The Search-and-Destroy Method for DAT Reading Comprehension

Short answer: the search-and-destroy method for DAT reading comprehension means you skip the slow, linear first read of the passage and go straight to the questions, then hunt the passage for the specific lines each question points to. It's fast and it works great on detail and fact-recall questions. It backfires on inference, tone, and main-idea questions, which is why the real skill isn't "always search-and-destroy" or "always read first" — it's knowing which question in front of you needs which approach.

We're the founders of DATPractice, and we both scored 97th-plus percentile on the DAT (legacy-scale 25 AA with a 30 in organic chemistry, and 27 AA with a 29 TS). We're now at the #1 dental school in the world, and Reading Comprehension is the section where students argue the most about "the right way" to do it. There isn't one right way. There's a right way per question type, and we'll break down exactly how to run it.

What the search-and-destroy method actually is

Search-and-destroy is a question-first approach. Instead of reading the whole passage top to bottom before touching question one, you read the question stem first, then go "search" the passage for the exact lines that answer it, and "destroy" it — pick the answer and move on.

The logic is simple: DAT RC gives you 50 questions across three passages in 60 minutes — roughly 20 minutes and 16-17 questions per passage. A full, linear read of a 900-1,000 word science passage can eat 3-4 minutes before you've answered anything. Search-and-destroy reclaims that time by only reading what a question actually requires.

In practice: read the question stem, note the keyword or concept it names, scan the passage for that keyword or a close synonym, read just the sentence or two around the hit to confirm, answer, repeat. Done well, this is genuinely fast. Done badly — searching for the wrong keyword, or searching when a question has no locatable "hit" at all — it can burn more time than just reading the passage would have.

Should I read the passage first or go straight to questions on the DAT?

This is one of the most argued-about questions in DAT prep forums, and most of the arguing happens because people treat it as an all-or-nothing choice. It isn't. The honest answer: it depends on what the question is actually asking you to do.

Some DAT RC questions are locatable — the answer lives in one identifiable spot, and your job is just to find it. Those reward going straight to questions. Others are synthesized — the answer requires holding the passage's overall argument or the author's attitude in your head, with no single sentence containing it. Those punish going straight to questions, because there's nothing to "search" for.

So the real rule isn't "read first" or "questions first" as a blanket policy for every passage. It's a per-question-type call you make in the moment, and it's fast once you've drilled the pattern.

When search-and-destroy works well

Search-and-destroy is at its best when the answer is anchored to specific, findable text:

• Detail and fact-recall questions. "According to the passage, what temperature does X occur at?" has one correct location, and scanning for the number gets you there fast.
• Definition-in-context questions. "As used in paragraph 3, the word 'labile' most nearly means…" only needs the sentence containing that word.
• "According to the passage" and "the passage states" questions. These explicitly ask you to locate, not interpret — exactly search-and-destroy's job.
• Questions that name a specific paragraph, figure, or study. If the stem tells you where to look, reading everything else first is wasted time.

On a passage stacked with these question types, going straight to questions saves real minutes you'll want later in the section.

When search-and-destroy backfires

The method falls apart on questions with no single findable "hit," because there's nothing to search for — you have to have already understood the passage:

• Inference questions. "The author most likely believes…" requires connecting ideas across the passage, not locating one sentence. Scanning for a keyword often lands you on a technically-true detail that isn't the actual inference being tested.
• Tone and attitude questions. Tone builds up across word choice throughout the whole passage. You cannot search for "the author's tone." See our full breakdown of how to answer DAT RC tone questions for this specific failure mode.
• Main idea and purpose questions. "The primary purpose of this passage is…" needs the shape of the whole argument, not one line.
• "Would the author agree/disagree" questions. These test whether you understand the author's position well enough to extrapolate it — impossible to fake by scanning isolated lines.

The trap we see most often: a student trains search-and-destroy on easy passages, gets fast and confident, then applies that same speed to an inference-heavy passage where it doesn't work. They pick the answer that "sounds like something from the passage" instead of the one actually supported by its logic — and those aren't always the same choice.

The decision rule: match the method to the question type

Here's the rule we'd give a friend studying for the DAT, instead of a dogmatic "always do X":

1. Skim first, always — but fast. Spend 45-75 seconds getting the passage's topic and rough argument. Not a full read, just enough that inference and tone questions aren't a blind guess.
2. Then go straight to questions. Decide in about one second whether each stem is a "locate" question (detail, definition, "according to the passage") or a "synthesize" question (inference, tone, main idea).
3. For "locate" questions, search-and-destroy. Scan for the keyword, confirm with the surrounding sentence, answer, move on.
4. For "synthesize" questions, lean on your skim. If unsure, re-read topic sentences rather than hunting for a single line — the answer isn't in one line.
5. Bank the "locate" questions first if a passage feels dense. Easy points first, then real thinking time on the harder synthesis questions.

This hybrid is faster than a full linear read on every passage, and more accurate than pure search-and-destroy on every question. Our guide to the DAT reading comprehension format, question types, and scoring breaks down how the 50 questions split across these categories.

Why practicing this on isolated passages doesn't prove it works

Here's where a lot of students go wrong after they understand the decision rule: they practice it on a handful of untimed, isolated passages and conclude it's "working." That doesn't actually test the method.

Search-and-destroy's whole value proposition is speed under pressure. If you're not timing yourself, doing one passage in isolation instead of three back-to-back, and the passage isn't calibrated to real DAT difficulty, you're not finding out whether the technique holds up — you're doing untimed reading practice with extra steps.

The real test is running this on a full 60-minute, three-passage, 50-question section, at real DAT difficulty, back to back with no do-overs. That's the only setting where you'll find out if your keyword-scanning is actually fast enough, if you're correctly identifying "locate" versus "synthesize" questions under a ticking clock, and whether fatigue on passage three makes you sloppier about which method you're even using.

This is also why we don't hand students one rigid "always do it this way" script for RC. The DAT rewards flexibility, and the only way to build that instinct is repetition on passages hard and timed enough to force real decisions. For the fuller strategy picture, our best DAT reading comprehension strategies guide covers the rest of the toolkit, and our skimming and keyword-highlighting techniques guide pairs directly with the skim step above.

Common mistakes that ruin search-and-destroy

• Searching for the wrong keyword. Stems sometimes use a synonym instead of the passage's exact word — scan for the literal stem word and come up empty, and you've wasted time.
• Applying it to every question type without checking. The single biggest cause of missed inference and tone questions is treating them like locate questions.
• Skipping the skim entirely. Going in cold makes even locate questions slower, since you have no mental map of where things live.
• Not confirming surrounding context. Finding the keyword isn't the same as finding the right answer — read the sentence before and after it, too.