Does cramming the night before an exam work?

Nate Kornell had students learn GRE-style vocabulary with flashcards. He compared two ways of dividing the same work. In one, a single large spaced deck, where each card comes back after a long while. In the other, four small massed decks, where cards come back quickly. He also compared spacing with cramming the night before (Kornell, 2009).

The result is clear. Spacing beat massing, and spacing beat cramming the night before. For 90% of the participants, spacing was more effective (Kornell, 2009). It is only one study, on vocabulary, with cards. It says nothing about the physics chapter you have to understand. But it points the same way as Cepeda’s meta-analysis, which finds that spacing beats massing across hundreds of measures (Cepeda et al., 2006). So cramming is not the best way to go about it, and the research says so clearly.

In the same study, 72% of the participants thought, after the first session, that massing worked better (Kornell, 2009). They were wrong. Cramming does have a real advantage in the very short term, everything is still fresh. Roediger and Karpicke measured it, at five minutes, rereading gave a better score than quizzing yourself. A week later it was the other way round, and the rereaders were still the most sure they would remember (Roediger and Karpicke, 2006).

That is the trap of the night before finals or the SAT. At two in the morning you reread, everything looks familiar, and you feel ready. That familiarity is a misleading signal. Karpicke and colleagues speak of an illusion of competence in students who reread a lot and quiz themselves little (Karpicke et al., 2009). If your notes feel clear, check it by saying them back without looking, not by rereading them a fifth time.

Sometimes the night before is all you have left. Then do the least bad cramming possible. First, quiz yourself instead of rereading. In Karpicke and Roediger’s study, once a word was known, studying it again had no effect on delayed recall, while retesting it had a large positive effect (Karpicke and Roediger, 2008). So spend your time on what you miss, not on what you already know.

Next, make several passes rather than one. Even if the delay is short, letting a little time go by between two passes is better than running them back to back. Cepeda showed that the best gap is a fraction of the time before the test, so it gets shorter when the exam is near (Cepeda et al., 2008). Finally, plan a moment to sleep. Sleep takes part in the consolidation of memories (Rasch and Born, 2013). An all-nighter sacrifices exactly that.

Here is a simple run-through. Start by listing the topics in the syllabus as short sentences, one idea per line. Then close your notes and try to say each one out loud or write it from memory. Mark the ones you miss. Open the notes again only for those, correct them, then come back to them after a break rather than in the next minute.

Keep the shakiest ideas for a last pass late in the evening, then stop and sleep. Think of the morning, rereading a sheet on the way there is harmless. Jenkins and Dallenbach observed as early as 1924 that people forget less when they sleep after learning than when they stay awake (Jenkins and Dallenbach, 1924). It only gives a reason not to sacrifice the whole night to the last review.