How to review MCAT practice questions

Write it down before you read the explanation

Once you have read the explanation you cannot recover what you believed a minute earlier. The correct reasoning overwrites it. You will remember having been confused rather than remembering what you were confused about, and those are not the same record.

So the first two steps of the procedure happen with the explanation still closed. Not to make the review harder — because after step three, the data no longer exists.

Some review guides already tell you to re-attempt a missed question before reading the explanation. That is good advice with a different purpose: it is aimed at reaching the right answer. This is aimed at preserving the wrong one — the specific belief you held, in your own words, so that you can later put it next to what actually happened and read the gap.

The point is not to re-solve it. The point is to record what you thought before the record becomes unrecoverable.

The review procedure

One question at a time. Steps 1 and 2 before you look at anything.

1. Write what you thought the question was asking. One sentence. Not the topic — the task. “Find which inhibitor type this is” is different from “find what happens to maximum velocity”, and mistaking one for the other is a whole category of miss.
2. Write which option you chose, and what made it attractive. In a sentence, in your own words. Not a tag. “B looked right because the passage mentioned surface area and B was the only option about surface area” is data. “Careless” is not.
3. Now read the explanation.
4. Name where your reasoning diverged — the first step that went wrong, not the subject it was about. Everything before that point was fine, and the fix belongs at the divergence, not at the topic.
5. Ask what your chosen option was built to catch. Wrong answers are engineered, and the one you picked was written for a specific error. See why the option you picked is evidence.
6. Decide what kind of failure it was: knowledge you did not have, a stem you read as asking something else, reasoning that went off the rails, or execution that fell over at the end. These four ask for four different responses, and only the first is fixed by more content review.
7. If you got it right, ask how. Solid, by elimination, or reached and then doubted. Right for a reason you can state is not the same as right — see the questions you got right.
8. Write the mechanism in one line, as a behaviour. “The most-rehearsed rule arrived before I finished the stem.” Something that happened, not something you are. You can practise against a behaviour; you cannot practise against a character trait.
9. Do not act on it yet. One occurrence is an event. Wait for the third — see when something becomes a pattern.

That is the whole procedure. It is slower than reading explanations and it is not slower than reading explanations twice, which is what happens when the same question type comes back.

Ask what your wrong answer was built to catch

MCAT wrong answers are not padding. Each one is written to catch a specific, predictable error — which means the option you chose is evidence about what you actually did, whether or not you can remember doing it.

An option that is the right number with the sign flipped exists to catch people who wrote the relationship down backwards. An option that is the correct answer to the second-to-last step exists to catch people who stopped early. An option that restates a familiar textbook phrase the passage never supported exists to catch people answering from memory.

So the question to ask is not why is C right. The explanation covers that, and reading it produces a comfortable feeling of understanding that changes nothing. The question is:

What would someone have had to do to find my option attractive?

This is also the step that rescues review when you genuinely cannot remember your reasoning. You may not recall what you were thinking. The option you selected is still sitting there, and it still narrows the possibilities to a handful. The full set of names for what it narrows to is in the mechanisms behind a careless miss.

Review the ones you got right

Most review advice already tells you to flag correct answers you felt unsure about, and that advice is right as far as it goes. It usually stops at a single flag — sure, or not sure — and a flag is not something you can act on.

A correct answer arrives in one of a few distinguishable ways, and they do not predict the same future:

• Solid. You can state the reason, and the reason is the real one. Nothing to do.
• Unstable. Your reasoning would not survive a small change to the stem. Invert the ratio or swap the inhibitor and the same approach produces a wrong answer.
• Unexamined. You matched the shape of the question and moved on without checking whether your reason was the actual one. It may have been. You did not look.
• By elimination. You removed three options and took what was left without ever affirming it. This works while distractors are visibly wrong, and stops working when they are not.
• Second-guessed. You got there, doubted it, and came back. The answer is right; the process that pulled you away is not, and it will not always return you.

Only the first needs nothing. The rest are the reason accuracy can look stable while a score does not move — they are misses that have not happened yet. Each one is unpacked, with what to do about it, in how to stop making careless mistakes.

You do not need to do this for every correct answer. Do it for the ones where you hesitated, changed your mind, or could not immediately say why the answer was right.

Write sentences, not tags

Almost every review system will tell you to classify each miss — knowledge, reasoning, reading, attention to detail — and to write the label down. Do the classification. Do not let it replace the sentence.

A label is a compression, and the information it throws away is the part that identifies the mechanism. “Reasoning error” is true of thousands of different failures. “I decided it was a competitive inhibitor as soon as I saw the graph and never read the second half of the stem” is one specific thing, and it tells you what to change tomorrow.

How you wrote it is itself evidence

The words carry more than their content. “Obviously B” and “I think B, went back and forth” describe two different states, and the second one is a finding even when B was correct. Hedging language records a reversal you would otherwise not have written down anywhere.

This is why full sentences are worth the extra seconds. A confidence checkbox records that you were unsure. A sentence records what the uncertainty was about.

What you never write is also data

Look at a month of your own notes and read what is missing. A student who has written “careless” forty times and “I did not understand this” zero times has told you something real — not about their errors, but about how they explain themselves to themselves. That pattern is invisible from inside any single entry, and it is one of the more useful things a review habit can surface.

The same rule applies to the sentence itself: describe what happened, not what you are. “I rushed” and “I am impatient” look similar and are not. The first names a behaviour that can be practised against. The second is a verdict, and there is nothing to do with a verdict.

When something becomes a pattern

A category on one question is not a finding. Every failure named on this site will happen to you once for a reason that never repeats — you misread a word, you were distracted, the question was genuinely ambiguous.

The working threshold: one occurrence is an event, two is a coincidence, three is a pattern. Below three, write it down and carry on. At three, change something.

A stricter rule applies to direction. To say a habit is improving or worsening you need three separated points in time, not three instances in one sitting. Three text not consulted misses in one passage set tells you about that passage set. The same three across three weeks tells you about you.

You may have read that patterns across practice tests are mostly noise — that with a few hundred questions spread over a handful of exams, any correlation you find is probably an accident. Across four full-lengths, that is a fair warning. It stops being one at the scale a study period actually runs to. Three separated points across nine weeks of sessions is not the same claim as three points inside one exam, and the threshold above is exactly what keeps the two apart.

Most of the value of this threshold is permission to ignore things. After a long block you will have a pile of individually explicable misses and an instinct to fix all of them. Most are events.

The same threshold does more work after a full-length, where the pile is larger and the temptation to rebuild everything is strongest. That case has its own method — how to review an MCAT full length — built around triage rather than the per-question pass above.

What is actually worth writing down

The standard endpoint of MCAT review is a tally: count which topics came up most, and study the one with the most marks. It is easy to maintain and it mostly reproduces the topic breakdown your question bank already printed on the score screen.

It also cannot see the thing you are looking for. A tally is organised by subject, and the mechanism behind a miss does not respect subjects — the same habit produces a biochemistry miss on Monday and a physics miss on Thursday, and a subject tally files them as two unrelated topics to review.

The columns most logs use cannot produce the answer

This matters more than it sounds. A record of question numbers, your letter, the correct letter and a topic yields topic counts and nothing else. No classification of the error, no pattern across sessions, nothing that says what to do differently — because none of the information that would identify the mechanism was ever captured.

What makes classification possible is the text: the stem, the options, and a sentence about why yours looked right. With those, the specific option you chose becomes readable evidence. Without them, there is nothing to read.

So if you keep notes, keep those things and skip the tidy columns. And be honest about the constraint: this is real work to do by hand, question after question, at the end of a study day when you are already finished.

Where this runs out

Running the procedure on tonight's block is the straightforward part. The hard part is the third occurrence — it usually lands weeks after the first, in a different subject, by which time the first two are not in front of you and the sentence you wrote about them is several pages back.

Running the procedure is one question's worth of the work behind a score that will not move. The wider practice is the same idea applied to a whole log: not labelling each miss, but finding the one mechanism that explains the most of them.

That is the specific problem CLIVEA is built for. Your tutor is in the session while you work, so the analysis is written as you study rather than becoming homework afterwards, and every session is read together with every session before it. You keep no log, which means the record does not depend on how tired you were when the block ended.

None of that replaces doing this yourself, and the procedure above works with nothing but the question in front of you. It is worth running tonight either way.