59
Convergent Validity
Several hypotheses predicted that DPICS II variables would correlate with the scores on the measures associated with child behavior problems. Because the conductdisordered children were over-represented in this study's sample (i.e., 50%), the groups were weighted to reflect the estimated prevalence of conduct problem behavior in the general population of preschool-aged children (i.e., 16%; Campbell, 1990). The created sample was normally distributed. The results of the analyses of the specific hypotheses are described below and are presented in Tables 6 and 7.
Hypothesis 3 predicted that the DPICS II variables would correlate significantly with the ECBI Intensity score. The ECBI Intensity score was negatively correlated with child Compliance, r (60) = -.42, p < .01, positively correlated with child Inappropriate Behavior, r (60) = .40, p12 < .01, and positively correlated with father Inappropriate Behavior, r (60) = .46, p <.01. The ECBI Intensity score, however, was not significantly associated with child Prosocial Behavior (with or without Information Descriptions), father Direct Command Ratio, or father Total Commands.
In Hypothesis 4, it was predicted that the DPICS II categories and summary
variables for child behavior (i.e., Compliance, Inappropriate Behavior, Prosocial Behavior) would correlate with the Child Domain score on the PSI. Child Compliance, r (60) = -.42, p = .001, and child Inappropriate Behavior, r (60) = .39, p = .01, were significantly correlated with the Child Domain score on the PSI. Child Prosocial Behavior