Abstract:

Durable learning—retaining knowledge long-term for future application—is a central goal of higher education. Cognitive research consistently shows that distributed practice (spacing repetitions over time), retrieval practice (self-testing), and interleaved practice (mixing topics) strongly enhance long-term retention. Yet many undergraduates rely on less effective habits such as cramming and rereading.

This study evaluated U-Behavior, a scalable, LMS-integrated intervention designed to coach students toward these high-impact strategies in authentic university courses. U-Behavior combines four elements:
1) A brief 10-minute tutorial (week 1) explaining the science and benefits of spacing, retrieval, and interleaving, and introducing the U-Behavior process and visualization tools.
2) Unlimited access to Retrieval Practice Activities (RPAs)—randomized, low-stakes online quizzes linked to course objectives, available all semester immediately after content is taught.
3) Two guided reflections (mid- and late-semester) in which students view their personalized RPA graph (a timeline of quiz attempts, accuracy trends, spacing intervals, and topic mixing) and describe planned changes to their practice habits.
4) Final RPA graph submission graded via a rubric that rewards high levels of distributed and interleaved practice (e.g., ≥3 attempts per RPA on separate days, topic alternation), while preserving student flexibility.

Three experiments tested U-Behavior. Experiment 1 (observational, science-of-learning course) showed students overwhelmingly massed practice on single days despite explicit instruction in effective strategies. Experiments 2 (randomized, introductory biomedical sciences) and 3 (quasi-experimental, upper-division microbiology) compared U-Behavior to a strong control: identical RPAs awarded for the highest score but without tutorial, graphs, reflections, or behavior-based grading.

U-Behavior produced large, significant increases in spacing and interleaving (ds > 1.75) without raising total quiz attempts—improving study quality, not quantity. Roughly half of the intervention participants reached the highest rubric level (consistent spaced, mixed retrieval).

Midterm and final exam scores showed no significant condition differences (likely due to the strong control and partial adoption in the intervention group). However, exploratory regressions revealed that higher spacing strongly predicted better exam performance—even after controlling for total attempts (Experiments 2 & 3) and university GPA (Experiment 3). In Experiment 3, greater spacing also predicted higher scores on an unannounced retention test one month post-course.

U-Behavior effectively bridges the knowing–doing gap for evidence-based study strategies. Through clear behavioral targets, real-time personalized feedback via intuitive graphs, structured reflection, and aligned grading, it empowers diverse undergraduates to self-regulate more effective durable learning practices. As a low-resource, adaptable LMS tool, U-Behavior offers strong potential to improve study quality and long-term retention across disciplines.

Keywords:

Spaced Retrieval, Practice, Self-Regulated Learning, Learning Analytics.