Anchoring and Adjustment

Understanding Anchoring and Adjustment Heuristic

In estimating the frequency or likelihood of an event, people use a starting point (called anchor) and then make adjustment up and down from this starting point. This mental shortcut or heuristic is called anchoring and adjustment.

Anchoring and adjustment heuristic is common to information-processing and decision-making processes.

It is a form of heuristic Opens in new window that the decision maker uses to “evaluate a sequence of information by choosing an initial estimate or anchor against which future adjustments are made as additional information is received.

A manager often makes a judgment by starting from some initial point and then adjusting to yield a final decision.

The initial point, known as the anchor, can come from the way a problem is framed, from historical factors, or from random information. Even when an anchor is absurd and people recognize it as such, their subsequent judgments are often very close to that starting point. That is, regardless of the initial anchor point, subsequent adjustments tend to be insufficient (Tversky & Kahneman, 1974).

Classic Studies of Anchoring & Adjustment

The accessibility of information can also lead to reliance on the anchoring and adjustment heuristic, in which people rely on an initial starting point making an estimate and then fail to adequately adjust their original decision (Mussweiler & Strack, 2000). In one study, researchers asked participants to estimate the number of states in the United States in 1840 (Epley & Gilovich, 2004).

Some participants were asked to make their estimation after reading that “the United States declared its independence on July 4, 1776.” This statement should remind participants that at this time the United States consisted of only 13 states (a low anchor). Other participants were asked to make their estimation after reading that “the United States will celebrate its 225th anniversary on July 4, 2001.” This statement should remind participants that currently the United States includes 50 states (a high anchor).

As predicted, participants who read the first statement gave a much lower estimate of the number of states than those who read the second statement: 21.3 versus 30.9. Thus, in both cases participants adjusted their response, but they did so insufficiently because the actual number of states in 1840 was 26.

Survey results can be biased due to the response options given: if you ask people how much TV they watch, providing a low anchor (e.g., “do you watch more or less than 5 hours?”) leads to lower reports than providing a high anchor (e.g., “do you watch more or less than 15 hours?).

People even fail to sufficiently adjust when the initial anchor is obviously wrong. In some study, some students were asked whether Mahatma Ghandhi died before or after age 140, and other students were asked if he died before or after age 9 (Strack & Mussweiller, 1997).

All students were then asked how old Ghandhi was when he died. Those who had been asked the first questions—with the anchor of 140—guessed on average that he was 67 years old when he died. Those who had been asked the second question—with the anchor of 9—guessed on average that he was 50 when he died. (For the record, Ghandhi was 78 years at the time of his death.)

There are numerous examples of bias resulting from anchoring and adjustment. For example, some school systems categorize children into certain performance categories at an early age. Whereas a child anchored in a low-performance group might meet expectations, another child of similar ability but anchored in a higher-performance category could be perceived as being a better performer simply because s/he was categorized as being a high performer.

Similarly, a low starting salary could be an anchor that a high-performing employee has difficulty overcoming even with substantial annual increases in terms of a percentage of base salary.

Both the source of an anchor and norms for adjustment might vary with cultural experience. For example, the willingness of new migrants from Hong Kong to Vancouver in the 1990s to pay far above market prices for residential property might be explained by this heuristic Opens in new window.

The Hong Kong Chinese might have anchored their initial estimate of the cost of housing in Vancouver in their previous experience. Subsequent estimate might still have been higher than reality because of the general tendency to make an insufficient adjustment mentioned earlier and a collectivist norm for avoiding extremes in evaluations.

In some cases, it makes sense to rely on the initial anchor. For example, when buying a house, the asking price (the initial anchor) is probably very relevant because it is based on a realistic appraisal of the selling prices of similar homes. However, people rely on anchors to make their judgments even when the anchor should clearly have no impact on their decision.

In a classic study, researchers spun a large wheel of fortune and asked people to evaluate whether the number on which the wheel stopped was higher than the percentage of African countries that belonged to the United Nations (Tversky & Kahneman, 1974).

In spite of the obvious irrelevance of the anchor, people gave a higher estimate when the wheel stopped on a high number than when it stopped on a low number.

Interestingly, in some cases, we can use people’s tendency to use anchoring in ways that are beneficial. In a recent study, Janiszewski & Uy (2008) demonstrated that home sellers get higher prices when they provide a precise number (such as “252,500”) than a rounded number (such as “250,000”). Why could the nature of the anchor in this case influence the final price of such an important purchase?

The authors propose that when people are bidding on something that costs a round number (such as $20.00), they think in terms of dollars (and then whether this object is actually worth $19 or $18 or $21). But a more precise number leads people to think in smaller denominations—in turn, if something is priced at $19.85, we think to be closer to the initial price when a precise anchor has been given as opposed to a more rounded anchor.