Timing in Study
We’ve already covered retrieval practice in this blog: retrieving from memory consolidates more than rereading, and it’s the effort of pulling out the information that leaves a trace. I won’t repeat the mechanism. What matters now is the when, because retrieving without spacing is like training without rest. It works, but at half the payoff.
Spacing isn’t a fixed table of “24 hours, 3 days, 1 week, 2 weeks, 1 month.” That sequence circulates through forums and exam-prep academies as a rule of thumb, and as a starting point it’s not bad, but there’s no experiment behind it that has actually tested it as such, with those exact numbers. What is behind it is an older, more solid body of research, and it says something different: the optimal interval between reviews isn’t a fixed number of days, it’s a proportion of the time you need to retain the information. Cepeda, Pashler, Vul, Wixted, and Rohrer reviewed 317 experiments drawn from 184 papers — 839 comparisons in total — on spaced versus massed practice, and the advantage of spacing showed up consistently in the vast majority [1]. But the exact figure for that proportion, and how it changes depending on how far off the exam is, comes from a later and more ambitious study: Cepeda, Vul, Rohrer, Wixted, and Pashler taught a dataset to more than 1,350 people, with a review session arriving between 0 and 3.5 months later, and a final test between one week and one year after that review. For an exam a week away, the optimal gap between reviews runs around 20-40% of the remaining time until the exam. For an exam a year away, it drops to 5-10% [2].
Translated into real timelines: eight months of civil-service exam prep isn’t planned the same way as two months of university entrance exams, and a week for an oral exam isn’t planned like either of those two.
Remember that on Pensar es Gratis we only cite reliable data and sources that pass our content filter. At the end of each post there’s a practical guide on the topic if it makes sense for one to exist.
Eight months, 70 topics (or 30)
You have 240 days, 70 topics, and case-study exercises. Or 30 topics and the same timeframe. The difference between 30 and 70 isn’t just volume: it changes how many reviews each topic can get before the exam. With 30 topics and eight months, each topic can be retrieved five or six times with wide intervals. With 70, the margin narrows, and planning has to be more aggressive with the first review.
Applying Cepeda’s proportion [2] to an eight-month horizon: the first review of a freshly studied topic can perfectly well wait three or four weeks, because trying to reinforce content that hasn’t even started to decay yet is time poorly spent. This has a consequence that’s unintuitive but well measured: in the same study, for one-year retention, the optimal gap ended up being several weeks, not days, and performance with that long gap was far superior to spacing just one day or one week apart [2]. Intuition says “review soon.” The evidence says “review in proportion to how long you need to remember,” and for eight months that means waiting longer than seems reasonable.
The most common mistake in long-syllabus exam prep is exactly the opposite: leaving review for the end. Someone who studies 70 topics in linear order and thinks “I’ll review in the last two months” arrives in March with the first 20 topics in ruins. Reverse planning helps: count the real weeks you have, subtract a 20% buffer for the unexpected, and reserve the last three or four weeks purely for intensive review and mock exams. For 70 topics, it probably pays off more to go through the entire syllabus three or four times with a modest accuracy bar per topic than to polish the first twenty topics to perfection and arrive rushed for the rest. That’s a reasonable conjecture, not a settled finding, but it rests on a well-documented pattern: successive relearning — retrieving each topic until you get it right a fixed number of times, and repeating the process in later sessions — greatly improves retention the first few times and then flattens out; pushing past three correct answers in a row in the same session barely adds anything, even though each repetition costs less time. A seventy-unit syllabus punishes a topic that was never reviewed far more than one mastered at 90% instead of 100%.
Case-study exercises aren’t retrieved by reciting. They’re retrieved by working through a closed case from an area you studied a week ago, without looking anything up. And it’s worth varying the example from one review to the next, rather than always repeating the same case: in a study where students watched a recorded lecture and then answered application questions, those who practiced with three different examples of the same concept transferred that knowledge much better to a new case two days later than those who practiced three times with the same example [3]. For an exam case study, this translates directly: don’t repeat the same practical case three times, look for variants.
Two months, eight subjects
This is university entrance exams. Two months, eight subjects, and the temptation to study one subject a week and leave the rest for later. It works poorly, for the same reason cramming works poorly: it feels productive and performs worse. Spacing across multiple subjects requires interleaving: alternating subjects within the same week so each one gets frequent retrieval instead of massed blocks separated by weeks of forgetting.
With eight subjects and sixty days, Cepeda’s proportion places the first serious review of each block of material between 12 and 24 days after first studying it, not a month out. That doesn’t take away value from retrieving earlier, briefly: an attempt to recall a block three or four days after studying it, even if incomplete, already leaves a trace and helps catch gaps early. These are two different things — a quick check a few days later, and the formal review with full retrieval two or three weeks out — and it’s worth not confusing them or substituting the first for the second.
For 8 subjects, the plan that works best reserves the last two weeks purely for timed mock exams per subject. Not for reading. For taking full exams, grading with a rubric, and retraining on typical mistakes. That last week’s retrieval isn’t about content: it’s about format and stamina.
One week, one topic
An oral exam, a single test, one topic. Spacing loses its lead role because there’s no time for long intervals. What’s left is intensive retrieval, at intervals of days rather than weeks.
The seven-day plan that holds up best is this: on day 1 you map out the topic and mark the highest-yield sections; on days 2 through 4 you do spoken retrievals of 90 seconds per section, no notes, with a timer; on day 5 you do a full mock oral exam with someone interrupting you; on day 6 you fix only the blocks that came up; on day 7 you do one last light retrieval and sleep.
On the oral exam specifically, here’s where a conjecture comes in, not a settled finding: nobody has specifically measured whether reciting a topic out loud produces the same kind of advantage as a written free-recall test, but it follows fairly naturally from the same mechanism. Roediger and Karpicke had students read a text and then, instead of rereading it, try to recall it on a blank page, once or three times, with no correction. At five minutes, whoever had reread more won out: 83% recall versus 71% for those who had only retrieved. At one week, the result flipped entirely: 40% for those who had only reread, 61% for those who had retrieved three times [4]. If that effect depends on active reconstruction of the content, rather than on the output format, rehearsing out loud without looking at the script should produce a similar advantage to that written test. That’s a reasonable hypothesis, not something tested specifically with oral exams.
If you get stuck during spoken retrieval, you don’t open the material. You note the block and move on. At the end you review only the blocks. The pattern that emerges is your map of real gaps, not the one you think you have — the same misleading sense of mastery that makes cramming feel better than it actually performs.
Rereading, highlighting, making outlines copied from the material: all of that generates familiarity, and familiarity gets mistaken for knowledge. The system of endless loops over the same topic, with no retrieval and no spacing, is the most popular way to study and one of the least effective. Planning that combines retrieval and spacing isn’t more comfortable. It’s more uncomfortable: it confronts you with what you don’t know. For 70 topics or 8 subjects, for eight months or one week, the principle doesn’t change. The interval changes. The density changes. The format of retrieval changes. What doesn’t change is that remembering is studying, and forgetting at the right time is part of the method.
References
[1] Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380. — Reliable. Meta-analysis of 317 experiments drawn from 184 papers, 839 comparisons in total.
[2] Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095-1102. — Reliable. Study with more than 1,350 participants recruited across several European countries; the optimal gap between reviews runs around 20-40% of the time until the exam for one-week retention, and 5-10% for one-year retention.
[3] Butler, A. C., Black-Maier, A. C., Raley, N. D., & Marsh, E. J. (2017). Retrieving and applying knowledge to different examples promotes transfer of learning. Journal of Experimental Psychology: Applied, 23(4), 433-446. — Reliable. Four experiments; variability in the examples used during retrieval practice improved transfer to new examples, compared with always repeating the same example.
[4] Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255. — Reliable. Compared studying four times versus studying once and retrieving three times; at 5 minutes repeated study won out (83% versus 71%), at one week repeated retrieval won out (40% versus 61%). The study measures retention at 5 minutes, 2 days, and 1 week; it doesn’t include a one-month condition.