# GMAT Review and Edit: Which Answers to Change | topin

On this page

## Know what the review screen actually permits

GMAC’s structure page, checked on 5 October 2026, says the review screen appears after every question in the section has been answered, provided time remains. You may review any number of questions and edit up to three answers per section. A bookmark does not let you leave an unanswered question behind during the first pass.

That changes the practical meaning of “return later”. You must enter your best available answer before continuing. If your plan requires six unfinished solutions to be completed during review, the plan has a capacity problem before the exam starts. The review period uses the remaining section time; it is not a separate untimed stage.

Treat review as a short verification queue. A question belongs in that queue when you can say what you would check: the denominator, a sign, a qualifier, a selected response or a case excluded by the wording. “It felt hard” does not specify an action and can consume the entire reserve without improving the answer.

The general [GMAT section guides](/gmat) explain the format. This article focuses on choosing between uncertain answers once review is available. The priority rules below are our practice framework, not a scoring rule from GMAC. No claim that changing a particular answer produces a fixed number of points is justified by this framework.

## Give each bookmark a reason you can act on

On permitted note-taking materials, keep a compact reference such as “Q6 base”, “V9 scope” or “DI12 second row”. Use whatever notation you can read without decoding it again. Do not write an essay about the question; the note should direct your attention to one unresolved decision when you return.

Separate incomplete reasoning from uncertain execution. An algebra solution with a possible transcription error is different from a problem where you never found an equation. The first may be resolved by comparing two lines of working. The second may need an entirely new approach. Review priority should reflect the remaining work, not the emotional intensity of the first attempt.

Also separate answer uncertainty from entry uncertainty. You may be confident about your reasoning but unsure whether you selected the intended option. That is a useful quick check. A long question where you selected an answer after several rejected methods may be less useful to reopen if the remaining time is short.

__Suggested bookmark categories; these are editorial decisions__
| Reason    | Return action                             | Priority when time is short        |
| --------- | ----------------------------------------- | ---------------------------------- |
| Entry     | Check chosen option against final working | High if a mismatch is plausible    |
| Execution | Recheck one denominator, sign or unit     | High if the calculation is short   |
| Scope     | Compare a qualifier with the evidence     | Medium; inspect exact wording      |
| Method    | Rebuild the whole solution                | Low unless a new approach is clear |

## Three original examples of a justified edit

Example one: a product price rises from £80 to £100\. You selected 20% as the increase. During review you notice you divided £20 by the final £100\. The question asks for growth from the original price, so the correct calculation is 20/80 = 25%. This is a justified edit because you can identify and repair the precise mistake.

Example two: a passage says some surveyed managers reduced travel after a policy change. You selected an answer saying most managers did so. Review reveals that “some” was expanded to “most”, without supporting numbers. You should reconsider that choice using the passage. You should not change merely because the answer includes a strong word; a strong word can be correct when the text supports it.

Example three: a table asks whether every eligible branch exceeded its target. You checked two branches and forgot the third. The third achieved £48,000 against a £50,000 target, so the universal statement is false. Reviewing the missing row changes the evidence. This differs from redoing the same two calculations and hoping that a different answer becomes more attractive.

Each repair ends with a second check: does the corrected result answer the actual question, and did you enter it in the intended location? A sound correction can still fail if you solve for the percentage but select the currency amount, or correct the wrong response in a multi-part item. Review should close the original uncertainty, not create an unchecked replacement.

## Create a review budget without forcing a late rush

Start by observing whether you currently complete a section with considered answers. If you regularly run out of time, a large review reserve is unrealistic. Repair the first-pass bottleneck before imposing a target such as five minutes for review. Our [working-professional study plan](/articles/gmat-study-plan-working-professionals) explains how to schedule targeted repair and later verification.

For a practice section that already finishes comfortably, try reserving two minutes. At the review screen, choose one question with a concrete short action, then reassess. Two minutes is an illustrative trial, not an official recommendation. You may find that one minute works better, or that a different section needs a different reserve.

Suppose you have 120 seconds and four bookmarks. Two are short checks estimated at 20 seconds each; two require rebuilding solutions estimated at 90 seconds each. Inspecting the short checks first preserves the possibility of several useful decisions. Starting with a rebuild risks consuming the whole budget before you even inspect an obvious entry mismatch.

Those estimates are uncertain, so use them as decisions rather than promises. If a supposedly short check becomes a second full solution, reassess quickly. Keeping the first answer is a legitimate outcome. Do not spend the last seconds repeatedly switching between two choices when you cannot explain what new evidence would settle them.

## Use evidence to distinguish correction from second-guessing

Before an edit, complete the sentence “My original answer fails because…”. A wrong base, an ignored condition or an unsupported inference supplies an answer. “The other option sounds better now” does not. This test is especially useful in Verbal, where rereading an attractive distractor can make its familiar wording feel like proof.

For [Critical Reasoning assumptions](/articles/gmat-cr-assumption-negation), return to the conclusion and the role the answer must play. For numerical questions, return to the requested quantity and constraints. Your original intuition is neither sacred nor automatically wrong; the relevant question is whether the available evidence supports the chosen response.

Do not infer scoring value from question position or apparent difficulty. This method does not rank bookmarks by an imagined adaptive penalty. A question with an easily repairable error can be worth checking wherever it appears. An unfamiliar question near the beginning does not deserve unlimited review merely because of its number.

Avoid inventing certainty to use all three edits. If two justified corrections are complete and the third bookmark remains unresolved, keeping the existing response may be the best available action. The official ceiling does not require you to fill a quota. Record the uncertain item for later study rather than letting the quota drive the answer.

## Measure whether your review decisions are helping

After each practice attempt, record what changed and why. Classify each actual edit as wrong-to-right, right-to-wrong or wrong-to-wrong. Also record checks that correctly left an answer unchanged. These categories describe your decision process; their count cannot be converted directly into an official total-score change.

For example, over three practice sections you might make four edits: two corrections, one harmful change and one unresolved replacement. Inspect the harmful change first. If it came from distrusting supported wording, your next practice goal is to require an evidence sentence before changing a Verbal answer. If it came from rushed arithmetic, shorten the correction and verify it.

Use fresh material when testing the revised process. topin’s [free full-length GMAT mock](/gmat/practice-test) is marked on the official scale and gives you a full rehearsal for section completion and review. A memorised question may help you practise the interface, but it gives weak evidence about whether your bookmark decisions work on unfamiliar problems.

Finish the audit with one change to the next session: fewer vague bookmarks, a smaller reserve, or a check for entry errors before rebuilding solutions. The point is to create a repeatable decision routine. A higher mock result is useful, but knowing which habit changed makes that result more informative than an isolated score screenshot.

## Try a full GMAT mock free

Timed like test day, every section scored, every answer explained.

[Take the free mock ](/gmat/login)

## FAQs

Can I skip a GMAT question and return during review?

You must answer all questions before the review screen becomes available. Enter your best available response, bookmark it if useful, and return only if time remains.

Must I change three answers in every section?

No. Three is the maximum. Use an edit only when you have a defensible correction; leaving a supported answer unchanged is a valid decision.

How much review time should I reserve?

Test a modest reserve after you can complete the first pass. Two minutes is an example to trial, not a universal target or official rule.

Should I always trust my first answer?

No. Check the evidence. A demonstrated calculation or reading error justifies reconsideration; unease without a specific reason is weaker grounds.

Which bookmarks should I review first?

Start with plausible entry or execution errors that can be checked quickly, then inspect more demanding uncertainties if time permits.

## Sources (checked 5 October 2026)

* [mba.com: Question Review and Edit access and limits (checked 5 October 2026)](https://www.mba.com/exams/gmat-exam/about/exam-structure)
* [mba.com: official report includes review and timing insights (checked 5 October 2026)](https://www.mba.com/exams/gmat-exam/scores/official-score-reports)

## Related articles

[All GMAT study guides ](/gmat)

* [GMATStudy plans & strategyChoose section order and breakCompare all six GMAT section orders using fatigue, warm-up and mock evidence, then choose where the optional break fits your actual testing routine.Updated 5 Oct 2026](/articles/gmat-section-order-break-strategy)
* [GMATStudy plans & strategyReverse percentage problemsSolve GMAT reverse percentage and successive-change problems with original worked examples, multiplier equations and checks for discounts, margins and bases.Updated 5 Oct 2026](/articles/gmat-percent-change-reverse-percentages)
* [GMATStudy plans & strategyRates, work and unitsSolve GMAT rates and work problems by tracking units, combining constant rates and handling changing teams, with original questions and checked solutions.Updated 5 Oct 2026](/articles/gmat-rates-work-units)
* [GMATStudy plans & strategyAssumption negation testUse the negation test on GMAT necessary assumptions with original arguments, quantified statements and precise distinctions from strengthening evidence.Updated 5 Oct 2026](/articles/gmat-cr-assumption-negation)