Most research on scaffolded instruction has been conducted with mothers rather than fathers and has carried the implicit assumption that it is superior to other methods. This study examined fathers' teaching style differences and the effects of those differences on children's recall and transfer of a strategy for solving balance scale problems. Participating were 11 father-daughter and 9 father-son dyads. Children ranged in age from 9 to 12 years. Two balance scales were used during instruction, one with four pegs to each side of the fulcrum, and one with 10 pegs to each side. During a laboratory session, fathers completed a questionnaire regarding his aspirations for his child and his involvement in his child's life. Prior to the teaching phase, fathers' understanding of the mathematical rule for solving balance problems was verified. The teaching interaction was videotaped. Following the interaction, children received a posttest of 4-peg problems to test strategy maintenance and were also posttested for transfer on a 10-peg scale and for far transfer using problems presented on paper. Teaching style was defined by the change over time in the proportion of words spoken by the father and child. The findings indicated that fathers were highly variable in the amount of time they took to teach the strategy, as well as in their teaching style. All fathers were equally effective in teaching the strategy to their children. A high correlation was found between frequency of help with homework and efficiency of teaching. (Contains 22 references.) (KB)