Note: I wrote this as a guest post for another blog, but the other blog didn't end up publishing it, so I'm putting it here.
Are your senator’s hands hot?
Have you ever seen a basketball player who, for just a few minutes, seems to be able to make every shot he attempts? Sometimes we say that this player is “on fire” or has a “hot hand.” Professional commentators and fans alike are quick to notice when a player seems to be “hot,” sometimes even to the point that they believe that when a player makes a single shot, he is automatically more likely to make his next shot.
Psychologists and others have studied people’s beliefs about hot hands for several decades. Much of the research has found that the hot hand is no more than a figment of the imagination. Thomas Gilovich of Cornell (with Robert Vallone and Amos Tversky), for example, interviewed basketball fans, and asked them to think of players who seemed to be the most likely to have hot hands. Most people were confident that basketball players really had hot hands after scoring, and could list several players that they thought were especially likely to get hot hands. Data on actual performance in basketball games, however, showed absolutely no evidence that any of these players ever had hot hands! Each shot seemed to be totally independent of every other shot, and success once didn’t seem to ensure success twice. People were perceiving something that simply wasn’t there.
The hot hand belief has many applications in the world and in public policy. For example, when people are choosing someone to manage their investments, they often choose someone who has done very well in the very recent past, even if other evidence indicates that their recent success was really just luck. They seem to implicitly believe that investment managers can have hot hands, and that their success yesterday automatically ensures their success tomorrow. In the world of policy, it’s easy to imagine many examples of hot hand biases. Imagine voters who re-elect someone because of a recent spate of good outcomes, thinking that these outcomes must continue because their leader is “on a roll.” Or, think of a leader who orders a general to continue a string of dubious military campaigns that have been successful in the recent past, thinking that the general will continue to have a hot hand in his future campaigns. Imagine the damage that could be done by the general himself, who might become cocky about his recent success and his own seemingly hot hand.
In graduate school I did some research about de-biasing the hot hand. My hypothesis was that a simple reminder of a track record can do the trick. For example, when people are predicting a coin toss and they’ve just seem four tails in a row, they might fall prey to the hot hand bias and predict that tails will definitely come next. If we simply remind them that the coin is fair (it has come up tails exactly 50% of the time in the past), they may realize their mistake and revise their prediction. In the domain of basketball, every time a player seems to be on a streak, we could remind commentators and fans about the player’s underlying skill level. If people realize that a player has made 73% of his thousands of shots for many years, they should understand that a streak of four shots in a row doesn’t mean much: he still has an exactly 73% chance of making his next shot.
How does this apply to policy? One possible implication is that voters and decision makers should be reminded about people’s track records before they vote or make decisions. If the ballot included information about a candidate’s record of success (for example in bringing about economic growth), voters would be less likely to focus on the fact that the candidate seems to be “on a roll,” and more likely to understand that the candidate’s future success will probably be close to his existing record. De-biasing voters could lead us to collectively choose better leaders.
Hot hand beliefs are strong and persistent, and like every psychological phenomenon, they can affect policy. More research is needed so that we can understand the hot hand bias and (ideally) eliminate it from our thought processes. I suggest that voters look closely at track record information before making decisions. If this works, it could be a simple and unobtrusive way to make the world a better place.
Post by Bradford Tuckfield
I offer a full range of data analysis services. To set up a free consultation, email me at firstname.lastname@example.org.