Two studies were undertaken within the framework of learning set or learning-how-to-learn methodology, with comparable sets of stimuli which varied from trial to trial rather than remaining constant. Two randomly assigned groups of tenth grade students used a different four-word set of difficult vocabulary items each day for 15 days. Time taken for the exercise and an achievement score provided two dependent variables. Two instructional methods were used: matching words to synonyms, and reading and rewriting definitions. In the first study students worked with one method for ten days, switched to the other for three and returned to the original for two. In the second study, students used the same method throughout. Results, illustrated by graphs, showed that students performed better in terms of achievement and time as practice continued. The synonym method appeared to be the more interesting, though not necessarily the better, of the two and provided immediate feedback to the student. Study 1 proved to be biased against the definition group, and the achievement test was changed in Study 2 to a sentence completion approach to be fair to both groups. The exploration of a wider variety of criteria is recommended for future studies. (MBM)