# GMAT Evaluate Questions: Test Both Outcomes | topin

On this page

## Name the conclusion before testing information

GMAC’s official Verbal description includes evaluating arguments and plans. All examples here are original teaching material, and the two-outcome method is our suggested reasoning process. The [Critical Reasoning overview](/gmat/verbal/critical-reasoning) covers the wider families; this article isolates questions asking what would be useful to determine.

Write the conclusion as one claim: “the new training caused higher sales”, “the policy will lower net cost” or “the survey supports a town-wide prediction”. Then write the evidence. The gap is the relationship that evidence has not yet established. Evaluation seeks information about that relationship.

A candidate question should be tested in both directions. If “yes” helps, what does “no” do? If “high” helps, what does “low” do? The answers need not be equally dramatic, but they should matter to whether the evidence supports the conclusion. Simply noticing that one answer seems favourable is an incomplete test.

Do not insist every useful evaluation is literally yes-or-no. A numerical result can be relevant when values above and below a threshold change the conclusion. The method is to compare meaningful possible results, not to force every answer choice into a grammatical binary.

## Evaluate a before-and-after claim with an original example

Original argument: a chain introduces a new staff-training course at its northern stores. Sales there rise 12%, while sales at the southern stores remain flat. Management concludes the course caused the increase. Which question is most useful in evaluating this conclusion?

Candidate A asks whether the northern stores also introduced a new discount during the period. Candidate B asks whether staff liked the course. Candidate C asks whether the company’s logo is recognised nationally. Candidate D asks whether the northern stores are older than the southern stores. Candidate E asks whether every manager has a university degree.

A targets an alternative explanation occurring alongside the proposed cause. If a northern-only discount was introduced, the sales comparison does not isolate training. If no such discount occurred, that particular alternative is reduced. A “no” answer does not prove training caused the gain; it still meaningfully changes the evidential situation.

B may matter to staff satisfaction, but liking the course is not the measured sales effect. D could become relevant with further information, but age alone does not identify a period-specific competing change. The best answer directly tests the gap in the given comparison rather than merely collecting another company fact.

## Use a break-even threshold for numerical evaluations

Original argument: a delivery company expects a route change to save 200 driver-hours annually, valued at £15 per hour. It concludes the change will reduce annual net operating costs. A useful question is the annual cost of implementing the new route system. The expected saving is £3,000 before that cost.

If annual implementation costs are £500, the stated numbers support a £2,500 net saving. If they are £4,000, the claimed saving does not cover the new cost. Those two possible results act differently on the conclusion. Asking the colour of the new routing screen does not establish the missing net-cost relationship.

__Original two-outcome tests__
| Candidate information        | Outcome one               | Outcome two            | Effect on the argument                |
| ---------------------------- | ------------------------- | ---------------------- | ------------------------------------- |
| Annual route-system cost     | £500                      | £4,000                 | Net saving versus net increase        |
| Concurrent northern discount | Introduced                | Not introduced         | Alternative cause versus one removed  |
| Survey group coverage        | Matches target population | Excludes a major group | Stronger versus weaker generalisation |
| Office wall colour           | Blue                      | Green                  | No stated connection to savings       |

The threshold is tied to the evidence supplied. Do not invent extra revenue or productivity gains to rescue the plan unless the prompt supplies them. The purpose is to evaluate the presented reasoning, not to design the best possible business plan. Our [assumption-negation guide](/articles/gmat-cr-assumption-negation) handles the related question of what such a net-saving claim requires.

## Test whether the measured sample fits the predicted population

Original argument: 80% of respondents to an online student newsletter say they want evening classes. A college concludes that most of all enrolled students prefer evening classes. A useful evaluation asks whether the newsletter respondents reflect the preferences of the full enrolled population.

One possible result is that comparable groups across the college have similar preference rates; that supports extending the survey result. Another is that the newsletter is read almost entirely by employed part-time students, while most enrolled students are full-time daytime students with different preferences; that weakens the extension.

A question about whether respondents understood “evening” can also matter if ambiguous interpretation affects the measured preference. Compare candidate choices against the precise gap. Do not reflexively select any answer containing “representative”, since a different option may test a more directly unresolved measurement issue in the actual prompt.

A large sample count alone does not answer every representativeness problem in this original argument. If the sample systematically omits the relevant population, adding more respondents from the same restricted group may leave the extension unsupported. The same scope discipline matters in [Reading Comprehension inferences](/articles/gmat-rc-inference-scope).

## Avoid one-direction and merely interesting choices

A common error is to choose information that would strengthen the conclusion if favourable but does little if unfavourable. For example, “do staff enjoy training?” might support participation if yes, but staff could learn and improve sales while disliking it. You have not shown that the answer settles the causal gap.

Another error is treating uncertainty as relevance. You may not know a company’s founding year, but unknown information is not automatically useful evaluation. The question asks which information matters to the reasoning, not which fact would make the story more complete.

Beware a choice that evaluates a side conclusion. If the argument concludes lower net costs, an answer about customer satisfaction may be tangential unless the prompt connects satisfaction to the cost calculation. Restate the actual conclusion whenever two plausible options seem to address different outcomes.

Also distinguish “proves the plan fails” from “helps evaluate”. A useful result may strengthen or weaken without conclusively deciding the claim. Rejecting an otherwise relevant question because a favourable answer is not a complete proof imposes a stronger standard than the task requires. Focus on its effect on support.

Original plan example: a town proposes replacing a bus route with smaller vehicles, arguing that lower fuel use will reduce total operating expense. A useful evaluation asks how many vehicles and drivers are needed to carry the existing passenger load. If the same staffing covers the demand, fuel savings are more informative; if twice as many drivers are needed, the stated saving may be offset.

The two outcomes do not require inventing exact salaries. They expose a missing cost relationship already relevant to the conclusion. By contrast, asking whether the smaller vehicles look modern addresses appearance without establishing the cost effect. If the conclusion instead concerned public approval, appearance might matter differently. The task is tied to the claim.

When a candidate question has several possible numerical answers, choose representative values on different sides of the relevant threshold. Do not compare two low costs and conclude the question is unhelpful because both support the plan. The meaningful contrast must include outcomes capable of changing the support.

## Use a two-column review record

In untimed practice, write each plausible candidate on a row with two possible outcomes and a short consequence for each. If both consequences are “no stated effect”, cut the candidate. If the consequences target different topics, check whether you have changed the conclusion halfway through the test.

Original mini-drill: a retailer predicts profit will rise after cutting prices because unit sales will rise by 30%. Useful evaluation asks about the original and new per-unit contribution and relevant costs. A unit-sales rise alone does not determine total profit. One outcome may preserve enough contribution; another may leave total contribution lower despite selling more items.

For timed practice, use concise contrasts: “discount yes → alternative cause; no → one alternative reduced”. You do not need a long imagined scenario for each answer. First eliminate off-topic choices, then compare the final contenders with the same conclusion visible.

topin’s [free full-length GMAT mock](/gmat/practice-test), marked on the official scale, can check whether this process fits your overall timing. Review mistakes by the missing outcome: did you test only “yes”, forget the threshold, or evaluate a different claim? A precise error label turns another set of questions into useful practice instead of repetition.

## Try a full GMAT mock free

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

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

## FAQs

What is the two-outcome test in GMAT CR?

For candidate evaluation information, consider two meaningful possible results and explain how each changes support for the conclusion.

Must an evaluate answer be yes-or-no?

No. A numerical result can be useful when values on different sides of a threshold change the conclusion.

Does a favourable result have to prove the conclusion?

No. It can strengthen support while leaving other uncertainty. Evaluate relevance to the reasoning, not absolute proof.

How is evaluation different from an assumption?

Evaluation asks what information would matter; a necessary assumption is a condition the reasoning requires. Match the method to the stem.

Why is an interesting company fact often wrong?

Unless it affects the evidence-to-conclusion relationship, it can add detail without helping evaluate the argument.

## Sources (checked 5 October 2026)

* [mba.com: Verbal argument evaluation and plan assessment (checked 5 October 2026)](https://www.mba.com/exams/gmat-exam/about/exam-content)

## Related articles

[All GMAT study guides ](/gmat)

* [GMATStudy plans & strategyReview and Edit decisionsUse GMAT Review and Edit with a clear bookmark system, worked correction examples and a practical review budget that protects first-pass completion.Updated 5 Oct 2026](/articles/gmat-review-edit-strategy)
* [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)