Anchoring in Choice Design

The idea that experts are immune to initial figures is a myth because of the anchoring bias. This is our tendency to rely heavily on the first piece of information (the “anchor”) when making a decision, even if that information is irrelevant. For example, a real estate agent might first show you a wildly overpriced house; every subsequent, more reasonably priced house will feel like a bargain in comparison. The first price anchored your perception of value.

Ask this: What was the first number or piece of information I saw, and is it shaping my judgment of what's reasonable?

Anchoring works by setting a cognitive starting point. Our minds grab onto this initial value and make adjustments from there. The problem is that these adjustments are often insufficient, keeping our final judgment closer to the anchor than it should be.

This is a powerful tool in choice design and policy. When a government agency presents a budget, the initial top-line number anchors the entire debate. If the first draft is $50 billion, negotiations might end at $45 billion. If the anchor had been $35 billion, the final number could have been much lower.

We also see this in technology subscriptions. A company might first display its '$299/year' plan. When the '$29/month' option is shown next, the number '29' feels small compared to the anchor '299'. This makes the monthly plan seem more affordable, even though it totals $348 annually. The initial large number anchors our perception of magnitude, not necessarily total value.

In legal contexts, the damages requested by a plaintiff's lawyer can anchor a jury's award decision. Even if the number is aspirational, it establishes a high reference point that can influence the final payout, demonstrating the effect's power in high-stakes environments.

Anchoring is a bias where we rely on an an initial value and make insufficient adjustments, a process called 'anchoring and adjustment' [5].

• In a seminal 1974 study, Tversky and Kahneman found that a random number skewed unrelated estimates, showing an anchor need not be relevant [1]. (1974)

• The bias impacts experts. Englich et al. (2006) showed that experienced judges' sentencing was swayed by arbitrary numbers from dice [3].

• Its robustness is well-documented; a review by Furnham and Boo confirms its power to influence novices and experts with even arbitrary anchors [4]. (2011)

The anchor's influence is not absolute. Its effect is weaker for experts with a firm, pre-existing sense of value. The bias can be mitigated by 'considering the opposite'—generating reasons why the anchor is wrong [5]. It is less pronounced for simple choices with clear metrics than complex, ambiguous ones.