Spacing the same problems across weeks
Ten permutation problems in one sitting did not beat five. Spreading practice across weeks did. The one-week test is the chart.
Textbooks still pile every problem of one type under the lesson you just read. It feels efficient: you are “in” the method, so you should drill it while the page is warm. Rohrer and Taylor tested that habit with a single kind of problem — counting permutations — and a test one week after practice ended.
Spacing won. Doing twice as many problems in one massed block did not.
Three groups, one skill
Students learned how to calculate permutations, then practised in one of three ways. Spacers spread their problems across sessions a week apart. Massers did a heavier block in one session. Light massers did about half as many massed problems. Everyone sat a test a week after their last practice. During practice itself, spacers and massers were close (about 83% and 82% on the later practice problems). The test is where the schedules split.
[[chart:ld-spacing]]| Schedule | Test accuracy |
|---|---|
| Spaced across weeks | 74% |
| Massed, 10 problems | 49% |
| Light massed, 5 problems | 46% |
More of the same hour is a weak buy
Massers and light massers were statistically hard to tell apart on the test, even though one group had done about twice the massed work. The paper treats that as a null effect of overlearning: once you can solve the type, another immediate copy does little for next week. The return came from leaving the method and meeting it again later, when producing it was slightly harder.
That is not a reason to do fewer problems forever. It is a reason to spend the second hour on a different day, or on a different skill, instead of grinding the same block until it feels automatic tonight and fragile on Thursday.
The study used one procedure and college students. It does not set a magic gap for every subject. It does knock down the excuse that “I already did a big set, so I am done for the week.”
A week that uses the result
- Learn the method on day one. Solve two problems cleanly. Stop the massed pile there.
- Two days later, solve two more with the notes shut.
- Near the end of the week, solve two more. If you miss, repair that miss — do not restart a twenty-problem block.
- Keep a small due pile so the gap is the schedule, not a mood.
Study Lab already spaces cards from the rating you give. You do not need a second calendar for every fact. For a procedure, the same idea is: a short pack today, the twin of that pack later, not a second copy of tonight’s worksheet. When you want the clock, use Practice Hub for a practice test for your exam rather than inventing a mock from the chat.
Read the companion on mixing problem types next. Spacing and mixing are different levers; textbooks usually skip both.
Where the second hour should go
Suppose you have ten permutation-style problems, or ten questions of any one method. The massed group in the study shows what happens if you do all ten tonight: practice looks fine, and the one-week test sits near the group that only did five. Split them. Two tonight, until you can solve the type once without the worked line beside you. Two more after a gap of at least a day. Two more near the end of the week. Hold the last four for the following week, mixed with a second method, rather than “finishing the worksheet” in one sitting.
The gap has to be real. A ten-minute break in the same evening is still massed. The spacers in the experiment came back on later days, after the method had started to fade. That fade is the moment the problem becomes retrieval instead of copying. If you never feel the fade, you are still in the massed block.
School homework will keep arriving in blocked sets. Do the set far enough to learn the method, then stop. Move the leftovers onto cards or a short list you will not open until the next session. Study Lab is the place those leftovers wait, because the due date is the gap. A second full worksheet of the same type tonight is the purchase the study found was weak.