Anchoring and Adjustment
In the study discussed here, participants described starting from a value and adjusting more often for self-generated anchors than for anchors supplied by an experimenter. The distinction suggests that different anchoring tasks may involve different processes.
THE IDEA
As predicted, participants were far more likely to describe a process of anchoring and adjustment when the anchor values were self-generated than when they were provided by the experimenter. Of those who appeared to know both self-generated anchors (n = 34), 94% made reference to anchoring and adjustment in response to at least one self-generated item, and 65% did so in response to both. In contrast, only 22% of participants (n = 50) described anchoring and adjustment for at least one experimenter-provided anchor, and merely 4% did so for both (see Table 7.1).
To assess the statistical significance of this pattern, we calculated the percentage of items for which each participant reported a process of anchoring and adjustment under the two conditions. Four participants were excluded because they knew neither self-generated anchor. The difference was substantial: participants reported anchoring and adjustment far more often when considering self-generated anchors (M = 73.9%) than when considering experimenter-provided anchors (M = 13.0%), paired t(45) = 8.56, p < .0001.
These self-reports suggest that self-generated and experimenter-provided anchors may involve different processes. The anchoring-and-adjustment heuristic, originally described by Tversky and Kahneman (1974), involves starting from an initial value and then adjusting to reach a final estimate. Critically, adjustments are usually insufficient, biasing the final answer towards the anchor. This heuristic has been demonstrated across a wide range of judgment tasks: people anchor on the probability of one component when judging the likelihood of disjunctive events (Rottenstreich & Tversky, 1997), on their own ability when making comparative judgments of peers (Kruger, 1999), and on the extremeness of evidence when forming confidence judgments (Griffin & Tversky, 1992). Classic experiments illustrate the power of anchoring even when the anchor is plainly irrelevant. In one study, participants first observed a random number generated by a spinning wheel, then estimated the percentage of African countries in the United Nations; those who saw a higher anchor gave higher estimates (Tversky & Kahneman, 1974). In a completely natural setting, a simple sign reading “Limit 12 per customer” led shoppers to purchase significantly more soup than a sign saying “No limit,” showing that an arbitrary purchase limit served as an anchor that guided behavior (Wansink, Kent, & Hoch, 1998).
Other proposed explanations of anchoring include selective accessibility, in which anchor-consistent information becomes easier to retrieve, and numeric priming, in which exposure to a value influences a later estimate. The present study suggests that these mechanisms may be engaged differently depending on how the anchor originates. Self-generated anchors appear to evoke the deliberate process of adjustment that participants can readily introspect about, whereas experimenter-provided anchors may operate through more automatic, less reportable pathways.
One might be concerned, however, about relying on participants’ self-reports given the widespread doubts about the accuracy of such reports (Nisbett & Wilson, 1977). An alternative Gricean interpretation notes that participants may simply fail to mention the experimenter-provided anchor because it was already stated in the comparative question, reducing the perceived need to repeat it. This interpretation is weakened by the fact that the same result pattern holds when scoring responses only for statements of adjustment, not the anchor itself. Distinguishing these explanations requires evidence beyond what participants choose to mention in a retrospective account.
When to use it
Negotiations: The first offer often serves as an anchor, influencing the final settlement.
Salary negotiations: An initial salary offer can anchor subsequent discussions.
Forecasting: Initial estimates in financial or sales forecasts can anchor subsequent revisions.
Everyday decisions: When estimating driving time, anchoring on a typical commute time and adjusting for traffic conditions.
Preparing an independent salary range before hearing an initial offer.
Revisiting a project estimate when its original assumptions have changed.
Determining the value of a product in a market with fluctuating prices
Checking whether revisions to a business forecast respond sufficiently to new evidence.
FROM THEORY TO EVERYDAY LIFE
In practice
Supermarket Purchase Limit
A study found that when a sign said 'Limit 12 per customer,' shoppers purchased more cans of soup compared to when the sign said 'No limit per customer.' The limit served as an anchor, influencing purchase quantity.
Wheel of Fortune
In Tversky and Kahneman's classic experiment, participants spun a wheel of fortune that landed on a number (e.g., 65). They were then asked whether the percentage of African countries in the UN is above or below that number, and to give an estimate. Those who received a higher anchor gave higher estimates, showing the influence of irrelevant anchors.
Antique Chair Valuation
When valuing an inherited antique chair, a person might anchor on the price of a similar chair seen elsewhere and adjust for differences. However, if they anchor on a very different chair (e.g., a signed designer chair) and fail to adjust properly, the estimate can be biased.
Used Car Negotiation
A used-car seller’s asking price may influence a buyer’s estimate. The buyer can also begin with an independently researched value and adjust for condition and mileage. Having an independent starting point does not guarantee freedom from the seller’s anchor.
Salary Negotiation
An initial salary offer can shape later discussion. Preparing a range based on the role and relevant comparisons gives a negotiator another reference point, though the offer may still influence the counteroffer.
A MOMENT TO REFLECT
Put it to work.
Think of a recent purchase you made. To what extent did an initial price anchor influence your willingness to pay?
Consider a time when you estimated a number (e.g., the cost of a project). Did you start from an initial figure and adjust? Did you adjust enough?
How might advertisers or salespeople use anchoring to influence consumer behavior?
How does the self-generation of an anchor affect the adjustment process compared to an externally provided anchor?
In what real-life situations might you be unaware of the influence of anchors on your judgments?
When someone says they did not use an anchor, what evidence would help distinguish an absence of influence from an influence they cannot report?
References
- Judgment under uncertainty: Heuristics and biases — Tversky, A. & Kahneman, D. (1974)
- An anchoring and adjustment model of purchase quantity decisions — Wansink, B., Kent, R. J., & Hoch, S. J. (1998)
- Lake Wobegon be gone! The 'below-average effect' and the egocentric nature of comparative ability judgments — Kruger, J. (1999)
- The weighting of evidence and the determinants of confidence — Griffin, D. & Tversky, A. (1992)
- Unpacking, repacking, and anchoring: Advances in support theory — Rottenstreich, Y. & Tversky, A. (1997)
- Telling more than we can know: Verbal reports on mental processes — Nisbett, R. E., & Wilson, T. D. (1977)
CONTINUE THE CONVERSATION
Source texts
- Heuristics and Biases The Psychology of Intuitive Judgment (Thomas Gilovich, Dale Griffin, Daniel Kahneman)
- Heuristics and Biases The Psychology of Intuitive Judgment (Thomas Gilovich, Dale Griffin, Daniel Kahneman)