Four experiments were conducted to examine the effects of various processing instructions on the rate of false recognition. The continuous single-item procedure was used, and false recognitions of four types were examined: synonyms, antonyms, nonsemantic associates, and homonyms. The instructions encouraged subjects to think of associates, usages (features), images, or rhymes. The only differences between the experiments were presentation rate and specific instructions included. Only the fourth experiment produced a statistically signficant instructions effect, with features, images, and rhymes instructions reducing false recognitions, of all types, and associates producing slightly more. The inconsistency of the effects over the four studies, however, suggests that a cautious interpretation is warranted. (Author)