation in purposeful activities with high visibility, and positively related to involvement in unstructured and unsupervised social activities, among both high school (Hawdon, 1996) and college students (Hawdon, 1999). A number of other studies have yielded similar findings. Participation in structured academic, extracurricular, community, and religious activities has been associated with low levels of delinquency, while unstructured and unsupervised peer interaction has been shown to increase adolescents’ risks for deviance (Agnew and Petersen, 1989, Anderson and Hughes, 2009, Barnes et al., 2007, Bernburg and Thorlindsson, 2001, Crawford and Novak, 2002, Eccles and Barber, 1999, Flannery et al., 1999, Fleming et al., 2008, Haynie and Osgood, 2005, Huebner and Betts, 2002, Mahoney and Stattin, 2000, Osgood et al., 1996, Thorlindsson and Bernburg, 2006, Vazsonyi et al., 2002 and Wong, 2005). Modeling the relationship between gender, routine activities and delinquency A few of the studies cited in the preceding section examined the relationship between gender, routine activities, and delinquency. Borrowing from the literature on gender and social control more generally (see Costello and Mereder, 2003, Jensen and Eve, 1976 and White and LaGrange, 1987), these analyses can be categorized based on the extent to which they emphasized mediating or moderating relationships between key variables. Mediating models specify the mechanisms through which independent variables influence dependent variables indirectly by identifying intermediate, or intervening, variables in a causal chain. For a mediating effect to exist an independent and dependent variable must be correlated, the third (mediating) variable must be associated with both the independent and the dependent variable, and one must be able to safely assume that the mediating variable is the result (rather than the cause) of the independent variable. When these conditions are met, mediating effects are detectable through a series of analyses in which variables are sequentially added into a statistical model (Baron & Kenny, 1986). Within the context of the literature on gender and routine activities, a mediating effect is presumed to exist when the association between gender and delinquency disappears when measures of routine activities are included in the analysis. If routine activity patterns account for the effect of gender on delinquency in this manner, this suggests that that males are more delinquent than females simply because they are more likely to participate in RAPs conducive to deviance. As such, the mediation hypothesis is what White and LaGrange (1987) termed a common causes argument. In support of the common causes position, Osgood et al. (1996) found that gender differences in participation in various routine activities, including unstructured peer interaction, accounted for much of the effect of gender on each of five types of deviant behavior (heavy drinking, marijuana use, the use of illicit drugs, crime and dangerous driving). Other analyses have, however, failed to show any mediating influences of routine activity patterns, including peer interaction (Anderson & Hughes, 2009), sports, religious activities and school clubs (Chapple, McQuillan, & Berdahl, 2005), on the gender-delinquency relationship. A second group of studies focusing on gender and routine activities emphasizes moderating over mediating influences. While mediating models link an independent variable to a dependent variable through one or more intervening variables, moderating models specify interaction effects (Baron & Kenny, 1986). When the effect of gender on delinquency varies across levels of participation in various activities, evidenced by significant cross-product interactions in a regression model, RAPs are said to moderate the gender-delinquency relationship. The moderation model proposes that gender differences in delinquency are due to a heightened reactivity to RAPs low in structure and visibility, or immunity to the protective effects of RAPs high in these attributes, among males relative to their female counterparts. In support of this model, Mahoney and Stattin (2000) have shown that low-structure recreational behaviors have a stronger effect on antisocial behavior among boys than among girls. Similarly, Crawford and Novak's (2002) study indicated that unstructured/unsupervised peer interaction during the sophomore year in high school increased the risk for subsequent drinking primarily among males. In their analysis of routine activities and delinquency in U.S. and three other countries, on the other hand, Vazsonyi et al. (2002) found few gender differences in the effects of various activity clusters (hanging out with friends, school-based and sports activities, and solitary pursuits) on measures of delinquency. Furthermore, there was no evidence that the effects of peer interaction on smoking, drug use, and general delinquency varied by gender in Barnes et al.'s (2007) study of adolescents’ use of time. Although an earlier analysis by Huebner and Betts (2002) yielded a larger inverse association between purposeful activities, such as clubs and hobbies, and a measure of general deviance among males than among females, these behaviors emerged as protective for both genders. Consistent with this, Fleming et al. (2008) found little evidence that the effects of after-school activities on delinquency varied by gender. Thus, to date, the literature on the relationship between gender, routine activities, and delinquency has been equivocal. Although there was some evidence that frequency of peer interaction, in particular, may increase the risk for deviance more for males than for females, this effect may be specific to social encounters characterized by low structure and visibility. The current study assessed the extent to which unstructured peer interactions, as well as a variety of other RAPs varying in both structure and visibility, mediated or moderated the relationship between gender and delinquency. Like Hawdon, 1996 and Hawdon, 1999 earlier work, many of the studies that examined gender and routine activities described above were cross-sectional in design, making it difficult to determine the causal direction of the relationships in question. This may be especially problematic when one considers the effects of peer interaction, as delinquent youth may be inclined to pursue encounters with friends that lack both structure and visibility (Crawford and Novak, 2008). Using data from the first and second follow ups of the National Education Longitudinal Survey (NELS:88), we assessed the effects of gender and a number of routine activity patterns on delinquency during the senior year in high school controlling for prior levels of deviant behavior.
نتیجه گیری انگلیسی
Results Bivariate correlations between RAPs (Time 1 – Time 2), between gender and both sets of RAPs, and the RAPs and delinquency are presented in Table 1. As shown in Column 1 of Table 1, correlations between each RAP (measured during the sophomore and again during the senior year in high school) were moderate in size, ranging from .29 (the social pattern) to .56 (the music/art pattern), suggesting a fair degree of continuity in the types of activities adolescents participate in over time. Table 1. Bivariate Correlations between RAPs, Gender, and Delinquency (n = 1,485 weighted) Time-1 RAPa (Matched) Female Time-2 Delinquency Time – 1 RAPs Academic .13*** -.10*** Religious/Community .11*** -.21*** Athletic -.25*** .05 Social -.05 .30*** Music/Art .16*** -.13*** Adults .08** -.08** Clubs .13*** -.08** Hobbies -.10*** -.10*** Nonsocial -.07** -.03 Time-2 RAPs Academic .35*** .06* -.20*** Religious/Community .51*** .04 -.22*** Athletic .53*** -.26*** .05 Social .29*** .02 .21*** Music/Art .56*** .09** -.13*** Adults .40*** .13*** -.04 School Clubs .40*** .14*** -.12*** Hobbies .51*** -.07** -.04 Nonsocial .35*** -.09** -.08** aCorrelations are for each Time-1 RAP and its Time-2 counterpart. *p < .05, **p < .01, ***p < .001. Table options With one exception, changes in RAPs over time did not significantly vary across gender. Interestingly, the social RAP was more stable over the two-year timeframe under investigation among males than among females. Additional analyses (data not shown) indicated that this was largely a function of gender differences in scores on this variable at Time 1. During the sophomore year, males participated in significantly more unstructured/unsupervised interactions than females. By the time respondents were seniors, scores on the social RAP were similar for the two genders. It is unclear whether this shift was due to the changes in question wording noted earlier or to the effect of gender on respondents’ social behaviors as they progressed through high school. Although the substitution of questions across waves (from an indicator of unstructured to any peer interaction) was less than ideal, it was less problematic in subsequent high-order analyses. Within the context of these models, the effects of the social RAP on delinquency were estimated with controls for structured peer activities in the form of the various RAPs. Preliminary bivariate analyses showed that males were significantly more deviant than their female counterparts (tTime 1 = 5.54, df 3645, p < .001; tTime 2 = 12.25, df 2813, p < .001). As shown in Column 2 of Table 1, there were also gender differences in scores on most of the RAPs. Moreover, all of the RAPs, with the exception of the athletic, the Time-1 nonsocial and the Time-2 adult and hobby orientations, were significantly associated with delinquency. Thus, the overall pattern of results in Table 1 met the preliminary criteria for mediating effects--in this case a negative (positive) correlation between the variable female and a particular RAP, and a positive (negative) correlation between that RAP and the measure of delinquency. A series of OLS regressions were used to assess the relationships between gender, RAPs and deviant behavior with controls for race/ethnicity, socioeconomic background, religiosity, peer support for substance use, and social bonds. The listwise deletion of missing cases yielded a final sample size of 1485 for each model. The initial analysis was conducted in a series of four steps. First, the deviance index was regressed on gender and the control variables, including the measures of bonds to society and Time-1 delinquency. In a second step, the nine Time-1 RAPs were added into the regression model. This enabled us to both assess the effects of early activity patterns on deviant behavior and to determine whether this set of variables mediated the gender-delinquency relationship. The results of these analyses are presented in Columns 1 (gender and the control variables) and 2 (gender, the control variables, and the Time-1 RAPs) of Table 2, respectively. Table 2. OLS Regressions Predicting Total Time-2 Delinquency Scores (n = 1,485 weighted) Step 1 Step 2 Step 3 Step 4 b Beta b Beta b Beta b Beta Constant .92 1.68 .50 .40 Female -1.16*** -.09 -1.04*** -.08 -1.08*** -.08 -1.13*** -.09 Asian -.53 -.02 -.47 -.01 -.32 -.01 -.32 -.01 Black -1.10* -.04 -.94 -.04 -.58 -.02 -.68 -.03 Hispanic .20 .01 .25 .01 .55 .02 .47 .02 Native -.33 .00 -.13 .00 .08 .00 .04 .00 SES .19 .02 .27 .03 .36 .04 .36 .04 Religiosity -.69** -.07 -.60** -.06 -.29 -.03 -.26 -.03 Peer Support 1.81 .19 1.73*** .19 1.61*** .17 1.62*** .17 Attachment .01 .01 -.01 -.01 .00 .00 -.01 -.01 Commitment .03 .01 .01 .00 .01 .00 .01 .00 Belief -.17*** -.09 -.16*** -.08 -.14** -.07 -.13** -.07 T1 Delinquency .55*** .55 .53*** .53 .52** .53 .53*** .53 T1 Academic -.01 .00 .03 .02 .03 .02 T1 Rel/Comm -.11 -.04 -.05 -.02 -.19* -.06 T1 Athletic .13* .04 .14 .04 .14 .04 T1 Social .17* .04 .07 .02 .27* .07 T1 Music/Art -.03 -.01 -.02 -.01 -.03 -.01 T1 Adult .14 .03 .16 .04 .15 .04 T1 Clubs -.01 .00 .01 .00 .01 .00 T1 Hobbies -.08 -.02 -.04 -.01 -.04 -.01 T1 Nonsocial -.11 -.02 -.06 -.01 -.06 -.01 T2 Academic -.17*** -.08 -.17*** -.08 T2 Rel/Comm -.21** -.07 -.21** -.07 T2 Athletic .03 .01 .03 .01 T2 Social .39*** .09 .38*** .09 T2 Music/Art .01 .00 .01 .00 T2 Adult -.16 -.04 -.16 -.04 T2 Clubs .04 .01 .04 .01 T2 Hobbies -.02 -.01 -.01 .00 T2 Nonsocial -.20* -.04 -.20* -.04 T1 Rel*Female .24* .06 T1 Social*Female -.37* -.07 R-Square .502*** .509*** .525*** .529*** Ch R-Square .006* .017*** .003** *p < .05, **p < .01, ***p < .001. Table options In a third analysis, we added the Time-2 RAPs into the regression model (Table 2, Column 3), enabling us to assess their individual effects with earlier activity patterns held constant and to determine whether they mediated the gender-delinquency relationship. Since correlations between RAP pairs were all below .60 (Table 1, Column 1), multicollinearity was not a problem. In a final set of regressions, we tested the significance of cross-product interactions between gender, RAPs (measured at Time 1 and then at Time-2) and delinquency in order to determine whether there were any moderating effects. Only those interactions that were statically significant are included in Column 4 of Table 2. As shown in Column 2 of Table 2, two of the Time-1 RAPs, the athletic and the social orientations, increased respondents’ subsequent levels of deviance in the predicted manner. None of the other Time-1 measures of routine activities (the academic, religious, artistic, adult, hobbies, school club and nonsocial orientations) significantly affected scores on the delinquency index in an additive fashion. Thus, the increase in the proportion of explained variation in levels of delinquency resulting from the addition of this block of variables into the statistical model was minimal (change in R2 = .006). Moreover, a comparison of the coefficients for the variable female between Columns 2 and 1 of Table 2 offered no indication of mediation. As routine activity patterns are presumed to affect youth deviance by providing contextual opportunities for misbehavior, it was expected that the later (Time-2) RAPs would be better predictors of Time-2 levels of delinquency. This was, in fact, the case. As shown in Column 3 of Table 2, four of the nine RAPs measured at Time-2 (the academic, religious, social and nonsocial patterns) significantly affected respondents’ levels of delinquency. As expected, given the structure and visibility of their component behaviors, the academic and religious patterns reduced, while the social pattern increased, participation in deviant activities. The inverse relationship between the nonsocial RAP and delinquency, on the other hand, was unanticipated, given the low structure and visibility associated with that activity cluster. Not surprisingly, the inclusion of the Time-2 RAPs in the statistical model ameliorated the effects of the earlier (Time-1) activity patterns. Including early (Time-1), as well as later (Time-2), RAPs in our analyses enabled us to assess the pathways through which gender influenced high school seniors’ levels of delinquency. A series of supplemental regressions (data not shown) indicated that gender had a number of indirect (but no direct) effects on three of the four Time-2 RAPs that were significantly related to Time-2 deviance (the academic, religious/community and nonsocial orientations). Through these paths gender was, at a minimum, three times removed from the outcome in question (e.g., females had higher scores than males on the Time-1 academic RAP, which increased their participation in these activities at Time 2, which subsequently reduced their risk for delinquency). Thus, the overall impact of the indirect effects of gender on Time-2 deviance was miniscule. The test for mediation described earlier (the comparison of coefficients for the variable female in the regressions with and without the Time-2 RAPs) confirmed this. As shown in Columns 2 and 3 of Table 2, there was very little change in the effect of gender on delinquency when the nine Time-2 RAPs were added into the analysis, suggesting that differential participation in routine activities did not explain the observed gender gap in deviant behavior. There was, however, some support for the moderation hypothesis. Although there was no evidence that gender moderated the effects of any of the Time-2 RAPs, on delinquency, as shown in Column 4 of Table 2, cross-product interactions between gender and the Time-1 religious/community RAP, and between gender and the Time-1 social RAP, were strong enough to reach statistical significance. In order to determine the direction of these effects, we used the procedure for interpreting interactions suggested by Ross, Mirowsky, and Huber (1983). Predicted deviance scores for males and for females were computed using the regression equation from Column 4 of Table 2. In each case, scores on the predictor of interest (the religious/community or the social RAP) were varied from one standard deviation below to one standard deviation above the sample mean, while holding all other variables constant at their mean score. Means and standard deviations on the model variables are presented in Appendix B. The effect of participation in religious and community activities on delinquency by gender is shown in Fig. 1. As indicated here, a two standard deviation increase on the religious/community RAP, measured during the sophomore year, decreased the risk for deviant behavior substantially among senior males (but not females) when current participation in religious and community activities (the Time-2 religious/community RAP) was held constant. Effects of Time-1 Religious/Community RAP on Delinquency. Fig. 1. Effects of Time-1 Religious/Community RAP on Delinquency. Figure options A similar pattern was observed when the nature of the interaction between gender, the social orientation, and delinquency was examined. Fig. 2 shows the estimated effects of a two standard deviation increase in the social RAP on deviant behavior by gender. While the social orientation had a minimal effect on delinquency among females, frequent participation in unstructured and unsupervised peer interactions during the sophomore year in high school substantially increased deviance among males two years later, when they were seniors, irrespective of their current levels of peer interaction. Effects of Time-1 Social RAP on Delinquency. Fig. 2. Effects of Time-1 Social RAP on Delinquency. Figure options Again, the Time-2 religious/community and social RAPs were strongly associated with delinquency, but these effects did not vary across gender. Overall, RAPs explained 2.7 percent of the variance in delinquency among the high school sample (Table 2).