--- title: "Anchoring Bias" description: "Anchoring bias is the tendency for an initial number to pull subsequent judgements toward it, even when that number is arbitrary or irrelevant. Estimates adjust away from the anchor but typically not far enough, so whoever states a figure first shapes the range everyone else reasons within." last_updated: "2026-08-02" --- # Anchoring Bias Anchoring bias is the tendency for an initial number to pull subsequent judgements toward it, even when that number is arbitrary or irrelevant. Estimates adjust away from the anchor but typically not far enough, so whoever states a figure first shapes the range everyone else reasons within. Relying too heavily on the first piece of information encountered (the 'anchor') when making subsequent estimates or decisions. *Example: A jacket marked down from $500 to $300 feels like a steal — only because $500 grabbed you first.* ## What it looks like **In a salary negotiation** > The role is budgeted at £45,000 — does that work for you? *Why it fails:* Every counter-offer now forms around £45,000. A candidate who would have asked £60,000 unprompted typically lands well below it, having adjusted from the stated figure rather than from their own valuation. **In a retail listing** > Was £299 — now £149. *Why it fails:* The crossed-out price sets the reference point for what the item is worth, so £149 reads as value gained rather than money spent. The anchor need never have been a real selling price. **In a project estimate** > Do you think this will take about two weeks? *Why it fails:* Asking with a number attached shifts the estimate toward it. The same question asked openly regularly produces materially different answers from the same engineer. ## Practice ### Question 1 In the classic demonstration, participants spun a wheel showing a random number, then estimated the percentage of UN member states that are African. Those who saw a high number gave higher estimates. What does this show? A. Participants believed the wheel had information about the UN B. A number known to be random still shifted numerical judgement, which is why anchoring is treated as a general feature of estimation rather than a reasoning mistake about relevance C. Participants were not paying attention to the task D. The estimates were correct and the wheel was irrelevant **Answer: B.** The force of Tversky and Kahneman's result is that the anchor was transparently random and still moved estimates — participants were not confused about the wheel's relevance. That said, this literature has been the subject of considerable replication scrutiny, and effect sizes vary by design. The core phenomenon of anchoring replicates robustly; the specific magnitude in any one study should be held loosely. ### Question 2 You are negotiating and want to avoid being anchored. What is most effective? A. Refuse to discuss numbers at all B. Determine your own valuation and walk-away point before the conversation, so you are adjusting from your figure rather than theirs C. Always let the other side name a number first D. Mentally halve whatever number they say **Answer: B.** Pre-committing to your own number gives you a competing reference point, which is the only reliable defence. Option D fails because arbitrary adjustment is still adjustment from their anchor. Option C hands them the anchor deliberately. Option A is impractical and forfeits the chance to set the anchor yourself. ### Question 3 Why does anchoring persist even among experts in their own domain? A. Experts do not experience anchoring B. Because adjustment from a starting point is how numerical estimation works, and domain knowledge does not remove the starting point's pull C. Because experts are overconfident D. Because experts have less time to deliberate **Answer: B.** Studies with estate agents estimating property values and judges considering sentence lengths have found anchoring effects among people with substantial domain expertise. Expertise narrows the plausible range but does not eliminate the tendency to adjust insufficiently from whatever figure is in play. ### Question 4 A restaurant adds a £95 dish to a menu where nothing else exceeds £30. Even if nobody orders it, what effect is intended? A. It signals culinary ambition to reviewers B. It anchors the price range so that a £30 dish reads as moderate rather than expensive C. It is required for licensing reasons D. It tests whether customers are paying attention **Answer: B.** This is anchoring used deliberately, sometimes called a decoy. The expensive item reshapes the perceived scale, so mid-priced options move from feeling like the top of the range to feeling like the sensible middle. Its value to the restaurant does not depend on anyone ordering it. ### Question 5 Which is the best description of why anchoring is hard to correct by simply being told about it? A. People forget the warning too quickly B. The anchor influences the range you consider plausible, so even deliberate correction tends to start from the anchored range C. Warnings make people anchor more strongly on purpose D. It only affects people who are unfamiliar with numbers **Answer: B.** Being warned does not give you an unanchored estimate to compare against — you still have to generate a number, and the anchor has already shaped what feels plausible. This is why the effective countermeasure is generating your own figure independently, before exposure, rather than trying to discount afterward. ## Frequently asked ### Does anchoring only affect numerical judgements? It is clearest and best documented with numbers, but related framing effects appear with non-numerical starting points too — an initial description or category can shape subsequent assessment in similar ways. The numerical case is where the effect is easiest to measure and most reliably demonstrated. ### How do I use anchoring ethically in a negotiation? Making the first offer is legitimate and often advantageous, provided the number is defensible and you can say why. The line falls at anchors constructed to mislead — fictitious original prices, or figures presented as market rates that are not. Anchoring with a real number is negotiation; anchoring with a fake one is deception. ### Is anchoring bias well replicated? The basic effect is one of the more robust findings in judgement research and has replicated widely across domains and populations. Specific effect sizes vary considerably by design, and some individual classic studies have attracted replication scrutiny, so treat the phenomenon as solid and any single reported magnitude as approximate. ## Related - [Framing Effect](https://www.criticalthinkingexercise.org/pages/glossary/framing-effect.md) - [Availability Heuristic](https://www.criticalthinkingexercise.org/pages/glossary/availability-heuristic.md) - [Base Rate](https://www.criticalthinkingexercise.org/pages/glossary/base-rate.md) - [Biases in Decision Making](https://www.criticalthinkingexercise.org/pages/exercises/biases-in-decisions.md) - [Common Cognitive Biases](https://www.criticalthinkingexercise.org/pages/exercises/common-cognitive-biases.md) - [Lesson: traps-and-pitfalls/cognitive-biases-decision-social](https://www.criticalthinkingexercise.org/pages/learn/traps-and-pitfalls/cognitive-biases-decision-social.md) ## Sources - [Judgment under Uncertainty: Heuristics and Biases (Tversky & Kahneman, 1974)](https://doi.org/10.1126/science.185.4157.1124) — Science ## Sitemap See the full [sitemap](https://www.criticalthinkingexercise.org/sitemap.md) for all pages.