A wheel of fortune stops on 65. Someone asks you: what percentage of African countries are in the United Nations? Your answer will be higher than it would have been if the wheel had stopped on 10.
You know the wheel result is random. You know it has nothing to do with African countries. And yet the number sticks — warping your estimate like a magnet distorting a compass needle. This is anchoring bias.
What anchoring is
Anchoring bias is the cognitive tendency to rely too heavily on the first piece of numerical information you encounter when making subsequent estimates or decisions. That first number becomes an anchor — a reference point your brain adjusts away from, but never far enough.
The key insight: you adjust from the anchor, you don't reset from zero. If the anchor is high, your adjustment ends up high. If the anchor is low, your final estimate ends up low. The adjustment is almost always insufficient, which means the anchor exerts a persistent pull on the answer even when you consciously try to counteract it.
The original study
Tversky, A. & Kahneman, D. (1974) — "Judgment under Uncertainty: Heuristics and Biases." Science, 185(4157), 1124–1131. The spinning wheel experiment described above is from this paper. Participants who saw the wheel land on 65 guessed that 45% of African countries were in the UN. Those who saw 10 guessed 25%. The random anchor shifted estimates by 20 percentage points. This paper, one of the most cited in social science, introduced the anchoring heuristic alongside representativeness and availability.
Three real-world examples
Salary negotiation. Whoever names a number first sets the anchor. If a candidate says $95,000, the entire negotiation gravitates around that figure. A hiring manager who intended to offer $80,000 is now psychologically pulled upward; a candidate who would have accepted $75,000 is pulled to defend the higher number. Studies consistently show that first offers in negotiation heavily predict final outcomes.
Retail pricing. "Was $199, now $89" — the $199 is the anchor. Your brain evaluates $89 relative to that reference, not relative to what the item is actually worth. Retailers design this deliberately. The anchor doesn't need to be a real price. It just needs to be a number that appeared first.
Medical diagnosis. A doctor who forms an early hypothesis tends to interpret subsequent information as confirming it — but anchoring also operates on the numerical side. An early lab result that seems high makes the doctor weigh all subsequent data relative to that baseline. In complex diagnoses, first impressions leave a trace that's hard to erase even with contradictory evidence.
Why you can't easily escape it
Being aware of anchoring bias doesn't make you immune. Research shows that even people explicitly told about the bias — even statisticians and finance professionals — still exhibit it. The bias operates at a level below conscious correction.
The mechanism is partly selective accessibility: once an anchor is presented, it activates related knowledge consistent with that anchor. Your brain retrieves information that makes the anchor seem plausible, which biases your final judgment before you've consciously reasoned through anything.
Partial defenses exist — consider opposite scenarios, consider multiple reference points, take time before committing to an estimate — but none eliminate the effect entirely. Anchoring is one of the most robust cognitive biases ever documented. Over 40 years of research has failed to find a reliable way to turn it off.