Performance in concrete and abstract tasks is examined systematically by varying the degree of abstractness of problem-solving and concept formation tasks. Four forms of a problem solving test were constructed. Each form of the test presented problem situations through four different modes: verbal stories, picture-book, color slides, three-dimension models. Advantaged and disadvantaged children from grades 2 and 4 were randomly assigned to test modes. Similar arrangements were made for testing concept formation. Stimulus material for the concept formation tests were presented via three modes: paper and pencil, motion picture film and actual objects. The degree of concreteness in the mode of presentation does affect the performance of children on the tasks. For only one of the the tasks, however, did the socioeconomic factor exhibit systematic relationships with the factor of concreteness. On the problem-solving task, all children performed best on the more concrete forms. Advantaged children out performed disadvantaged on all forms. On the concept formation tasks, however, the disadvantaged children out performed advantaged children on the most concrete test form. Additional factors which may have influenced test performance are discussed. A substantial reference listing, concept learning tests, and problem-solving tests are included. (Author/RC)