The aspect of the Unified Science and Mathematics for Elementary Schools (USMES) project described in this paper was undertaken in an effort to observe the problem solving behaviors of elementary school children. The Notebook Problem consisted of presenting a student with three notebooks arranged so as to differ from each other in terms of such dimensions as number of pages, number of lines per page, binding, price, etc.; the subject was asked to (1) select the most appropriate one for his class, and (2) indicate the reasons for his selection. Pretests and posttests were administered to randomly selected students from both control and USMES project groups. Scoring of responses was performed along the following lines: (1) whether any of the subject's reasons for selection were stated in measurable quantities, and (2) the highest level of warrant associated with the reasons stated. Representatives of the dimensions measurable were: (1) size-volume, (2) weight, (3) cost, etc.; while levels of warrant were determined by responses being: (1) personal opinion, (2) testable, or (3) had been tested. Chi-square analysis revealed significant improvement in pretest to posttest scores for the experimental group versus the control group. (Author/CP)