Retrieval practice effect (RPE) means that compared with restudy in the same amount of time, one or more tests (retrieval practice) of a learning content can effectively help people achieve a better performance in the final memory test. The typical RPE experiment includes four events: encoding, initial testing (retrieving), distraction task and final test. In most RPE related study, items actively retrieved from memory show the higher probability of being able to recall than items with additional study, even when retrieval practices are given without feedback. Early researchers believed that additional exposure or over learning (encoding) of the testing group made them outperform the restudy conditions. Another well-documented RPE account argued that, as the process of accessing information stored in long-term memory, retrieval is a critical variable for substantive long-term retention of memory. Only fairly recently, based on several influential experimental findings, researchers proposed that RPE is a combined effect of encoding and retrieval, rather than one single encoding or retrieval effect. Clarifying how encoding/retrieval processes and their possible interactions change the level of RPE will help to make a further understanding of it. So far, although theories about RPE abound, current researches shed light on this question are sparse.

The authors noticed a study conducted by Carpenter (2009) reported that although strong cues (e.g., Toast: Bread: facilitated recall of tested items initially, items recalled from weak cues (e.g., Basket: Bread) were retained better over time, such that this advantage was eliminated or reversed at the time of the final test. However, this reversal effect is not stably found across studies. The authors assumed that the entangled encoding and retrieval duration and their possible interaction in different researches might be the critical factors. To understand the inconsistencies in previous studies, it is necessary to clarify the changes in RPE under different encoding and retrieval duration conditions. Therefore, the present study will further investigate the impact of encoding and retrieval duration on RPE through two experiments.

Experiment 1 is a 2 (encoding duration: 3 seconds, 10 seconds) × 2 (retrieval duration: 8 seconds, 13 seconds)between-participants experiment design. 121 undergraduates learned 40 word pairs(lounge-couch). When participants learned all the pairs of words (encoded) , they retrieved the target words with the clue words. And after a 5-minute distraction task, they took the final test. The final memory test scores and retention rates were recorded. The results show that under the immediate test, extending encoding and retrieval time is beneficial to improve both memory scores and memory retention rate. Experiment 2 took a similar design, except that the final test was extended to 24 hours to examine the effect of encoding and retrieval duration on RPE in a delay test. We found that the memory advantage of the long-term encoding under the delay test disappeared with the extension of the test interval, while prolonged the retrieval time can effectively improve memory scores and memory retention rate. These data suggest that the effect of encoding duration on the RPE depends on the test interval; Retrieval duration is the key influencing factor of RPE. We discussed this under the framework of retrieval efforts, elaborative retrieval account, and episodic context account.