21. Anchoring
Wireless Philosophy, Cognitive Biases, 4:54. Original title: CRITICAL THINKING - Cognitive Biases: Anchoring [HD]
Summary
Laurie Santos (Yale University) explains anchoring, continuing her series on cognitive biases with a quick mental-maths demonstration.
A quick maths problem
Santos asks us to multiply, quickly and in our heads: eight times seven times six times five times four times three times two times one. When the psychologists Daniel Kahneman and Amos Tversky tried this with real subjects, the average guess was about two thousand two hundred and fifty. Asked instead to multiply one times two times three times four times five times six times seven times eight, the same subjects guessed much lower on average: about five hundred and twelve. Both problems are the same calculation in a different order, and the real answer for both is forty thousand three hundred and twenty: people are off by an order of magnitude, and, more strikingly, they answer the identical problem differently depending only on which end they start from.
Why: anchor first, adjust too little
Given enough time, you can multiply the whole chain out exactly. Under time pressure, you only multiply the first couple of numbers, for example eight times seven is fifty-six, and then guess the final answer must be "pretty big" from there; starting with one times two, which is only two, makes the eventual guess "pretty small" instead. That starting figure is the anchor: the first number we land on when making an estimate. Once we have an anchor, we adjust from it, but our minds are biased not to adjust as much as needed; like a real anchor, it keeps us stuck close to where we started.
Arbitrary anchors still work
Kahneman and Tversky also showed that this happens even with anchors that have nothing to do with the question. They had people spin a wheel numbered one to a hundred, then estimate what percentage of countries in the United Nations are African. People who spun a ten guessed about twenty-five percent; people who spun a sixty-five guessed about forty-five percent, purely because of the number they had just seen. In a related experiment, Dan Ariely and colleagues had people write down the last two digits of their social security number and then say whether they would pay that amount for a nice bottle of wine; people in the highest quintile of social security numbers turned out to be willing to pay three to four times as much for the identical bottle, so that someone who would otherwise pay eight dollars might offer twenty-seven instead.
Key takeaway
Because anchors, even arbitrary ones, pull our estimates toward themselves, salespeople and advertisers use them on us constantly, whether we are buying a car, a sweater or renting a hotel room. The next time you are handed an anchor, Santos suggests, take a minute to consider a very different number before you answer.