Figure 1. Unstandardized factor loadings and parameter estimates for Blacks and Whites with partial invariance. click this Separate parameters for Black and White groups indicate significant differences in multigroup CFA and structural equation modeling analyses (Black/White). … Discussion We examined two possible mechanisms for the influence of race on smoking in the 10th grade: (a) differences in the mean levels of risk and protective factors and (b) differences in vulnerability to risk and protective processes. Black teens were more likely than White teens to have parents who smoke, which should place Black teens at greater risk. However, Black parents reported higher levels of guidelines against substance use, which could protect Black youth from the negative influences of risk factors (Hill et al.

, 2005). This race difference is consistent with the finding of focus groups in which Black parents reported stronger feelings of efficacy (Clark et al., 1999). However, teen reports of family management and positive bonding did not differ by race nor did levels of deviant peer associations. We found measurement differences in the latent variable for parenting but not in the model relating risk and protective factors to teen smoking. Among the Black families, parent reports of guidelines were correlated with teen reports of attachment, monitoring, and discipline, forming a coherent construct reflecting positive family influences hypothesized to reduce the risk of smoking. Among White families, the parenting construct relied completely on the teen reports.

The parent reports of guidelines did not load on the parenting factor. We found no race differences in the magnitude of relationships between risk and protective factors and smoking. Across both race groups, we found positive family influences reduced deviant peers, as expected. The hypothesized mediated process by which parenting reduced association with deviant peers, which in turn reduced the likelihood of teen smoking, was confirmed. We found no evidence of direct effects of income or parenting in eighth grade on smoking in 10th grade. The effect of family income on teen smoking was mediated through parent smoking such that lower income predicted parent current smoking status, which in turn increased the likelihood of teen smoking. Family income was positively related to parenting, which in turn decreased association with deviant peers in both groups.

The strengths of the present study include parent and teen reports, a longitudinal design, and a sample with equal numbers of Black and White families. However, we failed to gather data on the smoking status of the teens�� peers (Unger et al., 2001). Despite this omission, we still observed the expected relationship between deviant peers Batimastat and smoking a year later. We also failed to ask questions of the teens and parents that dealt more specifically with efforts the parents make to discourage their teens from smoking.