How to activate students' natural desire to test themselves.

Testing oneself (i.e., doing retrieval practice) is an effective way to study. We attempted to make learners choose to test themselves more often. In Experiment 1, participants were asked how they wanted to study and were given four options: retrieval with no hint (e.g., idea: ______), a two-letter hint (e.g., idea: s____r), a four-letter hint (e.g., idea: se__er), or a presentation trial (e.g., idea: seeker). They tested themselves on the majority of trials. In Experiment 2, when the hint options were removed, they chose restudy rather than pure test on the majority of trials. These findings show that people prefer self-testing over restudy as long as they can get the answer right on the test. However, we would not recommend hints if they impaired learning compared to pure testing. Experiment 3 showed that this was not the case; the three retrieval conditions from Experiment 1 led to equivalent amounts of learning, and all three outperformed the pure presentation condition. We used different materials in Experiment 4 and found that the hints made retrieval slightly less beneficial when the hints made it possible to guess the answers without thinking back to the study phase (e.g., whip: pu__sh). In summary, hints catalyzed people's intuitive desire to self-test, without any downside for learning, thus making their self-regulated study more enjoyable and effective.

Investigating and explaining the effects of successive relearning on long-term retention.

Research on techniques for enhancing long-term retention has focused almost exclusively on single-session learning conditions. However, even the most potent initial learning manipulations typically do not yield retention levels sufficient for successful performance in many real-world contexts. In contrast, successive relearning (i.e., practicing to some level of mastery in multiple learning sessions) is a promising technique for enhancing long-term retention, but surprisingly few studies have been conducted to date. In two experiments, the advantage of successive relearning over single-session learning (i.e., relearning potency) was substantial, ds = 1.52 to 4.19. Additionally, the effects of initial lag on retention were sizable prior to relearning (ds = 1.07 to 1.87) but attenuated after relearning (ds = -.20 to .38; the extent to which relearning attenuates the effects of initial learning conditions is referred to as relearning override). To evaluate a possible explanation of these effects, we formally tested the relearning attenuatesdecay (RAD) model. The RAD model accurately accounted for relearning potency and relearning override effects (and several other secondary effects). We discuss the practical and theoretical implications of these findings and advocate that successive relearning is an important next frontier for educationally relevant memory research. (PsycINFO Database Record

Retrieval attempts enhance learning regardless of time spent trying to retrieve.

Attempting to retrieve information from memory is an engaging cognitive activity. We predicted that people would learn more when they had spent more time attempting to retrieve. In experiments 1a and 1b, participants were shown trivia questions for 0, 5, 10, or 30 seconds and then the answer was revealed. They took a final test immediately or after 48 hours. Retrieval enhanced learning, but the length of the retrieval attempt had no effect (i.e., final test performance was equivalent in the 5-, 10-, and 30-second conditions and worse in the 0-second condition). During the initial retrieval attempt, more time did increase recall, suggesting that participants continued to engage in productive retrieval activities when given more time. Showing the answer for longer (7 versus 2 seconds) increased learning in Experiments 2a and 2b. Experiment 3 examined the effect of retrieval success and Experiment 4 replicated the results using different materials. These results have direct implications for current theories of retrieval.

Effects of successive relearning on recall: Does relearning override the effects of initial learning criterion?

Retrieval practice improves memory for many kinds of materials, and numerous factors moderate the benefits of retrieval practice, including the amount of successful retrieval practice (referred to as the learning criterion). In general, the benefits of retrieval practice are greater with more than with less successful retrieval practice; however, learning items to a higher (vs. lower) criterion requires more time and effort. If students plan on relearning material in a subsequent study session, does the benefit of learning to a higher criterion during an initial session persist? In Session 1, participants studied and successfully recalled Swahili-English word pairs one, two, three, four, five, six, or seven times. In subsequent sessions, all of the pairs were relearned to a criterion of one correct recall at one-week intervals across four or five successive relearning sessions. Experiments 1 and 2 revealed that the substantial benefits of learning to a higher initial criterion during the first session do not persist across relearning sessions. This relearning-override effect was also demonstrated in Experiment 2 after a one-month retention interval. The implications of relearning-override effects are important for theory and for education. For theories of test-enhanced learning, they support the predictions of one theory and appear inconsistent with the predictions of another. For education, if relearning is to occur, using extra time to learn to a higher initial learning criterion is not efficient. Instead, students should devote their time to subsequent spaced relearning sessions, which produce substantial gains in recall performance.

Does testing impair relational processing? Failed attempts to replicate the negative testing effect.

Recent research on testing effects (i.e., practice tests are more effective than restudy for enhancing subsequent memory) has focused on explaining when and why testing enhances memory. Of particular interest for present purposes, Zaromb and Roediger (2010) reported evidence that testing effects in part reflect enhanced relational processing, which refers to the encoding of similarity among to-be-learned items. The multifactor account of testing effects (Peterson & Mulligan, 2013) further distinguishes between processing of cue-target relations (intraitem relational processing) and processing of relations shared by targets from different items (interitem relational processing). The intriguing claim of this account is that testing enhances intraitem relational processing at the expense of interitem relational processing. Confirming predictions of this account, Peterson and Mulligan (2013) found negative testing effects on final free recall and on a measure of interitem relational processing (the same measures on which Zaromb and Roediger found positive testing effects). The original intent of the current research was to resolve this theoretical debate by replicating and extending the findings of Peterson and Mulligan (2013) to identify the locus of the apparent inconsistency in the outcomes reported in these 2 studies. However, 5 high-powered experiments affording 8 comparisons of testing versus restudy did not replicate the negative testing effect on final memory performance nor on most measures of interitem relational processing. Thus, the weight of the evidence supports the conclusion that testing does not impair relational processing.

Does the benefit of testing depend on lag, and if so, why? Evaluating the elaborative retrieval hypothesis.

Despite the voluminous literatures on testing effects and lag effects, surprisingly few studies have examined whether testing and lag effects interact, and no prior research has directly investigated why this might be the case. To this end, in the present research we evaluated the elaborative retrieval hypothesis (ERH) as a possible explanation for why testing effects depend on lag. Elaborative retrieval involves the activation of cue-related information during the long-term memory search for the target. If the target is successfully retrieved, this additional information is encoded with the cue-target pair to yield a more elaborated memory trace that enhances target access on a later memory test. The ERH states that the degree of elaborative retrieval during practice is greater when testing takes place after a long rather than a short lag (whereas elaborative retrieval during restudy is minimal at either lag). Across two experiments, final-test performance was greater following practice testing than following restudy only, and this memorial advantage was greater with long-lag than with short-lag practice. The final test also included novel cue conditions used to diagnose the degree of elaborative retrieval during practice. The overall pattern of performance in these conditions provided consistent evidence for the ERH, with more extensive elaborative retrieval during long- than during short-lag practice testing.

Repeated retrieval practice and item difficulty: does criterion learning eliminate item difficulty effects?

A wealth of previous research has established that retrieval practice promotes memory, particularly when retrieval is successful. Although successful retrieval promotes memory, it remains unclear whether successful retrieval promotes memory equally well for items of varying difficulty. Will easy items still outperform difficult items on a final test if all items have been correctly recalled equal numbers of times during practice? In two experiments, normatively difficult and easy Lithuanian-English word pairs were learned via test-restudy practice until each item had been correctly recalled a preassigned number of times (from 1 to 11 correct recalls). Despite equating the numbers of successful recalls during practice, performance on a delayed final cued-recall test was lower for difficult than for easy items. Experiment 2 was designed to diagnose whether the disadvantage for difficult items was due to deficits in cue memory, target memory, and/or associative memory. The results revealed a disadvantage for the difficult versus the easy items only on the associative recognition test, with no differences on cue recognition, and even an advantage on target recognition. Although successful retrieval enhanced memory for both difficult and easy items, equating retrieval success during practice did not eliminate normative item difficulty differences.

When is guessing incorrectly better than studying for enhancing memory?

Recently, Kornell, Hays, and Bjork (Journal of Experimental Psychology: Learning, Memory, and Cognition 35:989-998, 2009) demonstrated that incorrect guessing can benefit subsequent memory to a greater degree than can an equivalent amount of study time. We explored this intriguing finding to determine which factors moderate the advantage of incorrect guessing relative to study. In contrast to the findings of Kornell et al., our Experiment 1 revealed that incorrect guessing resulted in worse performance than did studying and that the number of incorrect guesses did not moderate the effect. In contrast, Experiment 2 revealed that the timing of subsequent study moderated the effectiveness of incorrect guessing over study. Final test performance was greater for incorrectly guessed items than for prestudied items when a subsequent study opportunity occurred immediately after the pretrial, whereas the pattern reversed when subsequent study was delayed. This crossover interaction emerged largely because prestudy items showed a classic spacing effect, whereas the guess items did not. One plausible explanation for the absence of a spacing benefit for guess items is that delaying the subsequent study trial increases source-monitoring errors during retrieval, such that participants confuse their original guess with the correctly studied target. However, Experiment 3 provided evidence against this source-monitoring account. We concluded by discussing other possible accounts of why the timing of study could moderate the effectiveness of incorrect guessing.

Diagnosing criterion-level effects on memory: what aspects of memory are enhanced by repeated retrieval?

Previous research has shown that increasing the criterion level (i.e., the number of times an item must be correctly retrieved during practice) improves subsequent memory, but which specific components of memory does increased criterion level enhance? In two experiments, we examined the extent to which the criterion level affects associative memory, target memory, and cue memory. Participants studied Lithuanian-English word pairs via cued recall with restudy until items were correctly recalled one to five times. In Experiment 1, participants took one of four recall tests and one of three recognition tests after a 2-day delay. In Experiment 2, participants took only recognition tests after a 1-week delay. In both experiments, increasing the criterion level enhanced associative memory, as indicated by enhanced performance on forward and backward cued-recall tests and on tests of associative recognition. An increased criterion level also improved target memory, as indicated by enhanced free recall and recognition of targets, and improved cue memory, as indicated by enhanced free recall and recognition of cues.