Anchoring and Framing: Why the Same Choice Looks Different Depending on How It's Presented

In the early 1970s, two psychologists sat subjects in front of a wheel of fortune marked from 0 to 100. The wheel was rigged to stop on only one of two numbers — 10 or 65 — and the subject watched it land. Then came the actual question, which had nothing to do with the wheel: What percentage of African countries are members of the United Nations? People who had just watched the wheel land on 10 guessed, on median, around 25 percent. People who had watched it land on 65 guessed around 45 percent. A number everyone knew was random — produced by a spinning wheel in front of them — had nearly doubled the answer to an unrelated question of geography.

That experiment, reported by Amos Tversky and Daniel Kahneman in their landmark 1974 paper in Science, is the cleanest demonstration ever run of anchoring: the mind grabs whatever number is in front of it and adjusts away from that starting point — usually not far enough.

The mental model: a starting point you can't shake

Anchoring is the tendency to rely too heavily on the first piece of numerical information you encounter when forming an estimate. The anchor sets a mental starting line; you adjust from it toward your real answer, but the adjustment is systematically too small, so the final answer is pulled toward the anchor. The effect is so robust that it survives even when, as in the wheel-of-fortune setup, the anchor is transparently arbitrary and irrelevant.

Framing is the close cousin. Where anchoring is about a number that precedes your judgment, framing is about the words used to describe the options themselves. Tversky and Kahneman's later 1981 paper, "The Framing of Decisions and the Psychology of Choice," showed that logically identical choices flip depending on whether they are described in terms of gains or losses. The Library of Economics and Liberty's behavioral economics entry treats both as core evidence that human preferences are constructed in the moment, not simply read off a fixed internal ranking — which is exactly what the classical "rational agent" model denies.

The numbers behind the wheel

It is worth pausing on why the wheel result is so damaging to the rational model. The classical view holds that your estimate of UN membership reflects your knowledge, full stop. An irrelevant random number should move it by zero. Instead it moved the median answer from roughly 25 percent to roughly 45 percent — a 20-percentage-point swing driven by nothing but which number a wheel happened to display seconds earlier. The information content of that anchor was literally zero, and it still dominated.

The same machinery operates with real stakes. In follow-up studies, experts asked to estimate values — real-estate agents pricing a house, judges considering a sentence — were measurably pulled by anchors that should have been irrelevant to their professional judgment. Anchoring is not a novice's error that expertise cures. It is closer to a feature of how human estimation works.

Myth: "I'd never be fooled by the wording"

Most people believe they respond to substance, not packaging. The framing experiments say otherwise, and the demonstrations are uncomfortably simple.

Describe a ground beef product as "80% lean" and shoppers rate it as better-tasting and higher-quality than the identical product described as "20% fat." Describe a medical procedure as having a "90% survival rate" and patients choose it more often than the same procedure described as having a "10% mortality rate." In finance, a fund advertised with a "90% success rate" attracts more money than the same fund advertised with a "10% failure rate." And a subscription priced at "$30 a month" feels smaller than the same subscription priced at "$360 a year," even though they bill you identically.

The facts do not move across these pairs. Only the frame moves — and the decision moves with it. This is why the belief that you are immune is itself part of the vulnerability: the effect operates below the level where you feel yourself being persuaded.

Where this drains your money

Anchoring and framing are not laboratory curiosities. They are load-bearing tools in how products are priced and sold.

Consider the "original price" on a discount tag. A jacket marked "$200, now $120" feels like a $80 win, and the $200 is the anchor doing the work — whether or not the jacket ever genuinely sold for $200. Drop the anchor and the same $120 jacket is just a $120 jacket, evaluated on its merits. Consumer-protection regulators take this seriously enough to police it: deceptive reference pricing — inflating a fictitious "original" price to make a discount look bigger — is a recognized unfair practice, and the kind of disclosure manipulation the Consumer Financial Protection Bureau examines when it studies how the presentation of costs steers borrower decisions.

The pattern repeats across finance. A credit card's high "standard" interest rate makes a balance-transfer teaser feel generous by contrast. A car salesperson opens negotiations with a sticker price far above the expected sale price precisely to anchor the whole negotiation high; every dollar "off" that anchor feels like a concession won, even if the final price is ordinary. A 401(k) menu that lists a "suggested" contribution of 6 percent quietly anchors participants near 6 percent. And the original purchase price of a stock — discussed at length in studies of loss aversion — is an anchor that makes investors refuse to sell at $60 a share they bought at $80, treating a number the market has long since forgotten as the "real" value. The U.S. Securities and Exchange Commission's Investor.gov education materials repeatedly steer investors back to a holding's current prospects for exactly this reason.

How to drop the anchor and reframe the frame

You cannot turn anchoring off, but you can refuse to accept other people's anchors. The single most effective habit is to generate your own number before you are shown theirs. Decide what a car is worth to you, or what you will pay for a house, before you see the asking price; that self-generated anchor blunts the pull of the seller's. When a discount is dangled, ask what the item is worth to you at the new price, ignoring the crossed-out one entirely — because the only price you actually pay is the new one.

Against framing, the trick is to deliberately restate the choice in the opposite frame and see if your preference holds. Translate the "$30 a month" plan into "$360 a year." Convert the "90% success rate" into a "10% failure rate" and notice whether you still feel the same. Turn a "gain" frame into a "loss" frame. A preference that survives the flip is probably real; a preference that evaporates was being driven by the packaging.

The unifying lesson is humbling and useful at once. A wheel of fortune moved people's beliefs about the United Nations, and a single adjective moved how good a hamburger tasted. If a random number and a word can do that, the prices and pitches engineered specifically to move you deserve real suspicion. The defense is not to be smarter than the bias — it is to insist on your own reference point before anyone hands you theirs.