In one sentence
Teacher-made digital flashcards in a spaced-repetition app were linked to higher final-exam scores among first-year nursing students who actually used them — but merely being on a campus that was offered the deck did not move the class average.
What the researchers did
Ingebrigtsen, Miland, Bastesen, and Sæle studied a nationwide Norwegian first-year bioscience course (anatomy, physiology, biochemistry) with a shared high-stakes exam. They invited students at 19 campuses, cluster-randomised 10 to an intervention (a 90-minute lecture on effective techniques plus a teacher-built Anki deck of about 850–1000 cards), and compared them with the remaining campuses. 799 students completed surveys (about 25% of those invited; 87% women; mean age 23.8).
Logistics (including pandemic timetable changes) meant about 31% of students assigned to the intervention never received the lecture or deck. The team therefore analysed as treated: card users (n = 131, 32% of the intervention sample), non-users in the intervention arm (n = 279), and controls (n = 383). Cards were written to official learning outcomes, not a single textbook.
What they found
- Intention-to-treat (assigned intervention vs control): exam means 67.6 vs 65.9 — not statistically significant.
- Users vs non-users: 73.2 vs 65.6 (mean difference 7.6 points; d = 0.42, p < 0.001).
- Card users were about three times as likely to pass (odds ratio 2.84) and more than twice as likely to earn the highest grade (odds ratio 2.31).
- Flashcard use still predicted exam scores after the authors controlled for age, prior grades, study time, and how much of the material students said they had covered.
- Only about one third of the intervention group used the cards; among those who knew the app existed, adoption was about 54%. Reasons for skipping included other question banks, preferred methods, and a slow install.
What this means for learners and educators
- A well-mapped teacher deck can lower the cost of starting spaced retrieval — novices often lack time and domain knowledge to write good cards.
- Access ≠ use. Campus assignment did not raise scores; practice did, in this quasi-experiment.
- If you share Anki files, budget time for install friction and a clear “why this beats rereading” pitch, or many students will never open the deck.
Limitations and what we don't know yet
Use was self-reported, and users may differ in motivation even after statistical controls. The design is quasi-experimental, not a clean randomised comparison of forced practice. One national nursing bioscience exam may not match other fields. Long-term retention after the exam was not the outcome. The lecture and the cards were bundled, so we cannot split “knowing about spacing” from “doing the cards.”
FAQ
Should every course ship a 1,000-card Anki deck?
Only if someone can maintain quality and students can install it quickly. This study is one large nursing course, not a universal rule.
Did the cards cause the higher grades?
They predicted higher grades among users after several controls. Causation is still uncertain because students chose whether to use the deck.
How is this different from student-made cards?
Here the bottleneck the authors targeted was starting high-quality retrieval at all, not the extra learning that can come from writing cards yourself.