Perceptions of Randomness

One area of interest of mine is how people understand random events, and use the concept of randomness in their daily lives. To this end, there are several programs of study I have been working on.

Retrospective Gambler's Fallacy

Psychologists have known for a long time that people seem to distrust streaks in random series. In fact, people often behave as though streaks are more likely to end than chance would dictate. This is called the gambler's fallacy. A simple illustration of the gambler's fallacy is the fact that after a coin has landed on heads five consecutive times, people are more likely than chance to predict that the next flip will land on tails.

It turns out that people demonstrate the gambler's fallacy for events that have already occurred in addition to making predictions about the future (imagine I flipped a coin 6 times a half hour ago, the first 5 landed on heads, what do you think the 6th landed on). Since people's memories are reconstructive, it seems plausible to imagine that our memories of past events that we believed to be random will be biased by our perception of what randomness should look like. Therefore, if past random events exhibit streaks (as they inevitably will given enough time), we should have systematic memory errors. The extent of these biases are one topic of investigation.

Probability Matching

Imagine an unbalanced coin in which heads came up 60% of the time. If you were to try and predict the outcome of the coin, what percent of the time should you select heads to lead to optimal performance? The correct answer is 100%. However, most people probability match: they choose heads 60% of the time. Why is it that people use this non-optimal strategy?

One possibility is that people are looking for patterns in the data. That is, people may believe that randomness has a certain "look" that they can predict. One area of interest is investigating whether people are 1) looking for a particular "prototypical" set of outcomes and 2) Whether people are sensitive to deviations from true randomness, so as to be able to effectively pick up patterns.

Gambler's Fallacy and Hot Hand Bias

As described previously, the gambler's fallacy is a predisposition to believe that streaks are about to end. However another bias that has been identified in the literature is the hot hand bias: the belief that a person is "hot" or "on a roll" and that the streak is likely to continue. Although both of these biases have been demonstrated -- both in the laboratory and the real world -- they make diametrically opposed predictions.

It is still unknown exactly when and why people will exhibit each of these biases. Possible mechanisms include causal theories, motivational biases, sequential vs. simultaneous display, quality of "win" and many others. One area of interest is to try to tease apart when people exhibit the gambler's fallacy, and when they exhibit the hot hand fallacy.

What is randomness

What exactly is random? Most people agree that a coin flip is more random than whether Michael Jordan will make his next three pointer, but why is this so? A coin's outcome is predetermined -- if one knew all the forces that were acting on the coin (e.g. gravity, wind resistance, the flipper's finger) one could predict the outcome with certainty. What about this leads one to think it is random?

Although a formal definition of the abstract concept of randomness is best left to philosophers, psychology has done little to explore people's beliefs of what is random, and what elements predict those beliefs. To this end, one area of inquiry is to establish what factors are involved in people's a priori expectations of a random process.

I have many more research interests as well. go back to the homepage to look at them further.