Retrieval practice based on recognition memory: testing the retrieval effort hypothesis

We tested a core prediction of the retrieval effort hypothesis as an account for the testing effect (TE). Retrieval effort predicts that automatic, effortful retrieval should not lead to TE. Experiment 1 (N=76) showed that despite an encoding duration of three times less, object pictures retention is better after repeated testing under Old/New recognition conditions, than following repeated study. Experiment 2 (N=30) used a speeded and accuracy boosting procedure to rule out the contribution of recollection to retrieval. Retention of object pictures at 25 minutes and 6 months was similar after repeated testing or after repeated studying. These results call for a revision of the retrieval effort hypothesis as a mechanistic account of TE.