Inference Questions: The Right Answer Is BORING

Every week I watch the same thing happen. A sharp Test Taker reads an Inference question, knocks out three answers fast, and stares at the last two. One of them sounds like something a consultant would say in a board meeting. The other sounds like something a 9 year old would say after reading the passage out loud. They pick the consultant. They get it wrong. Then they tell me the question was 'tricky.'

It wasn't tricky. It was BORING. And on these questions, boring is exactly what the GMAT pays for.

What an Inference question actually asks

On the GMAT, an Inference question asks what MUST be true based on the text you were given. Not what is probably true. Not what a smart person would guess. What HAS to be true if every sentence in the passage is true. You'll see these in Critical Reasoning (a short argument followed by one question) and in Reading Comprehension (a longer passage followed by several questions). Both live in the Verbal section, which right now gives you 36 questions in 65 minutes, as laid out on GMAC's GMAT format page. The wording changes: "must also be true," "can be properly inferred," "most strongly supported." The job does not.

Here's the part people fight me on. In normal life, an inference is a leap. You see wet streets and you decide it rained. On the GMAT, that leap is a TRAP. Maybe a street cleaner went by. The GMAT's version of an inference is the tiny step nobody can argue with: the streets are wet. That's it. That's the whole prize.

Why the clever answer loses

The people who write these questions know you're smart. So they write wrong answers that reward smart people for going one step too far. A wrong answer will take the facts and add something: a cause, a prediction, a piece of advice, or a comparison the passage never made. It feels like insight. It's actually NEW information, and new information can be false.

The right answer usually does the opposite. It combines two facts you were already given, or it pulls out the obvious math the passage handed you. When you read it, your gut says "well, obviously." That reaction is a GOOD sign, not a reason to keep hunting.

Try one before you read another word.

Stop and look at what you just did. You didn't find the right answer by being brilliant. You found it by being stubborn: writing the facts down, predicting the obvious, then attacking every choice with a 'say...' world. That's the same instinct that wins in Quant when you TEST VALUES instead of grinding algebra. It's a process, and a process can be trained.

The 5 mistakes I see most (and what to do INSTEAD)

1. Picking the answer that 'goes somewhere.' Causes, predictions and big conclusions feel like the point of the passage. On Inference questions they're almost always the bait.

   INSTEAD: Ask one question of every choice: "Does this HAVE to be true?" Not "is this reasonable," not "would an expert agree." MUST.

2. Reading "most strongly supported" as "most interesting." That phrase gives you a little room, but not much. The supported answer is still the one closest to the text.

   INSTEAD: Treat "most strongly supported" almost exactly like "must be true." Pick the answer that needs the fewest extra assumptions.

3. Doing it all in your head. You read the passage, read the answers, then go back and re-read the passage. Twice. Your pacing 'problem' isn't a clock problem. It's a notes problem.

   INSTEAD: Write the facts as short notes while you read the first time. Two or three lines. Then you test answers against your pad, not your memory.

4. Ignoring strong words. Words like ALL, ALWAYS, MOST, CAUSED, SHOULD and ONLY are claims the passage has to back up. They usually can't be backed up.

   INSTEAD: When a choice uses a strong word, look for the exact sentence that proves it. No sentence, no answer. Meanwhile, weak words like "some" or "at least one" are your friends. They're easy to prove.

5. Trying to prove the right answer instead of killing the wrong ones. If you're hunting for the answer that 'feels right,' you'll land on the one that feels SMARTEST.

   INSTEAD: Build a quick world that makes each choice false, like we did with (C) above. The survivor wins. Usually it's the dullest one on the screen.

Reading Comp plays by the same rules

In Reading Comprehension, Inference questions hide behind phrases like "the author implies" or "the passage suggests." Same job. The right answer is a small, safe restatement of something the author actually said, often in different words. The wrong answers go past the author, or swap in what YOU know about the topic. If you nailed down the main idea before the questions, you already have the fence. Anything outside it is out.

Why this matters more, not less

GMAC has announced a shorter GMAT Focus Edition for late 2023, and its Verbal section drops Sentence Correction. Whichever version you end up taking, think about what that means: when the grammar questions go away, more of your Verbal score rides on reading questions like these. Getting comfortable with boring answers isn't a side skill. It's becoming the core of the section.

And let's be honest about where your points are. You're probably not missing Inference questions because they're impossible. You're missing them because you talked yourself OUT of the right answer. That's a gettable point you handed back. Do that four or five times across Verbal and you've given away real score for free.

Your challenge for today

Pull 10 Inference questions you've ALREADY done, from your own work or GMAC's official practice materials. Don't look at your old answers. For every wrong choice, write ONE sentence on paper describing a world where the passage is true and that choice is false. For the right choice, write the exact sentence or math that forces it.

I'll bet at least one of those 10 has a 'clever' answer you picked last time, and I'll bet you can't write a single sentence that saves it. Find that one. Look at it. That's the habit you're training out of yourself, starting tonight.