Mistakes: Your Best Study Data

Most students treat a wrong answer as a verdict: 'I'm bad at this.' They feel bad, move on, and make the same mistake in the exam. Treat it as information instead — 'which part broke?' — and you can diagnose it, fix it, re-test it, and turn it into a mark. Here's the uncomfortable flip side: a practice test you score 100% on taught you nothing new. It confirmed what you already knew. The value of practice sits almost entirely in the questions you got wrong, which means the mistakes are the point of doing it, not an unfortunate side effect.

Four kinds of mistake, and each needs a different fix. A KNOWLEDGE GAP means you never knew it — go back and learn it, then test it. An UNDERSTANDING GAP means you knew the fact but used it wrongly; re-learn the why rather than the what, and explain it aloud until the explanation holds. An APPLICATION GAP means you knew the content but didn't recognise where it applied — that's fixed by varied, interleaved practice, not by more re-reading. A CARELESS ERROR means you knew it and could do it and simply slipped; that isn't a study problem at all, it's a checking problem. Diagnosing wrongly is expensive: re-reading a chapter won't fix an application gap, and more practice questions won't fix a knowledge gap.

An error log is five lines per mistake. One: the question. Two: what I answered — exactly what you wrote, not a tidied-up summary. Three: the correct answer in full, with the reasoning. Four: why I got it wrong — which of the four types, and what precisely broke. Five: what I'll do about it, and when I'll re-test. Line four is the one everybody skips, and it's the only line that turns a mistake into a fix. Keep this up and by two weeks before an exam you're holding the highest-value revision document you own: a list of your gaps, in your handwriting, already diagnosed.

Individual mistakes are useful; patterns across mistakes are gold. Review your error log once a week and ask what keeps coming back. Is it always the same topic? Always the same type — say, always application gaps rather than knowledge gaps? Always word problems, or always the final step of long calculations, or always questions that appear late in a test when you're tired? Six separate wrong answers over three weeks very often trace back to one misunderstood idea. Fix that cause once and all six disappear together, which is a far better return than fixing six symptoms one at a time.

Here's the loop that makes all of this work. TEST to find the gap. DIAGNOSE which of the four types it is. FIX it with the matching repair. RE-TEST with a new question on the same idea — and this is the step almost everyone skips, because reading your own correction feels like fixing it. It isn't. A mistake is not fixed until you get a different question on the same idea right, without help. Then SPACE IT: come back to that idea in three days and again in a week, because a gap you closed once can quietly reopen. Test, diagnose, fix, re-test, space — round and round, and the log gets shorter every week.