ABSTRACT
Retrieval practice is effective for learning, and is often implemented in digital applications through multiple-choice questions (MCQs). Stepwise presentation, with answer options appearing after a brief delay, was recently shown to make MCQs more effective. We tested its effects in an authentic setting: a popular TV website with learning games. Users (N = 13,060) who started a quiz were randomly assigned to standard MCQs, learner-paced stepwise MCQs, or system-paced stepwise MCQs (fixed 6 s delay), all with feedback. Learner-pacing had a (small) positive effect on persistence as measured by the proportion of children completing the entire quiz compared to the other two conditions. Learner-paced delay times suggested that (most) learners processed the question before revealing the answers. The MCQ-condition did not affect short-term learning gains. Overall, learner-paced stepwise MCQs seem preferable in self-study settings, to allow learners to recall information from memory but not frustrate them by enforcing a fixed delay.