Choosing March Madness underdogs: Study finds people are riskier in gambles, forecasts made later in a sequence

Newswise — New research suggests that those who get hit with the annual case of March Madness are more likely to take a chance on higher seeds for matchups they happen to fill out within their brackets later on, rather than first. 

Jackie Silverman, assistant professor of marketing, co-authored a study in Management Science that found people are more risk seeking (more likely to predict a relatively improbable outcome) in their later forecasts and are thus more likely to make an incorrect prediction for them. For instance, individuals were more likely to bet that an “underdog” team will win in a sports tournament for predictions they made third, versus first, in a series of forecasts.

There can be negative consequences for this "prediction order effect"; by choosing the less likely option in these later predictions, people face greater chances of being wrong, which can reduce earnings and happiness.

Participants in the study made three consecutive predictions about which sports team would win in an upcoming game, with the order in which these predictions were made counterbalanced across participants. Although the outcomes of these games were independent and their order was random, participants were more likely to forecast that the underdog team would win for predictions they happened to make later in the sequence.

In a second related study, the researchers manipulated the order of the predictions between subjects by randomly assigning participants to forecast the outcome of one specific "focal" event either first or last within the sequence. That is, all participants again made predictions for the same set of three events, but researchers manipulated whether they predicted the outcome of the focal event before or after two other “nonfocal” events.

The results are likely to extend beyond gambling to other cases where people must make multiple forecasts, including decisions made by investors (who must make several predictions of how stocks will perform) and managers (who must forecast potential employees' performance before making hiring decisions).