Anchoring bias is the tendency to rely too heavily on the first piece of information you meet when you make an estimate or a decision. That first number becomes an anchor, and later judgments move too little away from that number. Tversky and Kahneman showed the effect in 1974 with a rigged wheel and a short multiplication task, and later studies found it in doctors, estate agents and judges.
What anchoring bias is
Anchoring bias belongs to the family of cognitive biases that shape judgments under uncertainty. For a wider map of that family, read our overview of cognitive biases. The core idea is simple: when you estimate an unknown quantity, you start from an initial value and then adjust. The adjustment is usually too small, so the final answer stays close to the anchor.
The anchor can be relevant or completely random. It can be a price tag, an opening offer, a first impression or a number from a wheel. It still pulls the estimate. The effect persists even when people know the anchor is arbitrary, which is why researchers treat it as a bias rather than a sensible use of a clue. A description of the phenomenon as reliance on the first information encountered appears in Simply Psychology's guide to the anchoring bias.
The two classic experiments
Tversky and Kahneman ran two short demonstrations that textbooks still use. Both show the same pattern: an irrelevant starting number moves the final estimate.
The wheel of fortune
Participants watched a rigged wheel stop on either 10 or 65. They first judged whether the share of African countries in the United Nations was higher or lower than that number. They then gave their own estimate. The group that saw 10 gave a median estimate of 25 percent. The group that saw 65 gave a median estimate of 45 percent. The wheel had no link to the question, yet it moved answers by twenty points. The numbers come from the original 1974 report as summarized in Simply Psychology's guide to the anchoring bias.
The short multiplication
Two groups of students had five seconds to estimate a product. One group saw 8 x 7 x 6 x 5 x 4 x 3 x 2 x 1. The other saw the same digits in reverse, starting with 1 x 2 x 3. The descending group gave a median estimate of 2,250. The ascending group gave a median estimate of 512. The correct answer is 40,320. Students did a few steps, reached a partial result and stopped too early. The figures are reported in the Anchoring effect entry on Wikipedia.
A later classroom version used the last digits of a social security number. People with higher digits bid 60 to 120 percent more for wine, chocolate and computer equipment than people with lower digits. The digits had no link to value. The result is summarized in the Anchoring effect entry on Wikipedia.
Why it happens
Researchers explain anchoring with two mechanisms. One fits anchors you make yourself. The other fits anchors someone gives you.
Anchor and adjust
The first account comes from Tversky and Kahneman. You start from an initial value and move away from it until the answer feels plausible. Then you stop. Because adjustment takes effort, it weakens under time pressure, distraction or memory load. Epley and Gilovich showed this pattern with self-made anchors in 2006. Their work is summarized in Simply Psychology's guide to the anchoring bias.
Selective accessibility
The second account comes from Strack and Mussweiler. When you test whether an anchor could be right, you pull up facts that fit it. Those facts stay active in memory and shape the final estimate. In their 1997 study, people judged whether the Brandenburg Gate was taller or shorter than 150 meters, then estimated its real height. Estimates moved toward the anchor. The pull was strongest when both questions used the same dimension. This account explains why even a random number from another person still works. For a related heuristic that also runs on what memory serves up, see how the availability heuristic works. The Brandenburg Gate result is described in Simply Psychology's guide to the anchoring bias.
Neither account covers every case alone. Current reviews treat them as complements: adjustment fits self-made anchors, accessibility fits given anchors.
Everyday examples
Anchoring shows up far beyond lab tasks. The same pull appears in shops, clinics, courts and pay talks.
- Shop prices. A struck-through list price makes the sale price feel like a gain, even when the first price says little about value. A limit sign such as 12 per person can raise how much people buy.
- Salary talks. An opening offer sets the range for the whole talk. A low first offer pulls the final pay down. A high but careful first offer pulls it up.
- Medical judgments. A doctor can stick with a first impression and weigh later signs too little. Researchers warned about this risk for missed diagnoses during the pandemic, when a suspected virus hid other causes.
- Court sentences. Judges who heard a high hypothetical demand gave longer sentences than judges who heard a low one, though all knew the demand was hypothetical. The mean sentences were about 28.7 months after the high anchor and about 18.8 months after the low anchor. The study is described in The Decision Lab guide to anchoring bias.
- House prices. Estate agents and students valued the same house differently when they saw different listing prices. Agents moved almost as much as students, yet most said price had played no part. The result is reported in the Anchoring effect entry on Wikipedia.
A broad review by Furnham and Boo in 2011 found the same pattern in legal judgments, purchases, salary talks and forecasts. Implausible anchors still moved answers. The review is cited in the Anchoring effect entry on Wikipedia.
Anchoring bias vs framing effect
Students mix these two biases. One question sorts most cases. Ask whether a first number pulled the estimate. If yes, suspect anchoring. Ask whether rewording would change the choice. If yes, suspect framing.
| Anchoring bias | Framing effect | |
|---|---|---|
| Trigger | An initial number or value | How options are worded |
| Question that reveals it | Was there a first number | Would new wording change the choice |
| Classic example | Wheel at 10 or 65 moves a later estimate | Meat at 80 percent lean feels better than at 20 percent fat |
| Fix that helps | Argue against the anchor | Restate the facts in neutral words |
The meat example comes from the FAQ in The Decision Lab guide to anchoring bias.
How to reduce the effect
You cannot remove anchoring by willpower alone. Warnings and rewards for accuracy shrink it only a little. One method has the strongest support: argue against the anchor on purpose.
- 1.Name the anchor in one sentence.
- 2.List two reasons the anchor could be too high.
- 3.List two reasons it could be too low.
- 4.Set the anchor aside and make a fresh estimate from other facts.
This consider-the-opposite routine forces your memory to serve up facts that do not fit the anchor. In studies, experts who listed counterarguments moved further from the anchor than experts who did not. The method is summarized in Simply Psychology's guide to the anchoring bias.
Treat single effect sizes with care. A well-powered 2025 replication found only about a 3.4 percent shift where an early report had found 31 percent. The direction held, but the size was far smaller. The replication is summarized in Simply Psychology's guide to the anchoring bias. Use study numbers as examples from named studies, not as fixed laws.
How to remember it for exams
Link the term to its two classic numbers: 10 versus 65 on the wheel, and 512 versus 2,250 in the multiplication task. Then link it to one expert case: agents or judges. If a question gives a first number and asks why the estimate stayed near it, answer anchoring bias and name insufficient adjustment or selective accessibility.
Save the term to a bookmark list such as Cognitive biases and drill it with the Question game, which needs at least four words in the list. Short repeated sessions beat one long reread.
Frequently asked questions
What is anchoring bias in simple terms
Anchoring bias means the first number you see pulls your later estimate toward it, and you adjust too little. A random wheel, a list price or an opening offer can all act as anchors.
What is a clear everyday example of anchoring bias
A sale tag works as an anchor: a television marked down from 1,000 to 700 feels cheap even when 700 is still high. A salary talk works the same way: the first offer sets the range for the rest of the talk.
How is anchoring bias different from the framing effect
Anchoring comes from a first number that pulls an estimate. Framing comes from wording that changes a choice, such as 80 percent lean versus 20 percent fat. Ask whether a number or a wording did the work.
How can you reduce anchoring bias
Name the anchor, list reasons it could be too high and too low, then estimate afresh from other facts. This consider-the-opposite routine has the strongest research support, though it reduces rather than removes the pull.
Sources
- 1.Simply Psychology (2026). Anchoring Bias and Adjustment Heuristic: Definition and Examples.
- 2.The Decision Lab. Why do we compare everything to the first piece of information we received.
- 3.Wikipedia. Anchoring effect.
- 4.EBSCO Research Starters (2020). Anchoring (cognitive bias).
- 5.Nelson University ThoughtHub (2016). The Effects of Anchoring Bias on Human Behavior.