The time course of repetition effects for words and unfamiliar faces.

The repetition effect on reaction time to words and unfamiliar faces was examined at lags of 0, 4, and 15 items between first and second presentations. For words, subjects made either a lexical decision or a decision based on the stimulus's structural attributes. In the lexical decision task, a significant repetition effect was found at all three lags for words, whereas for nonwords the effect was significant only at Lag 0. In the structural decision task, the repeated decision was facilitated for both words and nonwords only at Lag 0, despite a word superiority effect at all lags. Target faces were presented either zero, one, or five times before testing. Subjects made either structural discriminations (face/nonface) or recognition judgments. In the structural discrimination task, the effect of repetition was significant only at Lag 0 (regardless of the number of pretest presentations). In the recognition task, the repetition effect was longer lasting, and its magnitude increased with the number of presentations which, presumably, determined the strength of the episodic memory trace. These results are taken as showing that repetition effects, like other measures of memory, are influenced by the type of stimulus, its preexperimental history, the level to which it is processed, and the lag between the initial presentation and the test. The manner in which these variables affect performance, however, may differ across memory tests. The dissociations between performance among repetition tests, and between repetition tests and other types of memory tests, is interpreted according to a task-specific, component-process approach to memory.