This study investigates the impact of family configuration and parent education-income level on parental beliefs, the relationship between these beliefs and actual parental practices, and the effect of parental practices on children's problem-solving abilities. One hundred twenty intact families participated in the study. Forty families consisted of parents and an only child aged 3 1/2 - 4 1/2 years and 80 were three-child families with a middle child aged 3 1/2 - 4 1/2 years. In the latter group, half of the families had fewer than three years' spacing between oldest and middle children and half had greater than three years' spacing. Half of the families were working class, half were middle class. Parent questionnaires and interviews and seven tasks for assessing the child's representational abilities and problem-solving competence were employed. Results implicate social class and family configuration factors in the development of children's problem-solving abilities. Children from different family configurations demonstrate different strengths and weaknesses relative to other children, depending on the problem in question. At times, the results obtained contradict common stereotypes associated with family characteristics. For instance, only children tended to generate passive strategies less often than children with siblings. Passive approaches to problems and memory abilities differentiate the three family configuration groups. (Author/RH)