The present study examined psychopathology in high-risk children (ages 8 to 18) from families with a multigenerational history of alcoholism and contrasted them with low-risk children from community control families. Similar rates of childhood disorders were found for the high- and low-risk groups whether or not the children lived with an alcoholic parent. These findings suggest that the increased psychopathology commonly reported for children of alcoholics arises from comorbidity within the extended family as a result of assortative mating. When comorbidity is reduced through the selection of families with only alcoholism, a different symptom picture emerges.