simplesat

Behavioral leaks

How to Review an SAT Practice Test

Updated July 7, 2026

How do you review an SAT practice test?

You took the test, checked the score, groaned or shrugged, and reached for the next one. That reflex is where your points leak out: the number at the top teaches you almost nothing, and the review afterward is the actual studying. Here is the protocol: run a One-Sentence Autopsy on every miss, use the Rebuild to sort each one, then drop it into one of two buckets before you drill anything else.

Why the review is the studying

Taking a test measures you. It does not change you. Do ten in a row without reviewing and you have measured the same leaks ten times and fixed none of them. That is the quiet reason volume stalls out (the full version of that story is the score plateau guide): more tests re-measure what a real review would have fixed after the first one.

Two honest facts about the score itself. First, a single practice score can wobble from measurement noise alone, which is why College Board reports section scores as a range rather than a single point, so chasing the top-line number from test to test is chasing static. The signal is in the misses, not the score. Second, hours alone are not the lever: practice you never review is just more measuring.

The One-Sentence Autopsy ("careless" is banned)

For every question you missed, write one sentence naming the real reason you missed it. One sentence, plain language, before you move to the next. That is the whole tool, and it is harder than it sounds because your brain will offer you an escape hatch: "careless." "Careless" is banned. It is not a reason; it is a refusal to look. So are "silly mistake," "misread," and a bare "rushed." Each of those is a lid on the actual cause, and the actual cause is the only thing worth writing down.

Compare a banned autopsy to a real one:

  • Banned: "Careless, I knew this one."
  • Real: "I picked the choice that reused a word from the passage, but it answered a question the stem never asked." (a wrong-task miss)
  • Real: "I felt the answer was the second option, then talked myself into the fourth on review." (first-instinct override)
  • Real: "I read 'tempered' as 'strengthened' when it meant 'softened,' so I flipped the sentence's direction." (a vocab-in-context miss)
  • Real: "I did not know a colon needs a complete sentence on its left." (a rule I never learned)

Notice that the moment the sentence gets specific, it also tells you the fix. That is not a coincidence. The autopsy is the sorting step in disguise.

The Rebuild: re-solve it cold

The autopsy tells you what you think happened. The Rebuild tests it. Cover the answer choices and your previous answer, and solve the question again from scratch, as if you had never seen it. Predict the answer in your own words before you uncover the choices (that is Predict-Then-Peek, and doing it here strips out the anchoring that may have caused the miss the first time).

Two outcomes, and they mean opposite things:

  • You get it right cold. The knowledge was there. Something in your process lost the point, which makes it a behavior miss, a leak.
  • You still cannot get it, or you get it wrong again. The rule or skill is genuinely missing. It is a knowledge miss.

That is the entire function of the Rebuild: it is the instrument that sorts a miss into a bucket. Without it you are guessing whether "careless" meant "I panicked under the clock" or "I never learned this," and those two need completely different fixes.

Two buckets: knowledge or behavior

Every miss drops into exactly one of two buckets.

Knowledge bucket: a rule or skill you did not have. You could not re-solve it cold because the thing you needed was not in your head. The fix is learning, then drilling that one skill.

Behavior bucket: a leak. You had everything you needed and your process still lost the point. The fix is not more content, because you already knew the content. The fix is a process rule and catching yourself using it.

Most students find one or two buckets do most of the damage. That concentration is the good news: you are not behind on everything, you have two or three specific holes. Sort first, then decide where the hours go. Drilling random question sets before you have sorted is how people study hard and move nothing.

The 40-minute protocol

For one Reading & Writing section (both modules, 27 questions each), run four passes. The clock is a rough guide, not a stopwatch.

  1. Score it, then set the number aside (about 5 minutes). Write the score down and remember it lives inside a noise band. You are done looking at it.
  2. Autopsy and Rebuild, miss by miss (about 20 minutes). For each missed question: one autopsy sentence, then the Rebuild to sort it. Tag each miss K (knowledge) or B (behavior). Include the questions you got right but guessed on or felt unsure about, because a lucky right answer is an ungrounded point that will not survive next time.
  3. Tally the tags (about 10 minutes). Count the K-misses by skill and the B-misses by leak. One dominant bucket usually falls out.
  4. Write one next action (about 5 minutes). A single line: the one skill to drill or the one process rule to run next session. Not a list. One.

Run this today: take your last practice section and give exactly three misses the full autopsy-and-Rebuild. If even one of them re-solves cold, you just caught a behavior miss you were about to "fix" with more studying, which would never have touched it.

What to do with each bucket

Knowledge bucket. Drill that one skill in short daily sets, not another full test. If your K-misses cluster in punctuation, the fix is a handful of mechanical rules, not more volume (start with the grammar rules guide). Spacing beats cramming: three fifteen-minute sets across three days consolidate better than one long Sunday.

Behavior bucket. You cannot drill a leak away with content, because the content was never the problem. You install a rule and catch yourself using it. First-instinct override gets the Receipt Rule: change an answer only when you can point to the exact word that disproves your first pick. Anchoring gets Predict-Then-Peek. A time-sink gets the 90-Second Flag: if a question has not clicked by ninety seconds, flag it, guess, and move on (the full list of leaks and their fixes lives in its own guide).

Then re-measure. A miss is not closed when you understand the explanation. It is closed when the same question type comes back under a clock and you get it. Re-test the bucket, not the whole test.

Practice it now, free: start a practice section at simplesat.study, no signup, works as a guest. Review each miss with the autopsy and the Rebuild while it is still fresh.

Frequently asked questions

How long should reviewing a practice test take?

Budget about 40 minutes per Reading & Writing section, which is often as long as taking it. If your "review" is five minutes of staring at the score, you are not reviewing, you are re-measuring. The review is the studying, so it should cost real time.

Should I review the questions I got right?

Review the ones you guessed on or felt unsure about. A lucky right answer is an ungrounded point, and ungrounded points regress. Run the Rebuild on your unsure-rights: if you cannot re-derive the answer cold, treat it as a knowledge gap you happened to survive this time.

I review every test and keep missing the same thing. Why?

Usually because "reviewing" has meant reading the explanation and nodding, which cannot touch a behavior miss and does not count as spaced practice for a knowledge miss. Autopsy the miss to find its bucket, apply the matching fix (a process rule for behavior, short repeated sets for knowledge), then re-measure. Skip the autopsy and you keep applying the wrong fix to the wrong bucket.

Do I have to review the whole test at once?

No, and you should not. Review in one-section chunks while the test is fresh, ideally the same day. A miss reviewed cold two weeks later has lost the memory of what your brain actually did in the moment, and that memory is half the autopsy.

SAT® is a trademark registered by the College Board, which is not affiliated with, and does not endorse, this site.

← All guides