Data from this survey replicated the major outcomes of Kornell and Bjork (2007) and Karpicke et al. (2009): College students reported using a self-testing strategy (Fig. 2) largely for monitoring their learning progress, and also reported the use of a variety of other strategies, such as rereading and not studying what they already know. Consistent with expectations from SRL theory, the present study also revealed that some of these strategies are related to college students’ GPAs. Perhaps most impressively, the use of a self-testing strategy—which boosts performance when administered by an experimenter or teacher (Roediger & Butler, 2011)—is also related to student success when used spontaneously for academic learning. As is shown in Fig. 1, almost all of the most successful students (GPA > 3.6) reported using this strategy, and its reported use declined with GPA.
A major issue is the degree to which these benefits of self-testing will generalize to different kinds of tests (e.g., multiple choice, free recall, or essay), different course contents (e.g., biology, psychology, or philosophy), students with differing abilities, and so forth. Current evidence suggests that self-testing has widespread benefits across different kinds of tests, materials, and student abilities. For instance, self-testing by recalling the target information boosts performance on subsequent recall and multiple-choice tests of the target information, and it also boosts performance on tests of comprehension (for reviews, see Roediger & Butler, 2011; Roediger & Karpicke, 2006; and Table S1 from Rawson & Dunlosky, 2011). Nevertheless, it undoubtedly will not be useful for some courses, and if so, our present results may underestimate the power of self-testing, because the composite GPA would reflect courses in which testing would (and would not) matter. On the basis of this rationale and the positive evidence from the present study, future research should examine self-testing and grades for specific classes that vary in the degrees to which they afford self-testing as a potentially effective strategy.
Reported use of rereading was also related to GPA, which might be viewed as surprising, given that rereading does not always improve performance in the laboratory (e.g., Callender & McDaniel, 2009). When used correctly, however, it can boost retention and performance (e.g., Rawson & Kintsch, 2005), and the present rereading–GPA relationship may in part arise from students who read (a lot) versus those who do not read. In contrast to rereading, other reported strategies that presumably are effective did not predict GPA. In particular, the reported use of outlines and collaborative learning demonstrated slightly negative relationships, and the use of diagrams and highlighting were not significantly related to GPA. These outcomes are provocative, because many students believe that these strategies are beneficial when in fact they will not always boost learning. For instance, although studying with friends may have some benefits, students may not always collaborate appropriately when studying together. Also, highlighting by a textbook publisher or instructor can improve performance, but students’ use of highlighting has been shown to yield mixed results, depending on the skill of the user (e.g., Bell & Limber, 2010; Fowler & Barker, 1974). Thus, at least some of these strategies may actually be relatively inert when used by typical students. Based on the present study, however, it would be premature to conclude that these strategies hold absolutely no benefits for student success, because the survey did not measure how often a given student used each strategy and how well the strategies were used. Even self-testing (which was related to GPA) can be used ineffectively, such as when students test themselves by evaluating their familiarity with a concept without trying to recall it from memory (cf. Dunlosky, Rawson, & Middleton, 2005). An exciting avenue for future research will be to develop methods that allow researchers to describe students’ study behavior at a more fine-grained level, such as how often they use self-testing and exactly how they use it to monitor learning.
Although self-testing predicted GPA, the use of flashcards—a popular form of self-testing—unexpectedly did not. In fact, these two strategies were unrelated in the present study (r = .02, p = .68) and might be perceived as different by students. Among students who reported regular use of flashcards, approximately 30% did not report self-testing, which suggests that many flashcard users do not use them to self-test. Flashcards may often be used nonoptimally in vivo, such as when students mindlessly read flashcards without generating responses. Even when they are used appropriately, flashcards may be best suited to committing factual information to memory and not equally effective for studying all types of materials. In contrast, self-testing could also include answering complex questions or solving practice problems, which might encourage deeper processing and yield larger payoffs in performance across many types of materials and courses.
Even those strategies that best predicted GPA were only weakly predictive, which might suggest that students’ strategy choices have little consequence for their grades. Are other factors—such as motivation, interest, intelligence, environment, or competing demands—simply more important? Although possible, several reasons exist for why the correlations observed in the present study are expected to be small, even if some strategies are effective over a wide range of students, tests, and content (e.g., self-testing; Roediger & Butler, 2011). First, different students might have had different courses in mind (e.g., calculus vs. philosophy) when responding to the survey, which would create variability in responding and could obscure strategy–GPA correlations. Future research might overcome this limitation with test–retest methods, longitudinal follow-ups, or more context-specific questions. Second, any strategy could be used well or poorly. This variability in how well strategies are used would obscure how valuable they might be if used ideally. And, third, the present survey asked students to report whether they did or did not use a given strategy regularly (binary responses), rather than how much or how often a strategy was used. Future research will benefit from measuring the degree of usage (a continuous response scale), which might enhance the ability of study strategies to account for variance in performance.
A unique aspect of the present study was the investigation of students’ time management. Differences in scheduling did arise between the highest and lowest achievers, with the lower achievers focusing (a) more on impending deadlines, (b) more on studying late at night, and (c) almost never on planning their study time. Reports of spacing study (vs. cramming) were not significantly related to GPA, even though spaced (vs. massed) practice is known to have a major impact on retention (Cepeda et al., 2006). Although this outcome is surprising, cramming the night (and immediately) before an exam might support relatively good exam performance, even though students who use this strategy might remember little of the content even a short time after the exam. Furthermore, scheduling study sessions in a spaced manner may afford the use of other strategies, which themselves improve student success. Although these ideas are speculative, post-hoc analyses indicated that the reported use of spacing (vs. cramming) was significantly related to the use of more study strategies overall (r = .15, p < .009; combined Question 12 reports) and, in particular, was related to the use of self-testing (r = .11, p = .05) and rereading (r = .15, p = .007). These relationships are small, but they do suggest that spacing may support the use of more effective strategies.
In summary, low performers were especially likely to base their study decisions on impending deadlines rather than planning, and they were also more likely to engage in late-night studying. Although spacing (vs. massing) study was not significantly related to GPA, spacing was associated with the use of more study strategies overall. Finally, and perhaps most importantly, self-testing was a relatively popular strategy and was significantly related to student achievement.