Figure 1.

Experimental procedure. During the study phase (“encoding”), participants saw word pairs under deep or shallow processing tasks. During the subsequent test phase (“retrieval”), one word of the previously presented pairs was shown, intermixed with unstudied new words (“lures”). Participants indicated with one button press whether they thought the given word was new, the word was old but they did not remember the paired associate or the word was old and they recalled the paired associate. In the latter case, a second screen appeared to validate recall accuracy, providing three first-last letter combinations of which one corresponded to the target association. Analyses focused on correct identification of lures (CR), correct identification of old words without recalling the paired associate (IR) and correct identification of old words along with correctly recalling the paired associate (AR).