Anchoring Detection and Adjustment¶

Why It Matters¶

When making numerical judgments (valuations, predictions, estimates), people anchor on available numbers and adjust insufficiently from them. The anchor can be completely random (last digits of your phone number), strategically planted (first offer in negotiation), or contextually available (prior year's budget). Once anchored, adjustments are typically inadequate, leaving final estimates biased toward the anchor. This makes perception systematically manipulable and causes errors in judgment that persist even when the anchor's irrelevance is obvious.

The Key Move¶

When making numerical estimates, first notice if any number has been mentioned or is contextually salient. That's a potential anchor. Before adjusting from it, deliberately generate an independent estimate from first principles or a different starting point. If negotiating, notice who sets the first number and recognize its anchoring effect. If evaluating, check whether your "independent" judgment is suspiciously near a salient number. Make anchor-independence explicit: "Ignoring the $X figure mentioned, what would I estimate based on fundamentals?"

Classic Application¶

Tversky and Kahneman's wheel-of-fortune experiment. Participants watched a rigged wheel land on 10 or 65, then estimated the percentage of African nations in the UN. Those who saw "10" averaged estimates around 25%; those who saw "65" averaged around 45%. The wheel was obviously irrelevant, participants knew it was random, yet it anchored estimates. The effect persists even with expertise and monetary incentives—it's not about conscious reasoning but about automatic adjustment processes.

Surprising Application¶

Salary negotiations and career decisions. When asked "What are your salary expectations?", the initial number spoken becomes a powerful anchor—whether it comes from you or the employer. If you state first and too low, you anchor the negotiation below your market value. If the employer states first and you "adjust up," your adjustment is typically insufficient. The move: before any numbers are discussed, independently research market rates from multiple sources, establish your reservation price based on alternatives, and anchor yourself to that internal standard rather than whatever first number enters the conversation.

Failure Modes¶

Over-adjustment leading to contrarianism—deliberately avoiding reasonable anchors can lead to worse estimates than moderate anchoring. Ignoring that some anchors are informative—prior year's sales are actually relevant to this year's projection, unlike random numbers. Paralysis from realizing all starting points are somewhat arbitrary—you need some anchor to begin reasoning, the key is testing multiple ones. Believing awareness eliminates anchoring—knowing about anchoring reduces but doesn't eliminate its effects.

Go Deeper¶

Tversky & Kahneman, "Judgment under Uncertainty: Heuristics and Biases," Science 185(4157), 1974; Epley & Gilovich, "The Anchoring-and-Adjustment Heuristic," Psychological Science 17(4), 2006.

Anchoring - Behavioral Economics: Reasoning About Systematic Deviations