Selective retrieval practice can benefit long-term retention of untested associations: Evidence from a word triplet paradigm

It is now well established that LTM retention can be improved through methods ofoptimising learning and testing schedules that show promise in laboratory and applied contexts. Among these principles, there is a large and growing literature on retrieval practice (also known as test-enhanced learning or the testing effect), whereby later memory performance is improved when the individual carries out retrieval of material during learning and/or retention. One broad outstanding question concerns the task contexts in which retrieval enhances later memory and retention, and whether such effects are limited to the precise material being practiced, or if they generalise or ‘transfer’ to associated information. The current study reports two experiments examining this question, with a focus on whether retrieval can enhance associative connections within multi-feature memory representations over extended time periods of several days. We tested retention of associative links within word triplets shortly after encoding and after a one-week delay. Half of the participants also performed an additional intervening cued recall test 24 hours after encoding, in which a different associative link from each triplet was cued for retrieval. In both experiments, participants in the intervening group showed relatively better retention on the final one-week test. These findings indicate that retrieval benefits can transfer from tested to untested associations within word triplets when retrieval practice and testing is implemented over periods of several days.