Metacognitive awareness of the pretesting effect improves with self-regulation support

Abstract

The pretesting or prequestion effect refers to the counterintuitive finding that taking tests on information that one has yet to learn, during which many erroneous responses typically occur, can benefit learning relative to nontesting methods (e.g., reading) if the correct answers are studied afterwards. Using a knowledge updating approach that entailed two or three cycles of pretesting versus reading followed by a criterial test, we investigated (a) the extent to which learners develop metacognitive awareness of the pretesting effect through experience (as evidenced by predictions of criterial test performance) and (b) three forms of external support-namely, performance feedback (displaying criterial test performance for pretested versus read items), prediction reminders (displaying learners' predictions alongside performance feedback), and recall prompts (asking learners to remember criterial test performance during the first cycle prior to making predictions for the second cycle)-that might improve, or provide insights into, such awareness. Across five experiments, we found that learners generally lack awareness of the memorial benefits of pretesting, are predisposed to believing that reading is more effective even after repeatedly experiencing both techniques, and need support before they recognize that pretesting is more beneficial. Overall, these results underscore the challenge of, and highlight several means of dislodging, learners' inaccurate beliefs about the efficacy of pretesting.

Keywords: Errorful generation; Knowledge updating; Metacognition; Prequestions; Pretesting effect.

Fig. 1

Overview of experimental procedure. Note. Within each of two consecutive cycles, participants learned 32 word pairs via reading or pretesting (two lists of 16 pairs each, randomly intermixed), made global-differentiated predictions, and then took a 5-min delayed criterial test. Experiments 2–4 featured performance feedback after both criterial tests. The feedback in the first cycle of Experiments 3–4 also included reminders of participants’ original predictions. During the second cycle of Experiment 4, participants were further prompted to recall their test performance in Cycle 1 prior to making new predictions. Experiment 5 featured the same single-session, multicycle design but included three cycles instead of two

Fig. 2

Global-differentiated predictions and criterial test performance in Experiment 1

Fig. 3

Performance feedback (Experiments 2–5) and feedback with prediction reminders (Experiments 3–5). Note. The above screenshots display the different forms of external support that appeared after the criterial test in the first cycle of the indicated experiment (and, in Experiment 5, after the second cycle as well). After viewing performance feedback (or performance feedback with prediction reminders), participants were asked to indicate the performance difference between the read and pretested conditions (Experiment 2 and the performance feedback group in Experiment 5), or the degree to which their prediction matched their actual performance in each of the read and pretested conditions (Experiments 3, 4, and the performance feedback with prediction reminders group of Experiment 5). In Experiments 3–5, such feedback, where provided, was viewed once more before beginning the next cycle of the experiment. See https://osf.io/jxwcr for a larger version of this figure

Fig. 4

Global-differentiated predictions and criterial test performance in Experiment 2

Fig. 5

Global-differentiated predictions and criterial test performance in Experiment 3

Fig. 6

Global-differentiated predictions and criterial test performance in Experiment 4

Fig. 7

Global-differentiated predictions and criterial test performance in Experiment 5