By Alain Samson, Ph.D.


Games of chance are a crucible for human biases, errors and fallacies.  Two popular concepts in people’s perception of gambles are the hot hand and gambler’s fallacies, which may occur in a series of events with random binary outcomes, such as coin tosses.

People succumb to the hot hand fallacy if they believe that, in a sequence of independent events, a person’s chances of winning are better after a previous win rather than a loss – “lucky streaks”.

The gambler’s fallacy is the other side of the same coin. It’s the belief that the last outcome of a series is unlikely to occur again. If the roulette ball has already landed on red several times, some people (excluding most Psychology Today readers, of course) may think that it’s more likely to land on black next time.

New research by Sandra Laporte and Barbara Briers, published in the Journal of Consumer Research, investigates these fallacies from an interpersonal perspective. They show that being similar to a previous winner can have radically different effects on people’s participation likelihood in sweepstakes – it all depends on the attributions people make for the winning outcome.

One of the reported studies was done as part of a an actual sweepstake offered to a company’s customers. The researchers sent two versions of an email inviting customers to participate in the sweepstake. One version included the previous winner’s first name, Philippe, while another (neutral) version made no such mention. Results showed that actual customers named Philippe participated almost twice as frequently when they received the “Philippe” email (.99%) than when they didn’t (.50%).

To demonstrate the role of attributions in these outcomes, the researchers conducted a study that also measured people’s belief in luck. Unlike chance, luck is more about personal attributes that may influence events. In this experiment, gender was used as the similarity variable (the two last winners’ names were either female or male). On average, people were almost twice as likely to participate in the sweepstake when their gender matched that of previous winners. This difference was significant for respondents with high belief in luck, but disappeared for those with low belief in luck.

According to the authors, highlighting a previous winner implicitly fosters some kind of “human cause” behind sweepstake outcomes, which leads to an ‘interpersonal hot hand’ effect. But what happens if people are reminded that these events are purely a matter of chance?

This question was investigated in two additional studies. One experiment primed half of the participants with the idea of randomness by asking them to construct sentences out of several words that also contained words like ‘chance’ or ‘random’. Another experiment made randomness salient by depicting the sweepstake as a virtual roulette game. As hypothesized, the researchers found the opposite effect of similarity under those conditions – an ‘interpersonal gambler’s fallacy’. The intention to participate or attractiveness of the sweepstake was lower among persons whose gender matched that of previous winners.

The results of this research are not only interesting from an academic perspective, but also reflect the practice by state lotteries or sweepstakes companies to depict past winners as “real people” that may share the target population’s demographic characteristics or lifestyle. This practice makes participation more attractive by making a potential win both tangible (see availability heuristic) and personal. However, as Laporte and Briers’ research shows, the effect of similarity may reverse if people’s attributions change. Under those conditions, anonymity or dissimilarity would encourage greater participation.