Chapter 15: Evaluate the Argument

The job: Fix It or Break It. Find the question whose answer would most help you decide whether the argument holds.
Frequency: roughly zero to one per section (rare). Difficulty: medium.

What this question type asks

Evaluate questions hand you a flawed argument and ask which choice it would be most useful to know, or most relevant to investigate, in order to judge that argument. The choices are usually phrased as questions: “Whether the new policy was tested in cold climates,” “How many of the surveyed workers had children.” You are not asked to strengthen or weaken the argument directly. You are asked which piece of information, once you had it, would let you do either.

That makes Evaluate a close cousin of three types you already know. The argument has a gap, an unstated assumption bridging premise to conclusion (see Chapter 11, Necessary Assumption). The right answer points straight at that gap. So Evaluate is really Strengthen (Chapter 13) and Weaken (Chapter 14) bundled into one: the correct choice is the question that could strengthen the argument if it came out one way and could weaken it if it came out the other.

Because the flaw is always an assumption error, everything you know about gaps applies. Causal leaps, surveys that may not be representative, predictions that assume the future matches the past: each one suggests the very question worth investigating. Find the assumption, and you have usually found the answer before you read the choices.

How to recognize it

“The answer to which of the following questions would be most useful in evaluating the argument?”
“Which of the following would it be most helpful to determine in order to assess the argument?”
“The answer to which of the following questions would be most relevant to evaluating the conclusion drawn above?”
Choices are typically written as questions (“Whether…”, “How many…”, “Did…”).

Do not confuse Evaluate with two look-alikes. A Strengthen or Weaken stem asks you to pick a statement that helps or hurts; an Evaluate stem asks you to pick a question whose answer would help you judge. And an Assumption stem asks for the unstated premise itself; Evaluate asks for the issue you would investigate to test that premise. Same underlying gap, different deliverable.

The method (step by step)

Read the stem first so you know you are evaluating, then read the stimulus for its argument skeleton: premises, conclusion, and the gap between them (the universal approach, Chapter 7).
Name the assumption. State in plain words the thing the author is taking for granted. This is your prephrase. If the argument is causal, the assumption is usually that the effect is not fully explained by an alternative cause and that the causation is not reversed (Chapter 5). If it is a survey or prediction, the assumption is usually that the sample or the past is representative. In everyday terms, the assumption is the hidden if the argument rests on: “the new lighting is why the museum’s attendance rose” only holds if a blockbuster exhibit did not open the same month.
Prephrase the question. Turn that assumption into something you would want to look up. “Could something other than the drug explain the recovery?” “Were the people surveyed typical of the whole city?”
Eliminate fast, then apply the Variance Test to your keepers. This is the signature technique for the type, and you should not waste it on every choice. First cut the obvious losers (off-topic, already-granted premises). For each choice that survives, run the test below.

The Variance Test

For a surviving choice, take the question it raises and plug in two opposite extreme answers to it. Then ask what each does to the conclusion.

If one extreme answer helps the argument and the other hurts it, the question is relevant. The conclusion varies with the answer, so knowing it matters. This is your correct choice.
If the conclusion is unaffected either way, both extremes leave it exactly where it was, the question is irrelevant. Kill the choice.

The test is mechanical and decisive. Suppose the conclusion is “the fertilizer caused the higher yield” and a choice asks “Whether the treated field received more rainfall than the untreated field.” Yes, far more rainfall: now rain might be the real cause, and the argument weakens. No, identical rainfall: an alternative cause is ruled out, and the argument strengthens. The answer swings the conclusion both ways, so the choice is correct. Now try “Whether the fertilizer was expensive.” Yes, very expensive and no, cheap both leave the causal claim exactly as strong as it was. No variance, so the choice is wrong.

Worked example

A regional health authority introduced a free flu-shot program at workplaces last winter. At the companies that took part, employee sick days dropped by 18 percent compared with the previous winter. The authority concludes that offering flu shots at the workplace is an effective way to reduce employee absences.

The answer to which of the following questions would be most useful in evaluating the argument?

(A) How much did the health authority spend per employee on the flu-shot program?
(B) Did the participating companies also change their sick-leave policies last winter?
(C) What percentage of employees at the participating companies worked full time rather than part time?
(D) Was last winter’s flu season milder overall than the previous winter’s across the entire region?
(E) Do employees generally prefer receiving flu shots at work rather than at a clinic?

Find the gap. The premise is a correlation: companies got the program, and sick days fell 18 percent. The conclusion is causal: the flu shots reduced the absences. The author assumes the shots, and not some other change, produced the drop, and that the comparison year was otherwise comparable. That is a textbook causal assumption (Chapter 5). Prephrase: “Was there some other reason sick days fell last winter?”

(D) is correct. Run the Variance Test. Yes, the whole region had a much milder flu season: then sick days probably fell everywhere, program or no program, and the argument collapses. No, the season was just as severe as the year before: then an alternative explanation is ruled out and the causal claim stands stronger. Opposite answers push the conclusion in opposite directions. This is exactly the alternative-cause question the gap demanded.

(A) is wrong. Cost is out of scope. Very expensive and very cheap both leave untouched the claim that the shots worked. Effectiveness is the conclusion, not cost-effectiveness. No variance.

(B) shows variance but loses to (D) on scope. A sick-leave policy change would be an alternative cause, so this question is genuinely relevant. But it probes only an internal change at the participating companies, while (D) tests the broader, more likely confound: a region-wide change in the flu season itself, which would explain the drop even if nothing about the companies changed. When two choices both show variance, pick the one that addresses the most central alternative explanation. Here that is (D). (If (D) were absent, (B) would be the answer.)

(C) is wrong. The full-time/part-time mix describes the workforce, but plug the extremes against the conclusion. Almost everyone worked full time and almost everyone worked part time both leave the reported 18 percent drop exactly as it was, and the question is what caused that drop. Without a comparison to the prior winter or a reason the mix changed, this does not distinguish flu shots from policy changes or a mild season. No variance on the causal claim.

(E) is wrong. Employee preference is off-topic. The conclusion is about whether the program reduces absences, not whether employees like it. They love it and they hate it leave the effectiveness claim untouched. No variance.

Trap patterns

No-variance / out of scope. The most common trap. Both extreme answers leave the conclusion where it was. Tells: cost, popularity, preference, or any topic that never connects to the premise-conclusion link. (Choices A and E above.)
Already granted. The question probes a premise the argument already takes as settled. There is nothing to determine, because the stimulus stipulated it. Investigating it cannot move a conclusion that was built on it.
Topical but irrelevant. The choice uses the argument’s vocabulary (here, “flu shot,” “employees”) so it feels on-point, but its answer does not bear on the gap. (Choice C: it is about the program, yet tells you nothing about cause.) Run the test rather than trusting the topic. Everyday version: arguing over whether a new coaching staff turned the team around, the question “how many players are on the roster?” is full of team vocabulary yet tells you nothing about whether the coaching caused the turnaround.
One-directional. The choice’s answer could help the argument but never hurt it (or vice versa), so it behaves like a one-sided Strengthen or Weaken rather than a true Evaluate. A real Evaluate answer must be able to cut both ways.

Drill

For each, find the gap, then use the Variance Test to choose the question most useful in evaluating the argument.

1. A bookstore moved its café from the back of the store to the front entrance. In the following month, total book sales rose 12 percent. The owner concludes that placing the café at the entrance boosts book sales.
Which would be most useful to determine?
(A) Whether café revenue also rose after the move.
(B) Whether a popular new release went on sale the same month the café moved.
(C) Whether customers found the new café layout attractive.
(D) How much it cost to relocate the café.
(E) Whether the store’s competitors also have in-store cafés.

2. A study found that adults who drink herbal tea daily report lower stress levels than adults who do not. A wellness columnist concludes that drinking herbal tea daily lowers stress.
Which would be most useful to determine?
(A) Whether herbal tea is more expensive than ordinary tea.
(B) Whether the herbal-tea drinkers in the study also tended to have calmer daily routines than the non-drinkers.
(C) How many cups of herbal tea the daily drinkers consumed.
(D) Whether the study’s participants enjoyed the taste of herbal tea.
(E) Whether non-drinkers had ever tried herbal tea in the past.

3. A city predicts that a new express bus line will cut downtown traffic, because a survey of current car commuters found that 40 percent say they would switch to the express bus if it existed.
Which would be most useful to determine?
(A) Whether the city can afford to operate the new bus line.
(B) Whether the express bus will run during rush hour.
(C) Whether commuters who say they would switch actually do so when a new transit option becomes available.
(D) Whether the proposed buses would be electric or diesel-powered.
(E) Whether other cities have built similar express bus lines.

Drill answers

1. (B). Causal argument: café placement caused the sales rise. The gap is an alternative cause for the 12 percent jump. Yes, a popular new release launched that month: the release, not the café, may explain the rise (weakens). No new release: that alternative is ruled out (strengthens). Variance, so correct. (A) café revenue says nothing about why book sales rose. (C) attractiveness does not tell you what caused the sales. (D) cost is irrelevant to whether the move worked. (E) competitors’ cafés do not bear on this store’s sales change.

2. (B). Correlation (tea drinkers report less stress) to causation (tea lowers stress); the classic gap is a third factor causing both. Yes, the drinkers also had calmer routines: a common cause could explain the link (weakens). No, their routines were just as hectic: that alternative is ruled out (strengthens). Variance, so correct. (A) cost is out of scope. (C) quantity does not distinguish tea-as-cause from a third factor. (D) taste is off-topic. (E) past trials do not bear on whether daily drinking lowers stress.

3. (C). The prediction assumes stated intentions match real behavior, that the 40 percent who say they would switch actually will. Yes, stated intentions reliably match behavior: the prediction holds (strengthens). No, people routinely overstate willingness to switch: the figure is inflated and traffic may barely change (weakens). Variance, so correct. (A) affordability does not bear on whether the line would cut traffic if it ran. (B) rush-hour service is tempting, but nothing says the line skips rush hour, and running at rush hour leaves the say-versus-do gap untouched, closer to an already-granted assumption than live variance. (D) vehicle type may matter for emissions or operating choices, but it does not address whether car commuters who say they would switch will actually do so. (E) other cities do not tell you what these commuters will do.

Key takeaways

Evaluate asks for the question whose answer would most help you judge the argument, not a statement that helps or hurts. The right answer points at the argument’s key assumption.
It is Strengthen and Weaken fused: the correct choice’s answer could firm the argument up one way and tear it down the other.
The Variance Test: plug two opposite extreme answers into the question a choice raises. Opposite effects on the conclusion means relevant and correct; no effect either way means out of scope and wrong. Apply it only to your remaining keepers.
Most Evaluate arguments are causal, so the question worth asking is usually about an alternative cause, reversed causation, or a confounding third factor (Chapter 5).
Kill choices about cost, popularity, preference, or a premise already granted. They fail the Variance Test because the conclusion does not move either way.
If two choices both show variance, pick the one that addresses the most central alternative explanation.

Put it to work: drill Evaluate questions in the practice platform, free during early access, with every answer choice explained.