The testing effect-more learning by testing as compared to restudying-is a well-established finding. A typical testing procedure in the context of meaningful learning comprises a recall task after an initial study phase. Different theories refer to different mechanisms when explaining the positive effects of such recall tasks. In the context of learning from expository texts, we tested three mechanisms as suggested by a variety of prominent approaches: the elaborative-retrieval theory, the theory of transfer-appropriate processing, and the unspecific-goal perspective. We experimentally varied the type of testing task (short-answer task vs. free-recall task, both compared to a restudy task) in a within-subject design (N = 47 university students). We replicated the testing effect. We found no evidence for a transfer-appropriate processing effect or an unspecific-goal effect. The testing effect disappeared when statistically controlling for mental effort. Initially non-tested material was also fostered by testing (spreading activation effect). These findings indicate that testing helps learning when learners must invest substantial mental effort, as suggested by the elaborative retrieval theory. For educational purposes, testing tasks should be assigned that require the learners to invest substantial mental effort.
Keywords: elaborative retrieval; meaningful learning; mental effort; retrieval practice; testing effect.