Sort your wrong answers before you study them

The standard advice after a bad exam is to go through your mistakes. Fine. But most people do this by copying the correct answers out neatly, and that does almost nothing.

Rewriting a correct solution produces the feeling of understanding it — of course it does, it's right there — without touching whatever caused the mistake. And that's the problem, because the cause is different every time.

The four kinds

1. I didn't know it.

You could not have produced this. The concept wasn't there, or the method was never learned.

This is the only category where "review the material" is the right response. It's also the minority of most people's mistakes, which is inconvenient, because it's the one everyone assumes they're making.

2. I knew it but misread the question.

You solved a problem, competently. It was not the problem on the page. The question asked for the marginal effect and you gave the total; it said "which is not true" and you found one that was.

Studying more content cannot fix this. It's a reading habit, and it's fixed by a reading habit: underline what's being asked before you start, and reread the question after you've written your answer. Sounds trivial. Recovers real points.

3. I knew it and slipped.

Right method, right setup, dropped a negative sign in line four.

Everyone writes this off as carelessness and moves on. Don't — slips cluster. Almost everybody has one or two specific places where they consistently break: distributing a negative across parentheses, transcribing a number from one line to the next, unit conversions.

If you log these for two exams you'll see your pattern, and once you know your pattern you can slow down at exactly that step instead of trying to be globally more careful, which nobody can do.

4. I ran out of time.

You never got there. Or you got there with four minutes left and wrote something incoherent.

This is a pacing problem and it's fixed by timed practice, full stop. Studying more content makes it worse if anything, because more knowledge with no time budget means you spend even longer on question 2.

Why the sorting is the point

Three of those four are not content problems. If you treat every wrong answer as a content problem, you'll spend your prep on the one category out of four that studying fixes, and the other three will show up again on the next exam, identically.

I found this out by doing exactly that. Two exams in a row, same course, same result, and the second time I finally sat down and sorted. Of eleven wrong answers, four were things I genuinely didn't know. Three were misreadings. Three were sign errors. One was the last question, unattempted.

The four I "should" have been studying were less than half. The seven I'd been ignoring were the difference between the grade I got and the grade I wanted.

Actually doing it

Keep one page per course. Not a notebook of neatly rewritten solutions — a log.

For each mistake, three things:

1. What the question asked, in a few words
2. Which of the four kinds it was
3. One sentence on the fix

That's it. Fifteen minutes after an exam or a returned problem set.

Then, before the next exam, read the log. Not the correct answers — the log. It tells you what to practice, and it's specific in a way that "review chapters 4 through 7" never is.

Getting the data in the first place

You need your graded exams back and a lot of courses don't hand them out permanently.

If yours doesn't, exam review is what office hours are for. Most instructors will let you look at your exam in their office. Bring paper and the four categories, and go through your wrong answers sorting them. Twenty minutes, and you'll leave with the most useful study document of the term.

Do the same with returned problem sets and quizzes during the semester, not just after exams. The pattern is visible in low-stakes work too, and finding it in week five is worth substantially more than finding it in week fifteen.

A note on the fourth category

Running out of time is the one people are most embarrassed by and it's the most mechanically fixable.

Take a practice exam under real time. Note where you were at the halfway mark. If you were 30 percent through the paper at the halfway point, you have a pacing problem, and pacing responds to practice fast — usually within two or three timed sessions.

The specific habit worth building: decide your time budget per section before you start, write it on the exam, and enforce it. When a question exceeds its budget, leave it, move on, come back. Most people know this and very few do it, because abandoning a half-finished problem feels like giving up. It isn't. It's the only way to make sure that the questions you can definitely do get done.