Abstract
Numerous studies have demonstrated that retrieval from long term memory (LTM) can enhance subsequent access to that memory, a phenomenon labeled the testing effect (Carrier & Pashler, 1992). However, the testing effect has been primarily been studied with verbal/categorical stimuli. Consequently, it has not yet been determined whether the testing effect impacts the probability of successful memory retrieval or the quality of the retrieved representation. To answer this question, we examined LTM for the color of 400 unique images. After every 10 images subjects (n=20) either recalled each of the ten colors or were given a chance to restudy the items. After a 30 minute delay, subjects selected the color of each of the previously learned images from 360 degrees of continuous color space. Using a mixture model analysis (Zhang and Luck, 2008) we determined the probability and quality of the recalled color. A robust testing effect was observed; subjects recalled a significantly higher proportion of items that they had previously retrieved relative to items that they had restudied (p <.01). Interestingly, this boost in storage probability was accompanied by a decrease in mnemonic precision such that mnemonic precision was significantly worse for items subjects were asked to retrieve from memory relative to restudy (p <.05). This decline in precision appears to be at least partially explained by a tendency for delayed recall responses to be biased towards the (imperfect) responses that were made in the immediate recall portion of the retrieval condition (r=.26), a tendency that was absent in the restudy condition (r=.03). Thus, the testing effect enhances the probability of successful recall but it comes with a cost; the learned associations are shaped by the imperfect representations that are formed during retrieval practice.
Meeting abstract presented at VSS 2014