In one sentence
A simple grading rule — points for each day you complete a practice set, not for every question — pushed students to space their work and raised exam scores in a within-course RCT, with the biggest unlock for lower-GPA learners.
What the researchers did
YeckehZaare and Resnick tested a grading incentive they call Counting Days in two randomised controlled trials. In the counting questions control, students earned points for each practice question answered. In the counting days treatment, points came for each day on which a student answered a set of questions — rewarding spread over time rather than binge volume.
One RCT randomised 143 students within a single course. Another randomised 71 instructors (between-instructor design). The within-course experiment could compare exam learning; the instructor experiment compared practice behaviour (days and questions) when instructors used different incentives.
What they found
- In the within-class experiment, the counting-days group earned higher exam scores, mediated by spreading practice across more days.
- Spacing helped lower-GPA students especially: the correlation between course exam scores and prior-course GPA was significantly lower under counting days — meaning prior GPA predicted the new exam less tightly.
- In the between-instructor experiment, learning outcomes could not be compared across instructors, but both number of practice days and number of questions practiced were significantly higher under counting days.
What this means for learners and educators
- If your platform or course awards credit for raw question count, you may accidentally reward massed cramming; switching the unit of credit to days active is a low-tech nudge toward spacing and retrieval.
- Students who historically struggle (lower GPA) may benefit most when the incentive unlocks distributed practice rather than only rewarding speed.
- Instructors adopting the rule saw more days and more questions — so volume need not collapse when you stop paying per item.
- Pair the incentive with a clear “set” definition so gaming (tiny daily check-ins) stays aligned with real practice.
Limitations and what we don't know yet
The instructor-level RCT could not compare learning outcomes across classes — only practice metrics. Course context, subject matter, and exact point weights matter and may not travel unchanged. Mediation via spacing days is persuasive inside one course but still sits inside a graded ecosystem where motivation is entangled with points. Long-term retention beyond the course exam is not reported in the abstract. Implementation fidelity across 71 instructors will vary.