This study focuses on the effectiveness of "Solutions Unlimited," a computer software program developed to teach problem solving to middle level students. Fifty-one fourth, fifth, and sixth graders (21 girls and 30 boys) attending a private school in a small mid-west community were the subjects for this experiment; none had received prior training in problem solving. For a pretest, all students completed the Purdue Elementary Problem Solving Inventory (PEPSI), subscales of which include such things as sensing that a problem exists, judging information, analyzing details, solving problems, and verifying solutions. Subjects in the treatment group then worked in groups of three on one lesson of "Solutions Unlimited" each week for eight weeks. The control group participated in a creative dramatics group during the same period, and was given an opportunity to use the computer software following the study. Results of the study did not indicate any significance in the difference scores of the treatment group; both groups generally achieved similar change scores. There is speculation that the design of the PEPSI was not appropriate for the age group involved, and also that problem solving may need to be integrated into the entire curriculum rather than being taught for 45 minutes once a week. The strength of the computer program used is in its simulations, which create realistic life situations. It is suggested that follow-up activities should be used to stimulate the possibility of transfer of problem-solving skills learned from the computer simulations to new and novel situations. (7 references) (CGD)