This investigation examined strategic and semantic aspects of the answers given by preschool children to class inclusion problems. The Piagetian logical formalism for class inclusion was contrasted with a new, problem processing formalism in three experiments. In experiment 1, it was found that 48 nursery school subjects nearly always performed better on percept inclusion than on concept inclusion. This result supports problem processing formalism and contradicts logical formalism. Experiment 2 used 11 of the same subjects to investigate three questions: whether the children's counting strategies would produce the same response patterns as in experiment 1, whether the answer "the same number" (essential to any correct coextensive comparison) was available in their response repertoire, and whether expected responses to coextensive problems in concept and in percept sets would be obtained. Results offered consistent experimental support for SCAN and MATCH components of the problem processing model. Experiment 3 utilized 48 new subjects and a design which crossed four categories with four problem types, to clarify the reasons for the difference observed between the difficulty of percept and concept problems. Interpretations of the results are discussed in terms of the children's semantic strategies and counting strategies. The general conclusion offered is that problem-solving strategies, not logical deficits, are the source of young children's inclusion of errors. (GO)