Cognitive biases
The anchoring effect: why the first number shapes your decisions so much
You are looking at an apartment. The listing shows “€280,000” crossed out, with a new price next to it: €245,000. The thought arrives before you have calculated anything: “Not bad, they knocked off thirty-five thousand.” Meanwhile, the main question — what the apartment is actually worth — has not even been asked.
This is the anchoring effect: the first number we see or hear becomes a starting point. From then on we estimate from that number, not from zero, and usually we do not move far enough away from it. Even when the number is obviously random.
Below, we look at what research has shown, where anchors show up in everyday life, and what helps you depend on them less. In the middle of the article you can try the classic experiment on yourself.
Contents
In short: how the anchoring effect works
What the anchoring effect is
The best-known experiment was described by Amos Tversky and Daniel Kahneman in a 1974 paper. Participants watched a wheel with numbers from 0 to 100 spin. The wheel was rigged: for some people it stopped on 10, for others on 65. Then people were asked what percentage of United Nations member countries are in Africa.
The number on the wheel had nothing to do with the question, and participants could see that. Even so, the typical answer (the median) in the group that got 10 was 25%, and in the group that got 65 it was 45%. The gap was almost twofold, and a random number created it.
Since then the effect has been tested hundreds of times, with prices, deadlines, distances, and fines. In a large replication project, where dozens of labs repeated the same experiments, anchoring turned out to be one of the most reliable results. That is rare in psychology: a classic effect that survived strict rechecking.
Try it yourself
The experiment takes about a minute. Here is what will happen.
Step 1. The wheel stops on 10 or on 65: some players get one number, others the other. The number has no connection to the question, and you know that.
Step 2. We ask whether the share of African countries in the UN is higher or lower than that number. This question makes you take the random number seriously for a second and hold it up against your answer. The original study was set up the same way.
Step 3. You give your own estimate.
Step 4. We compare your answer with the correct one and with the answers of the 1974 participants.
Why a question about Africa? Almost nobody knows the exact answer, so you have to make a rough guess. And when we are unsure, an anchor can influence the estimate more easily: in experiments, people who knew the topic better were less affected. If we asked how many days are in a week, no wheel could throw you off.
If your answer did not move toward the number on the wheel, that does not mean the effect is not real. It shows up on average across many people, not necessarily in each one. A single answer cannot show it at all: you do not know what you would have said with a different number. That is why the demo ends by showing how the answers of the two groups in the original experiment diverged.
Why the first number sticks
Picture a boat at anchor. The chain does not hold it in one place; the boat can drift. But only as far as the chain is long. This is an image, not a mechanism, yet it captures what the experiments show: the estimate moves, but stays tied to the first number.
Researchers offer two explanations, and according to a review of anchoring research, both find support.
Insufficient adjustment. A seller says a used bicycle costs €500. You think: “No, that’s a lot. 450? Still a lot. 350 — maybe that’s fine.” You started at 500 and moved down until the price stopped looking clearly too high. And you stopped at the edge of “normal,” not in the middle of it. If the seller had said 150, you would have moved up from 150 and stopped as soon as the price no longer looked suspiciously low, say around 200.
Selective accessibility. When you ask yourself, “Maybe it really is 65%?”, memory obligingly supplies arguments for a large number: “Africa does have a lot of countries.” When the anchor is 10, other thoughts come up: “Well, it’s only one continent.” These arguments do not go away, and they color the final estimate. This is why the trick of “recalling arguments against the anchor” works: it pulls the other side out of memory. In an experiment by Thomas Mussweiler and colleagues, such a simple procedure noticeably weakened the influence of the anchor.
Where anchors show up in life
Shopping. The “old price,” the manufacturer’s suggested price, the first figure a seller names at a market. Even if you never planned to pay that much, the starting number sets what counts as expensive and what counts as a bargain. That is why a discount from an inflated price feels like a gift.
Salary and budgets. In negotiations, the first number named becomes the frame of the conversation: after that, both sides discuss how far to move from it. Research on negotiations shows that the first offer is linked to the final agreement. The same happens with last year’s project budget or the first time estimate named at a meeting: they are rarely recalculated from scratch, and more often adjusted slightly.
Judging people and work. The first assessment you hear about a person or a project can act as an anchor for the ones that follow. “A colleague said it’s just average” — and you are already reading the text looking for confirmation. Here the anchoring effect overlaps with confirmation bias.
Numbers in the news. A striking number in a headline sets the scale against which we then compare everything else. How strongly this affects your opinion of a particular story depends on the topic and on what you already know. But it is always worth asking yourself, “What should I compare this figure with?”
Does experience protect you?
It is tempting to think that an anchor only works on beginners. The data do not agree.
In the study by Northcraft and Neale, real estate agents toured the same house and received identical materials. Only the listing price differed. The professionals’ estimates shifted along with that price, and most of the agents said they had not relied on it.
In an experiment with legal professionals, experienced judges and prosecutors proposed a sentence based on the description of a fictional case. Before that, they rolled rigged dice. A higher number on the dice went together with a harsher sentence. This is a lab situation, not real sentencing, but it shows that experience does not switch the effect off.
The conclusion is not “nobody can be trusted.” It is more modest. Knowledge helps, but it does not free you from the anchor. What helps is a procedure, not confidence in your own objectivity.
How to depend less on anchors
The anchoring effect is one of the biases on the cognitive bias map. There you can also see which other biases affect money and decisions. If you are curious how we judge our own knowledge, read about the Dunning–Kruger effect.
Frequently asked questions
Are “anchoring” and “the anchoring effect” the same thing?
Yes. In English you will see “anchoring,” “the anchoring effect,” and “anchoring bias.” They all describe one phenomenon: the first number becomes the starting point for an estimate.
If I know about the anchoring effect, does it stop working?
Mostly, no. In experiments, a simple warning about anchors changed the result very little. Concrete actions work better: your own number in advance, independent reference points, and arguments against the anchor.
Can I use an anchor in my favor?
Yes, and salespeople and experienced negotiators do it all the time. The honest way is to name first a number you can justify. An inflated anchor may work, but it destroys trust once it is exposed.
Can an anchor be something other than a number?
Most research uses numbers: prices, percentages, deadlines. But a first impression, a first assessment, or the first solution you think of can work in a similar way, as the thing everything else is then compared with.
