simplesat

Behavioral leaks

The Receipt Rule: Never Change an Answer Without Proof

Updated July 6, 2026

When should I change an answer on the SAT?

Only when you can point to the exact word that proves your first pick wrong — the Receipt Rule. A feeling that another choice sounds better is not a receipt; a concrete misread is. The rule turns an unlimited question, am I sure, into a binary one: do I have a receipt, yes or no.

Two minutes left in the module and question 14 suddenly looks wrong. You stare, you doubt, you switch — and the score report says your first pick was right. Again. This page is the rule that ends that loop.

First, the honest science — you might not need this page

The folk advice on answer-changing is wrong in both directions, so start with what the research actually says. The "first instinct fallacy" literature finds that on average, students who change answers switch wrong-to-right about twice as often as right-to-wrong. "Never change your answers" is a myth; for most people, reviewing helps.

But that's an average, and it hides a minority with the reverse pattern: strong first reads, and second-guessing that switches them off. The research does not say who is in that minority. Only your own answer sheet can. (Why the leak exists at all, and why wrong answers are built to survive a doubting second look, is covered in the first-instinct override guide.)

The grade-twice self-test: run this first

The self-test: on your next two or three practice sections, write your first pick on scratch paper before you allow yourself any change. Afterward, grade each section twice — once as if you had kept every first pick, once with your final answers.

  • Finals win: your reviewing works. Keep it. Forcing the Receipt Rule on yourself would cost you the wrong-to-right changes you're earning.
  • Firsts win: you have the leak, and everything below is for you.

The Receipt Rule

You may not change an answer unless you can point to the exact word that proves your first pick wrong.

Not a feeling that another choice sounds better. A word — in the passage, the sentence, or the question stem — that you can put your finger on and say: this is what I missed the first time.

The rule works because it converts an unlimited question ("am I sure?") into a binary one ("do I have a receipt — yes or no?"). A well-built wrong answer can manufacture doubt forever. It cannot manufacture a word that isn't there.

Receipts vs. vibes

Receipts — legal grounds to switch:

  • A word you misread: "I read until as unless."
  • A rule your first pick breaks, named: "the subject is committee — singular — so were is wrong."
  • A question-stem detail you skipped: "it asks what the critics claim, not what the author claims."
  • A mechanical failure: your pick leaves a comma holding two complete sentences together.

Vibes — illegal grounds:

  • "The other one sounds more sophisticated."
  • "This feels too easy for a question this late."
  • "I've picked C three times in a row." A letter pattern is not evidence.
  • "I'm just not sure anymore." Not being sure is the default state on a hard question, not information.

Two switches, worked

Legal. The sentence: "The board of directors are expected to announce a decision on Friday." Your first pick kept are — it sounded fine next to directors. At review you find the real subject: board. A group noun, and groups are always singular (the Liars List: each, every, everyone, anyone, a team/company/group). A word, a nameable rule — a receipt. Switch to is.

Illegal. The sentence: "The senator's response was measured: she acknowledged the criticism without conceding the argument." The question asks what measured most nearly means. Your first pick: restrained. At review, calculated starts looking attractive — colder, smarter, more like a hard answer should look. So hunt for the receipt: which word proves restrained wrong? There isn't one. "Acknowledged without conceding" is restraint, described. All you have is a feeling about calculated's personality. No receipt, no switch.

The difference runs in seconds: the legal switch ends with a finger on a word; the illegal one ends with an adjective.

Running it under time pressure

  1. Mid-section: answer and move. Don't relitigate finished questions between items — that's a time-sink wearing a new costume.
  2. Flag genuine unknowns. If a question isn't clicking, flag it, pick something, and go — the 90-Second Flag. Flags are for unresolved questions, not answered ones you felt weird about.
  3. At the review screen, open flagged questions only.
  4. On each flag, ask once: can I point to the word that proves my current pick wrong? Yes — switch once and move on. No — leave it and get out.

Review-screen discipline: flagged only

Here's the part that feels wrong and isn't. If the grade-twice test showed your firsts winning, your unflagged answers are closed cases: you made them with the passage fresh in working memory and your reasoning intact. Reopen one at minute 30 and you bring none of that — just the choices and free-floating doubt. You aren't checking your work; you're giving the trap a rematch under worse conditions. When the flags are done, stop.

If your score is stuck, start here

Switching off right answers is a classic way real studying stops showing up in scores: points arrive from content work and leak back out at the review screen. It's one of three behavioral leaks — habits that cost points independent of knowledge — mapped in the behavioral leaks guide. If you've been flat for weeks despite honest work, run the sorting method in the plateau playbook: a plateau made of behavior doesn't respond to knowledge medicine.

Frequently asked questions

Should I change my answer on the SAT?

On average, yes — research finds answer changes go wrong-to-right about 2:1. But measure yourself first with the grade-twice self-test: a minority shows the reverse pattern, the research doesn't say who, and if that's you, the Receipt Rule is the fix.

When is it OK to change an answer on a test?

When you have a receipt: an exact word you misread, a named rule your first pick breaks, or a question-stem detail you skipped. If you can't point to one, the change is a vibe — and vibes are how right answers get switched off.

Is the first-instinct fallacy real?

Yes — it's the finding that "always trust your first instinct" is bad advice on average, because changes help about twice as often as they hurt. The nuance: that's an average, a minority shows the opposite pattern, and the studies don't identify who. Your own graded sections do.

Should I review every answer at the end of a module?

If first-instinct override is your leak, no — review flagged questions only. Your unflagged answers were made with the passage fresh in working memory; rereading them later with nothing but doubt gives the traps a second chance under worse conditions.

Does the Receipt Rule slow you down?

It saves time. "Do I have a receipt?" resolves in seconds and ends the two-minute stare-downs at the review screen. It also kills mid-section relitigating, which is where time-sinks like to hide.

SAT® is a trademark registered by the College Board, which is not affiliated with, and does not endorse, this site.

← All guides