Violent Socialization and Current Life Circumstances

Model 10 displays the results of combining all level-1 variables with violent socialization. Social desirability was included as a control in this model as a control. The results show the effects of violent socialization on youth criminal behavior remain constant: violent socialization continues to be significantly associated with increases in youth criminal behavior scores even when controlling for other theoretical constructs, both in the community (Y₀|) ^an^ within families (Yiq)- For each 1 point increase in the violent socialization in the community scale, mean youth criminal behavior scores tend to significantly increase, by an estimated 0.11 points for each 1 increase in the violent socialization in the community scale (Y_ol = 0.113, p<0.001) on average, net of violent familial socialization and other controls. For each 1 point increase in the violent familial socialization scale, mean youth criminal behavior scores tend to significantly increase, by an estimated 0.11 points (Y_lo= 0.153, p< 0.001) on average, net of violent socialization in the community and other controls. The predicted mean youth criminal behavior score for the prototypical male respondent of average age experiencing violent socialization, but has average levels of attachments, self-control, socioeconomic status, and socially desirable responses whose peers engage in criminal behavior, is approximately 6.1 points on average.³

These results offer further support for the first and second hypotheses. In the nations where violent socialization processes are used, both in the community and within the family, youth in that nation tend to engage in more criminal behavior (Hj and H,). This finding holds true even in the presence of effective social control markers. This finding holds true regardless of other theoretical premises used to explain youth criminal behavior. Violent socialization may just be part of the life circumstances for some youth. This contention is further supported by the finding that youth criminal behavior is variable. Estimated youth criminal behavior scores within nations continue to significantly vary (o²= 2.763, p < .001) and differences in youth criminal behavior scores across different nations continue to be significant (T_O(1 = 0.106, p<.001, net controls. The results also offer continued support for the third hypothesis. Youth criminal behavior significantly varies from nation to nation as a function of violent familial socialization processes (T_ol= 0.013, p < .01). But Model 10 does not fully embrace the norms of violence thesis. Families are nested within nations. Some nations have structural norms supporting violence. These structural norms must be accounting in assessing the norms of violence thesis. The intercept variance (TM rejecting the null hypothesis indicates there is still additional “explainable” variation to account for youth criminal behavior.

Violent Socialization Transnationally

Examining national-level averages does not provide nation specific-information to examine the details of variation. The analysis needs to include each individual nation. In order to determine if violent socialization processes predict criminal behavior for respondents within each nation, separate regression models were run for each nation using OLS regression. By separating respondents by nation, the results can be compared to show the specific differences, across nations, in the relationship between violent socialization processes and respondents’ individual criminal behavior.

The results of the OLS regression are shown in Table 5.4. By separating respondents by nation, the results show the relationship between violent

Table 5.4 Criminal Behavior Regressed on Violent Socialization. Attachment, and Self-Control for Each Nation in the IDVS

• ***p<0.00l **p<0.0l *p<0.05 ~p
• (Continued)
• 88 Analyzing Norms of Violence

Analyzing Norms of Violence 89

Table 5.4. (continued)

• ***p<0.00l **p<0.0l * p<0.0S ~ p
• (Continued)
• 90 Analyzing Norms of Violence

TNZ(n = 208)

SAF (n= 124).

VNZ (n = 249)

• *** p<0.001
• ** p<0.0l

p<0.05

— p <0.10

Analyzing Norms of Violencesocialization processes and respondents’ individual criminal behavior varies trans-nationally. There are respondents in some of the nations in which their criminal behavior as a youth is not significantly associated with violent socialization processes, either in the community or within their individual family. These nations include Guatemala (GTM), India (IND), South Africa (SAF), Tanzania (TAN), and Venezuela (VZA). For the other nations, respondents’ experiences with violent socialization processes significantly predict youth criminal behavior.

There was only one nation in which only community-based violent socialization processes are significantly associated with youth’s criminal behavior: England (ENG).⁴ Looking at England, increases in violent community socialization are significantly associated with increases in individual criminal behavior scores. With each one-unit increase in violent community socialization, youth criminal behavior scores increase by about 0.19 points (b₍ =0.194, p<0.001). These increases occur even in the presence of the protective factor of selfcontrol b₄ = -0.151, p<0.05).’ Violent socialization within the communities in England significantly explains about 27% of the variance in youth criminal behavior (R² = 0.267, p< 0.001), net of controls. The predicted youth criminal behavior score for the prototypical respondent who experienced violent community socialization is approximately 9.57 points (bj + b^-l-constant). This is about 4 points higher than the mean youth criminal score across all nations. So, the respondents who experienced violent community socialization tended to report more criminal behavior compared to respondents who did not experience violent community socialization processes.

In the majority of nations, violent familial socialization alone was significantly associated with youth’s criminal behavior. The results show that respondents who experienced more violent familial socialization processes tended to report having significantly more involvement in criminal behavior prior to age 15. These countries include Brazil (BRZ), China (CHN), Germany (GER), Greece (GRC), Israel (ISR), Lithuania (LTH), Mexico (MEX), New Zealand (NZL), Portugal (PGL), Romania (ROM), Russia (RSS), S. Korea (SKR), Switzerland (SWZ), and Taiwan (TWN).¹’ For example, respondents in New Zealand (NZL) who experienced violent familial socialization processes tended to have the highest predicted youth criminal behavior scores, net of controls, at 11.11 compared to youth in other nations (b₍+ b, + b₄+constant). This score is more than double the mean youth criminal score across all nations (see Table 4.3 for descriptive statistics). In New Zealand, self-control offered a slightly protective effect on youth criminal behavior (b₄ =-0.208, p<0.05) whereas attachments within the family had no statistically significant effect. The lowest youth criminal behaviors scores among those who experienced violent familial socialization are in Portugal. The predicted youth criminal behavior score was only slightly above the mean, at about 6.3 points, net of controls (b,+ b, + b₄+constant). In New Zealand, attachments within the family had a slight protective effect on youth criminal behavior in Portugal (b. = -0.112, p < 0.01) whereas self-control has no statistically significant effect.

There were several nations in which both violent socialization within the community and violent familial socialization were significantly associated with youth criminal behavior. These nations are Australia (AST), Belgium (BEL), Canada (CND), The Netherlands (NTL), Scotland (SCT), Singapore (SNG), Sweden (SWD), and the United States (USA). Again, comparing the highest levels of youth criminal behavior to the lowest levels, respondents in the United States who experienced violent socialization processes tended to have the highest predicted youth criminal behavior scores, net of controls, with a prototypical score of 10.16 compared to youth in other nations (bj + b^ + ^ + b^ constant). This is taking into account the protective factors of attachments (b, = -0.116, p<0.001) and self-control (b₄ = -0.218, p<0.001), which were both significantly associated with less youth criminal behavior. This score is still more than double the mean youth criminal score across all nations (see Table 4.3 for descriptive statistics). The lowest youth criminal behavior scores among those who experienced violent socialization are in Sweden. The predicted youth criminal behavior score was only slightly above the mean, at about 6.3 points, net of controls (b₁ + b₂+b, + b₄ +constant). This includes the protective factor of self-control (b₄=-0.114, p< 0.001).'

These results continue to provide support for the norms of violence thesis. To some degree, violent socialization processes are associated with youth criminal behavior across most nations. The effect of violent socialization on youth criminal behavior varies from nation to nation.

In 22 nations, respondents’ experiences with violent socialization processes were significantly associated with their criminal behavior as youth. Further analyses were conducted to look for specific patterns: group analysis by region.

Using the regional categorizations from the IDVS (see http://pubpages.unh. edu/~mas2/Nationsl.htm), the results seem to reflect similarities by region with few exceptions. In the African nations, violent socialization processes were not significantly associated with respondents’ criminal behavior during youth. In the Asian nations, violent socialization processes were significantly associated with respondents’ criminal behavior in all nations but Japan and India. In all the European nations, violent socialization processes were significantly associated with respondents’ criminal behavior during youth.⁸ In both Australia and New Zealand, violent socialization processes were significantly associated with respondents’ criminal behavior during youth. This is also true for the North American nations. It is only the Latin American nations that produced strong mixed results: in half the nations, violent socialization processes were significantly associated with respondents’ criminal behavior during youth. It is worth noting that in the Middle Eastern nations (i.e., Iran and Israel), the overall regression models were not statistically significant and therefore cannot be included in this OLS regression analysis. Unfortunately, there tend to be disparities between the Middle Eastern nations and the other nations in the IDVS. For example, the two Middle Eastern nations (Iran and Israel) have a religionspecific context to the nation’s structure. Thus, further research is necessary to discover what societal influences may be present in the Middle Eastern nations to explain youth criminal behavior.

These geographical-regional results partially support the first and second hypothesis: youth from families that use violent socialization processes will engage in more criminal behavior, compared to youth from families that use non-violent familial socialization processes, even in the presence of effective familial socialization processes (H ), and youth in communities that use violent socialization processes will engage in more criminal behavior, compared to youth in communities that do not use violent socialization processes (H,). This analysis was useful to support the need for multilevel modeling regression analysis. Multilevel modeling regression analysis can determine if there are significant variations in youth criminal behavior scores across different nations and if violent socialization processes could account for this variation.