All tips
The single pattern that separates 22s (430) from 25s (490)

DAT RC inference questions, and how to stop overthinking them

Inference questions reward boring, defendable answers. They punish creative ones. Here is how to build the discipline that fixes this for good.

inference

Inference questions are the single most under-diagnosed killer on the DAT RC section. Students miss them, blame the passage, and move on. The passage was usually fine. The technique was the problem.

Inference accuracy is the cleanest signal of where a student is on the score curve. Students under 22 (430 on the new scale) typically run sub-60 percent on inference. Students above 25 (490) are usually north of 80 percent. Almost nothing else correlates that tightly.

Inference rewards boring answers. It punishes interesting ones.

What an inference question really asks

Inference questions use language like "the passage most strongly suggests," "it can be inferred that," or "the author would most likely agree that." They are asking you to take one step beyond the literal text. Not three.

Most wrong answers on inference questions are technically true, technically supported by some interpretation of the passage, but require more inference than the test rewards. The right answer almost always feels too obvious to be the answer. That is the trap.

The two failure modes

Inferring too far

You read a passage about coral reefs and the author mentions that overfishing has reduced predator populations. The question asks what the author would most likely agree with. You see a choice that says "stricter international fishing quotas would help reef ecosystems recover." Reasonable. The passage said overfishing is bad. So this must be right.

Wrong. The passage said overfishing reduces predator populations. It did not endorse a specific policy. The right answer is something like "ecosystems are sensitive to disruptions in predator populations." Boring. Defendable. Right.

Not inferring at all

The opposite failure. You scan the passage looking for the literal sentence that matches an answer choice. None match exactly. You guess.

Inference questions deliberately do not have a literal-match answer. The right answer is a paraphrase or a one-step extension. If you find a literal match, it is usually a trap.

The trap
Picking a choice that requires you to assume something the passage did not say. If you need a 'because of course' or 'obviously this implies' to defend the answer, you are inferring too far.
The fix
Ask yourself: would the author definitely agree with this, or would they just probably agree? Definitely beats probably. Boring beats interesting.

The defendability test

Before locking an inference answer, ask yourself this question: if a strict literature professor read the passage, would they grade my answer as well-supported? Inference answers should be defendable from the text in one or two sentences. If you need a paragraph to defend it, you are wrong.

What to do this week

  1. 1
    Pull your last 3 timed sections.
    Find every inference question. Tag each one as right or wrong.
  2. 2
    For every wrong inference question, identify which failure mode hit you.
    Inferred too far, or did not infer at all. Most students show one pattern strongly.
  3. 3
    Take 20 inference questions from your next round of practice.
    Before locking each answer, write down in one sentence why the passage supports that choice. Make it a habit.
  4. 4
    Track your inference accuracy.
    Move it from sub-60 percent toward 80 percent and your overall score will move 2 to 3 points by itself.

See all 7 question patterns for the broader framework.

Related: DAT RC tone questions · Why you are bad at DAT RC