Reviewing the LSAT Questions You Got Right — But Slowly

A question you answered correctly in two and a half minutes and a question you answered correctly in fifty seconds look identical on a score report. They are not the same event. The first tells you something the second does not: your method for that question type is not yet reliable, and you compensated by spending time you will not always have.

Reviewing only misses means this signal is discarded. And it is a large signal — on most sections, the questions that consumed disproportionate time outnumber the questions actually missed. A student reviewing a section with three wrong answers may have six or seven slow-but-correct questions, each pointing at a specific gap.

The reason this matters more than it sounds: those slow questions are the mechanism behind the misses. Time spent grinding question 9 is time unavailable at question 23, so a miss late in the section is often produced by an unreviewed struggle early in it. Reviewing the late miss in isolation studies the symptom.

This is a different review filter than the one most methods use, and the distinction is worth being precise about. Blind review asks you to flag questions you felt unsure about, then rework them untimed. That filter is confidence, and it catches a real category — questions where you guessed and happened to be right.

Elapsed time catches a different and partially non-overlapping category: questions you were entirely confident about and still took ninety seconds too long on. Confidence and speed come apart in both directions. You can be certain and slow, which usually means your process has redundant steps. You can be fast and uncertain, which usually means you are pattern-matching without verification.

The two middle rows are the ones this review pass is built to surface, and neither shows up on a score report.

You need per-question times, or a usable approximation of them. If you practice on a platform that records time per question, the data already exists. If you work on paper, the low-effort version is to note the clock reading at questions 5, 10, 15, 20, and at the end, which gives you time per five-question block and localizes the damage well enough to act on.

A simpler manual method that requires no timing infrastructure: while working, put a small mark next to any question where you felt yourself settle in and work rather than move through it. You will remember at the time; you will not remember afterward. The mark takes half a second and does not disturb your pace.

Set a threshold rather than eyeballing it. Roughly ninety seconds per Logical Reasoning question is a reasonable working average, so flag anything that ran meaningfully past it. The exact number matters less than applying it consistently, because the value is in comparing your own questions to each other over time.

The wrong question is "was I right?" You already know. The useful question is "where did the time go?" — and the answer is almost always one of a small number of places.

Re-reading the stimulus. If you read it more than twice, the first read was passive. The fix is not to read faster but to read once with the intent of being able to state what the argument is doing before moving to the stem.

Cycling through answer choices repeatedly. If you visited choice B four times, you were not eliminating — you were re-considering. Elimination requires a stated reason. "I don't like it" is not a reason, and an answer eliminated without one gets revisited, which is where the time goes.

Stalling between two finalists. This is its own failure mode and it has a decision rule: prefer the answer that requires you to assume less. If you spent forty seconds oscillating and then picked correctly, you got the point without the method, and next time the coin may land differently.

Reconstructing a rule you have not internalized. If you paused to re-derive how to translate "unless," the question was not slow — your conditional translations are. That is a drill target, not a review note.

Write the answer down in a sentence. "Spent 40 seconds re-reading because I skipped the pivot word" is actionable. "Was slow" is not.

Suppose a section produced two misses and five flagged slow questions. Standard review handles the two misses. Here is the rest of the pass.

Sort the five by question type. If three of them are Necessary Assumption, you have found something a score report would never show you: your necessary assumption method works and is slow, probably because you are running the negation test on all five choices rather than on the two survivors. That is one specific, fixable habit, and it was invisible in a section you scored well on.

If the five are spread across five different types, the cause is more likely general — usually the re-reading pattern above, or a reluctance to commit to an elimination.

Then check placement. If four of the five sat in the first twelve questions, the two misses at the end were probably manufactured by that early time loss, and fixing the slow questions fixes the misses without any work on the missed question types at all.

Finally, redo two or three of them a week later, timed, and see whether the time came down. This is the only part of the process that confirms the fix worked. Note that you are checking whether the reasoning is faster, not whether you remember the answer — pick questions old enough that recall is unlikely, or focus on how quickly you can justify each elimination rather than on which letter you land on.

Treating slowness as a stamina or focus issue. It occasionally is, but the far more common cause is a method with unnecessary steps, and the two have opposite fixes. Look at the specific questions before concluding you need more endurance work.

Trying to fix slowness by going faster. Speed on the LSAT comes from resolving questions more decisively — eliminating with reasons, committing to a diagram, using a rule instead of oscillating — rather than from reading or thinking at a higher rate. Deciding to hurry produces misreads on questions you already understood.

Reviewing every slow question. If nine questions ran long, review the three worst and look for what they share. The goal is a pattern, not a complete accounting.

Skipping this pass on a good section. A section you scored well on with heavy time pressure is exactly where this review has the most to offer, because the errors have not surfaced as misses yet.

Recording the observation without a next action. "Slow on Parallel Reasoning" is a finding. "Drill ten Parallel Reasoning questions focusing on matching structure before reading answers" is a plan.