Date of Award




Document Type

Master's Thesis

Degree Name

Master of Arts (MA)


Department of Psychology


Cognitive Psychology

Content Description

1 online resource (ii, 38 pages) : illustrations.

Dissertation/Thesis Chair

James H Neely

Committee Members

Heather Sheridan


mediator effectiveness hypothesis, paired-associate learning, testing effect, Paired-association learning, Learning, Psychology of, Memory, Association tests

Subject Categories

Cognitive Psychology


The current experiments examined Pyc and Rawson’s (2010) mediator effectiveness hypothesis (MEH) as one potential explanation of the beneficial effects of testing relative to restudying in paired-associate learning. In two experiments, participants learned cue-mediator-target triads (e.g., lion-tiger-stripes) broken down into separate components. Specifically, participants learned and then restudied or were tested on cue-mediator (List 1) and mediator-target (List 2) word pairs in Session 1, and in a Session 2 given 48 hours later learned cue-target (List 3) word pairs. According to Pyc and Rawson, compared to restudying, testing on List 1 cue-mediator and List 2 mediator-target word pairs should lead to better List 3 cue-target recall due to testing strengthening both links of the cue-mediator-target chain. Experiment 1 results indicated that participants who were tested on List 2 mediator-target word pairs performed better on List 3 cue-target recall; however, target recall was the same whether or not participants could successfully recall the List 1 mediator, suggesting little support for the MEH. Experiment 2 explored the possibility that the review test on the List 2 mediator-target word pairs increased List 3 cue-target recall by increasing the accessibility of the target. Along with the same mediation triads used in Experiment 1, Experiment 2 included nonmediation triads. Consistent with Experiment 1, the results did not support the MEH. Overall, when the cue-mediator and mediator-target links were learned separately, the current experiments do not provide support for spontaneous mediation with weakly related English word pairs for which mediation should be quite useful.