Gender, Warmth, Rejection, and Violence
Many authors have posited that the strength and importance of interpersonal relationships may vary by gender. There are extensive evolutionary discussions that predict that this is so (e.g., Savage & Kanazawa, 2004). Studies of social support and other measures of social relationships, disaggregating boys and girls, often uncover disparities in effects and effect sizes. For example, Rubin et al. report a significant negative correlation between maternal warmth and externalizing for girls but not boys in a Canadian sample of toddlers (Rubin, Hastings, Chen, Stewart, & McNichol, 1998). Based on this assumption, we might hypothesize that one reason for gender differences in antisocial behavior is that females have stronger social bonds or that social bonds exert a greater impact on females than males (discussed earlier this chapter).
To assess this possibility, the findings of studies reporting data for males and females were separately examined. The findings on associations between both warmth and rejection and violent behavior were mixed for both males and females, but a gross analysis suggests that if either sex is more adversely affected by low parental warmth or outright rejection, it would be the boys. For females, coefficients were in the expected direction in about 2 out of 4 studies of warmth, but in no study was the PoC statistically significant (Li, 2007; Ribeaud & Eisner, 2010; Scholte et al., 2007; Shekarkhar & Gibson, 2011). For parental rejection, we find only 3 studies, and the findings just barely lean in support of an association between rejection and violence for girls, so strong conclusions cannot be drawn until more research has been produced.
There were more studies of males than females, so we can get a better understanding of the state of the research from this body of work. Studies using “warmth” had mixed findings but weigh slightly in the direction of an inverse association between parental warmth and violence in children. Approximately 6 out of 9 studies report associations in the expected direction, and 4 of these report a statistically significant PoC (Loeber & Dishion, 1984; McCord, 1996; Tupin et al., 1973; Walsh, Beyer, & Petee, 1987). Loeber and Dishion (1984) found significant associations between both maternal and paternal warmth and male violence, but they used only simple comparisons and no controls for confounding factors. Similarly, McCord (1996) reports significant simple associations between warmth and violence in her male sample. For Ribeaud and Eisner (2010), parental warmth and violence were correlated in the right direction but were not statistically significant for males. Several studies report findings that contradict an association between warmth and violence among males (Li, 2007; Scholte et al., 2007; Shekarkhar & Gibson, 2011). If the association between warmth and violence among males were a strong one, we would expect more consistent findings to have emerged by now. It is possible that measurement problems or some other methodological obstacle are responsible for the inconsistency.
Findings on rejection and violence among male subjects are also mixed but we see an emerging pattern. Because 3 out of 5 studies report a PoC that is in the expected direction (Akse et al., 2004; Dutton et al., 1996; Shekarkhar & Gibson, 2011), and the remaining 2 are not strongly in the opposing direction (Brennan et al., 1999; Huesmann et al., 2006), and in consideration of the quality of the methodologies used, we lean toward concluding that the association holds for males. In addition, studies comparing violent to nonviolent offenders were in the predicted direction.
The weight of evidence regarding warmth and rejection is weaker when we look at studies of males and females separately than it is when we look at combined samples. The emerging data suggest that the effects of parental warmth and rejection on violent aggression are more consistent for males than females.