Evaluate questions ask a strange-sounding thing: not what helps or hurts the argument, but what you would need to know to judge it, the missing measurement on which the conclusion’s fate turns. They are rare, they unsettle students who’ve never seen one, and they are nearly free points for anyone who knows the one tool built for them: the Variance Test. Run it correctly and the credited answer identifies itself.

What an Evaluate Question Actually Tests

Gap-awareness, one notch more abstract than weaken or strengthen. Those types hand you facts and ask their effect; evaluate hands you questions and asks which question’s answer would swing the argument. The credited choice names the issue whose resolution could move the conclusion in either direction, the live wire. Wrong answers name issues whose every possible answer leaves the argument exactly where it stood.

How to Identify Them

Stems: “The answer to which one of the following questions would be most useful in evaluating the argument?” or “most helps in determining whether the conclusion is warranted” or “it would be most important to determine which one of the following.” The choices themselves are usually phrased as questions, the type’s visual signature.

The Lovare Method, Four Steps

One: state conclusion and support plainly. Two: name the gapsame diagnosis you’d run on a weaken question; the argument’s vulnerable joint is where useful information lives. Three: pre-phrase the live question. “What I’d want to know is whether…” Four: run the Variance Test on finalists. Give each candidate question its two extreme answers, a strong yes and a strong no. If one extreme strengthens the argument and the other weakens it, the question is live and credited. If both extremes leave the conclusion unmoved, the question is decoration.

A Worked Example (Original)

Meridian Farms argues: “Our drip-irrigation pilot field used 30 percent less water than our sprinkler fields this season while producing the same tomato yield. Converting all fields to drip irrigation will therefore cut our water use dramatically without hurting harvests.”

The gap: was the pilot field representative? Candidate (A): “Did the pilot field’s soil retain moisture better than the soil in most of Meridian’s other fields?” Variance Test, answer yes: the savings may not transfer; the argument weakens. Answer no: the pilot looks representative; the argument strengthens. The needle moved both ways, live wire, credited. Candidate (B): “Did neighboring farms also consider drip irrigation this season?” Yes: nothing. No: nothing. Dead wire, however topical. The test is mechanical, which is precisely the gift.

The Variance Test, Properly

Two errors corrupt it. First, soft extremes, testing “somewhat” answers instead of the poles; always push each candidate question to its strongest yes and strongest no, because only the poles reveal swing. Second, single-direction satisfaction, crediting a question whose “yes” weakens but whose “no” merely… does nothing. A genuinely useful question moves the argument both ways across its answers; one-way questions are usually narrower traps orbiting the real issue.

Withheld Tip: evaluate questions are weaken questions in disguise, pre-phrase the weakener you’d look for, then find the answer choice asking whether that weakener is true. Students fluent in the three causal doors can convert that fluency here at a one-to-one exchange rate.

The Traps, Named

The Dead Wire asks a question no answer to which moves the conclusion. The Background Check requests context, history, popularity, cost, orthogonal to the logic. The One-Way Street swings the argument in a single direction only. The Pre-Answered asks something the stimulus already settled.

How to Train It: The Loop Applied

Because the type is rare, drill it in concentrated sets rather than waiting for organic encounters, ten evaluate questions in one sitting builds more pattern than three months of scattered ones. Log misses by which Variance Test error occurred: soft extremes or one-way satisfaction. A clean blind-review record with timed misses means the test exists in your toolkit but not in your reflexes; rehearse the poles aloud, “strong yes… strong no…”, until the ritual survives the clock. In Priority Stack terms, a rare type with a mechanical solution is exactly what you perfect after the high-frequency core is stable: cheap insurance, never the main investment.

If You Only Do 3 Things

1. Pre-phrase the gap exactly as you would on a weaken question.
2. Run the Variance Test at the poles, strongest yes, strongest no, on every finalist.
3. Demand two-way swing; one-way movement is a trap wearing usefulness.

Evaluate Questions: Quick Answers

How rare are evaluate questions, really?

Among the least frequent Logical Reasoning types, often one per test, sometimes none. Their cost-benefit is unusual: minutes to master, points that arrive as pure surplus when they appear.

Is the Variance Test ever wrong?

When run at true extremes with both directions checked, it tracks the credited answer with near-perfect reliability, the failures in practice are soft extremes and one-way satisfaction, both procedural.

What if two questions both seem to swing the argument?

Compare swing sizes. One question typically moves the conclusion’s core support; the other jostles a side detail. The credited answer is the bigger lever, re-anchor on the main gap and the tie breaks.

Related Playbooks

• LSAT Weaken Questions: The Complete Guide
• LSAT Strengthen Questions
• Logical Reasoning: The Complete Strategy Hub
• The 12-Week LSAT Study Schedule

Lovare’s Take

Evaluate questions panic students exactly once. After the Variance Test, they become the section’s most honest transaction: a mechanical procedure, applied at the poles, returning the answer. Rare points, fully purchasable in an afternoon, the kind of trade a disciplined Priority Stack never leaves on the table.