The present study examined mothers' use of corporal punishment and adolescents' aggression as predictors of mother–youth conflict during early adolescence. Particular attention was given to the potential mediating role that adolescents' hostile attributions of intent (HAI) regarding mothers' behavior might play in connections between corporal punishment, youth aggression, and mother–adolescent conflict for European American (EA) and African American (AA) youth. Data were collected from 268 12- to 14-year-olds (154 European American; 114 African American; 133 girls; 135 boys) and their mothers over a period of 2 years. Questionnaires completed by both mothers and adolescents were used to assess maternal corporal punishment and adolescent aggression, and interviews concerning hypothetical situations were used to assess adolescent HAI in year one. In both year one and year two mother–adolescent conflict was observed in a laboratory interaction session. Data revealed that adolescent HAI mediated the link between maternal corporal punishment and mother–adolescent conflict for EA, but not AA youth. Adolescents' HAI mediated the link between adolescent aggression and mother–adolescent conflict for both EA and AA families.
The transition to adolescence is marked by a small, but significant, increase in conflict between parents and children (Laursen et al., 1998 and Smetana et al., 2006). To some extent, these changes are considered to be a normative result of the cognitive, social, and physiological transformations that coincide with puberty (Conger and Ge, 1999 and Steinberg, 1987), as well as disturbances in self-concept development that occur during the transition to adolescence (Rosenberg, 1985). However, there are wide individual differences in the rate and intensity of parent–adolescent conflict across families (Smetana et al., 2006). Furthermore, parent–child conflict can take different forms (Smetana, 1996), with some forms falling into a category of developmentally appropriate disagreement surrounding issues of autonomy and independence (Adams & Laursen, 2007), and other forms of conflict characterized by atypical hostility and coercion (Conger & Ge, 1999). It is the more hostile and coercive form of conflict that past research associates with adolescents' emotional distress (Chung, Flook, & Fuligni, 2009), declines in academic performance (Dotterer, Hoffman, Crouter, & McHale, 2008), and problems in adult romantic relationships (Overbeek, Stattin, Vermulst, Ha, & Engles, 2007). Consequently, the focus of the current study is on understanding factors that are associated with parent–child hostile and coercive conflict during the transition to adolescence.
Adolescents' hostile attributions of mothers' intent
According to Social Information Processing (SIP) Theory (Crick and Dodge, 1994 and Dodge, 1980), one mechanism that may account for individual differences in parent–adolescent conflict is the cognitive attributions that adolescents make regarding parental behavior (Dodge, 2006). Attributions are naïve causal explanations that individuals form in order to better comprehend, predict and respond to their environment (Kelley, 1967). Formally labeled “hostile attribution bias,” such misattributions are viewed as being a SIP error in which youth interpret ambiguous situations as being hostile (Burks et al., 1999 and Weiss et al., 1992). It is also possible, however, that hostile attributions are not always in error or mis-attributions, but rather reflect accurate expectations of parent behavior (MacKinnon-Lewis, Lamb, Hattie, & Baradaran, 2001). When an adolescent is confronted with an ambiguous situation, a knowledge structure will be referenced to help the adolescent interpret and decide on the appropriate course of action. Adolescents who are prone to accessing hostile knowledge structures may be more likely to attribute hostile intent to the ambiguous situation, which in turn increases the likelihood that they will choose a hostile response.
To examine possible ethnic and gender differences on demographic characteristics, we performed a 2 (EA vs. AA) × 2 (male vs. female) MANOVA, with child age, mother age, family income and mother education as the dependent variables. There was no significant main effect of ethnic group, Wilks's λ = .78; F(2, 262) = 2.31, ns, or gender, Wilks's λ = .86; F(2, 262) = 1.73, ns. Nor was there a significant interaction between race and gender, Wilks's λ = .72; F(2, 262) = 2.57, ns. These results suggest that AA and EA families, as well as families with boys and girls, in the sample did not differ significantly on demographic characteristics.
The means and standard deviations of all the primary variables used in analyses for boys and girls of both ethnic groups are depicted in Table 1. Zero-order Pearson Product Moment correlations across ethnic groups are presented in Table 2 (correlations for the AA group are depicted above the diagonal). Gender, a categorical variable, was included in the correlation analysis to provide an indication of whether girls or boys scored higher on a particular variable. Prior to testing the hypothesized models a thorough graphical inspection of the data were completed, with screening for outliers and evaluation of traditional convenience assumptions such as normality and linearity of relations. No outliers were found for any variables, however, four variables: mother report of corporal punishment at T1, adolescent report of corporal punishment at T1, mother report of adolescent aggression at T1, and adolescent self-report of aggression at T1 (all were positively skewed) were found to exceed the cutoffs suggested by West, Finch, and Curran (1995). Logarithmic Transformations were carried out, after which all distributions approached normality (maximum skew = 1.6 and maximum kurtosis = 1.7).
