“…This model, however, may be useful in characterizing certain psychosocial outcomes. For example, although they did not directly test for mediation, Schafer, Birchler, and Fals-Stewart (1994) demonstrated that low verbal and executive functioning of polysubstance abusers negatively affected interpersonal relations when anxiety and depression were controlled; yet when neurocognitive functioning was controlled, affective state was not predictive. That is, wives of polysubstance abusing husbands evidenced more frequent negative behaviors (e.g., criticizing, ignoring, interrupting) and less frequent positive behaviors (e.g., agreement, smiling) as their husbands' cognitive functioning decreased.…”