Why testing yourself beats re-reading

5 min read

Re-reading worked solutions feels like learning and is not. What the evidence says about retrieval practice, and how to build revision around it.

The problem with feeling prepared

A student reads through their notes and worked solutions, follows every line, and feels ready. Two weeks later, in a timed paper, the same question stops them dead. Nothing has gone wrong with their memory. The revision method simply never asked them to do the one thing the exam asks: produce the answer from nothing.

Psychologists call this the fluency illusion. Recognising a solution is easy and feels like knowing. Generating it is hard and feels like failing. Students consistently mistake the first for the second, and revise accordingly.

What the research shows

In a well-known experiment, Roediger and Karpicke (2006) had students learn material and then either restudy it or test themselves on it. The students who restudied were more confident about how much they would remember. On a test a week later, the students who had practised retrieval remembered substantially more. Confidence and actual learning pointed in opposite directions.

In 2013, Dunlosky and colleagues reviewed ten common study techniques against the available evidence. Several of the most popular, including highlighting, re-reading and summarising, were rated low utility. Practice testing was one of only two techniques rated high utility, holding up across ages, subjects and time frames.

Robert and Elizabeth Bjork describe the underlying principle as a desirable difficulty: conditions that make practice feel harder and slower, while making what you retain considerably stronger.

Why this matters in maths specifically

Maths papers almost never ask you to recognise a method. They present an unfamiliar problem and expect you to select a method and execute it under time pressure. Reading a worked solution rehearses recognition only. Working a problem from a blank page rehearses selection and execution together, which is the actual exam task.

There is a second benefit that matters just as much. Retrieval exposes what a student cannot yet do; re-reading hides it. A student who re-reads finishes a session with a list of topics covered. A student who self-tests finishes with a list of questions they got wrong, which is a far more useful thing to own two months before an exam.

How to revise with retrieval practice

• Work from a blank page. Close the book, cover the solution, and rebuild the question from scratch. If you cannot start, that is information, not failure.
• Mark honestly. Award nothing for 'I would have got that'. An examiner will not.
• Redo failures later, not immediately. Reading a correction fixes very little. Reproducing the same question two days later, unaided, is what makes it stick.
• Brain-dump before revising. Write everything you can recall about a topic before opening any notes, then use the gaps to direct the session.
• Treat past-paper questions as the main revision activity, not the final check at the end.

A simple weekly loop

Choose the topic, attempt questions cold, mark them strictly, note every failure, then re-attempt those specific questions 48 hours later without looking at anything. Repeated across a term, that loop reliably outperforms any amount of reading through neat worked examples.

It will feel harder and less satisfying than traditional revision, and progress will feel slower in the moment. That discomfort is the mechanism doing its work, not a sign that it is going badly.