Does the testing effect work for your exams Image: University of Illinois Library, CC BY 2.0, via Wikimedia Commons

Does the testing effect work for your exams

Most students revise by reading their notes again, often with a highlighter. Asked which method works, they choose rereading, and they are wrong by a wide margin. The experiment that shows it is simple enough to repeat in a classroom, and it has been repeated in hundreds.

The 2006 experiment

Henry Roediger and Jeffrey Karpicke, at Washington University in St Louis, gave undergraduates a short science passage of about 250 words to read for seven minutes. One group then read it again for another seven minutes. The other group spent the seven minutes writing down everything they could remember from it, with no feedback. Five minutes later, the rereading group recalled 81% of the ideas and the recall group 75%. Two days later the order had reversed: 54% against 68%. A week later it was 42% against 56%.

In a second experiment the passage was studied four times, or studied three times and recalled once, or studied once and recalled three times. A week later the three groups remembered 40%, 56%, and 61%. The group that had read the passage only once and spent the rest of its time trying to recall it remembered half as much again as the group that had read it four times. Asked to predict their own scores, the students had predicted the opposite. The paper, "Test-Enhanced Learning", appeared in Psychological Science and is at the journal.

The testing effect, or retrieval practice, is the finding that pulling something out of memory strengthens it more than putting it in again. It is one of the best replicated results in cognitive psychology. A review in 2013 by John Dunlosky and colleagues, published in Psychological Science in the Public Interest and free to read, rated ten study techniques by the evidence. Two were rated as having high utility: practice testing and distributed practice. Rereading, highlighting, and summarizing were rated low, because they do little for the time they take and students believe they do a lot.

Why rereading feels better

The second reading of a passage is fluent. The sentences are familiar, nothing is hard, and the fluency is read as knowledge. A recall attempt is the opposite: it is slow, it fails often, and it exposes every gap. The feeling of difficulty is the feeling of learning happening, and the feeling of ease is the feeling of recognition, which is a different and much weaker thing. Students who have been told the figures still report that rereading feels more productive, which is why the method has to be chosen on the evidence rather than on how it feels.

The effect is larger when the retrieval is harder, up to a point. Recalling without cues beats multiple choice; recalling after a delay beats recalling at once; and getting feedback after the attempt adds more, especially for the items you got wrong. It also transfers: students who practise recalling facts do better on later questions that ask them to apply those facts, which was the worry about the method, and a 2011 study in Science found that one recall session beat a session of drawing concept maps, the technique many teachers prefer, on inference questions a week later.

What spacing adds

The second high-utility technique is older. Hermann Ebbinghaus found in the 1880s that the same number of repetitions produced better retention when they were spread over days than when they were crammed into one session. The spacing effect has been found in hundreds of experiments since, and a meta-analysis of 254 studies by Nicholas Cepeda and colleagues in 2006, at the journal, put the benefit at about 10 to 15 percentage points of recall for the same study time. The best gap between sessions depends on how long you need to remember: roughly a day or two for a test next week, a week or two for a test in a month, and a month or more for something you want to keep for a year.

Spacing works for a reason that follows from the testing effect. A session after a gap starts with some forgetting, so the material has to be partly retrieved rather than only recognized, and retrieval is what strengthens it. Cramming avoids the forgetting and so avoids the retrieval. It produces a high score on the next morning and a low one a month later, which is the pattern every student has seen in themselves.

A plan built on both

Put the two effects together and a revision plan writes itself. It costs no more time than rereading; it only moves the time around.

  1. After each lesson or unit, close the book and write down, from memory, what it said. Use the headings as cues if you must. Then open the book and mark what you missed. Ten minutes, the same day.
  2. Turn the material into questions, not summaries. A list of 20 questions on a unit, with the answers on the back of the page or in a flashcard program, is the raw material for every later session.
  3. Answer the questions two days later, a week later, and a month later. Say or write the answer before you look. Any card you get wrong comes back sooner; any card you get right twice goes further out. Free spaced-repetition software such as Anki keeps the schedule for you and is the tool of the week for the memory unit of the masterclass.
  4. Mix subjects within a session. Interleaving topics is harder than working through one at a time and is rated as a promising technique in the 2013 review, for the same reason: it makes each retrieval harder.
  5. Use past papers as recall practice, under time pressure, before you feel ready. Feeling unready is the condition under which the method works.

What it does not do

Retrieval practice strengthens what you have already understood. It will not explain a proof or a mechanism that you never followed, and it does not replace working the examples. Its effect is also smaller for very complex material, and some studies with long, elaborate texts have found little benefit, which is an open question in the field. For the memorization that every subject contains, from the dates of history to the stages of mitosis to the vocabulary of a language, the evidence is as clear as evidence in psychology gets.

Every unit of the masterclass ends with a quiz for this reason, and Unit 5 gives the experiments in detail, together with the forgetting curve they are a response to. A student who has read it and still rereads has at least made an informed choice.

Study the mind properly

The Psychology Masterclass: weekly units, quizzes, a final exam, and a certificate.
About the program