No interaction effects statistically significant at the 0.01 level were found between maternal and paternal alcohol use or between alcohol use and parental mental distress or the child's gender, age, social network or mental distress. An interaction effect was found between paternal alcohol use and relationship dissolution on attention (Wald Type III = 14.766; p = 0.005). Children of alcohol abusing fathers with dissolved relationships had higher levels of attention problems than expected from the totalled main effects of abuse and relationship dissolution (additional effect: d = 0.57, C.I. 0.15 - 0.99, p = 0.007). The effect of seeing parents drunk varied with paternal alcohol use category on attention (Wald Type III = 181.93; p < 0.001) and conduct problems (Wald Type III = 81.41; p < 0.001). Post hoc tests included too many group combinations to provide meaningful results, but seeing parents drunk tended to be more predictive of these problems if the father did not participate or if the father was in the at risk group.
