This study-was designed to test the hypothesis that descriptive learning cycles are neither sufficient to stimulate students to reason at a formal operational level nor to encourage facility with the processes of scientific investigation. A 6-week long, three-investigation unit on simple machines drawn from a ninth-grade physical science curriculum was selected for the study. Students in the course were assigned to one of three instructional groups: descriptive group (DE), question design group (QD), and hypothesis testing group (HT). Each group completed identical exploration and invention activities. Each group participated in qualitatively distinct activities during the expansion phase. The DE students completed the activities outlined in the curriculum (a descriptive learning cycle). The QD group designed and conducted experiments to answer a question posed by the teacher. The HT group generated hypotheses concerning a problem, then designed and conducted experiments to test those hypotheses (a hypothetico-deductive expansion). The effects of the treatments were assessed in a pretest–posttest format using Lawson's Seven Logic-Tasks, the Test of Integrated Process Skills, and Lawson's Revised Classroom Test of Scientific Reasoning. Analyses of the data indicated that the HT group exhibited a significant increase on the Test of Integrated Process Skills and on Task 1 of the Seven Logic Tasks during the 6-week period.