Rehearsal of associations stored in long-term memory.

A 2 by 2 by 2 design in Experiment I allowed comparison of rehearsers and controls assigned to one of four types of paired associate lists differing in imagery. After reaching a partial learning criterion, rehearsers engaged in covert rehearsal of associations drawn from long-term memory. Rehearsers were superior to the controls on posttreatment trials, but higher imagery was not associated with greater rehearsal gains. In Experiment II, postrehearsal success was strongly related to the relative frequencies of overt rehearsal. Rehearsals were more frequent on known associations, but rehearsal frequency also was related to new associative gain. Overall, the results support the contention that effective rehearsal may be undertaken on associations retrieved from long-term memory.