Students in three mathematics classes were assessed on two aptitudes, field independence and general reasoning, and randomly assigned to either an expository or a discovery treatment. The expository treatment used a deductive sequence of instruction and provided maximal guidance for the students. The discovery group used an inductive sequence with minimal guidance, and provided calculators to help students discover concepts and rules independently. The topic of instruction involved errors in measurement and calculations with approximate data. There was a significant interaction with general reasoning on the retention test, as predicted. There were no interactions with field independence. (Author/MP)