Probability matching and statistical naïveté

People often probability match: they select choices based on the probability of outcomes. For example, when predicting 10 individual results of a spinner with 7 green and 3 purple sections, many people choose green mostly but not always, even though they would be better off always choosing it (i.e., maximizing). This behavior has perplexed cognitive scientists for decades. Why do people make such an obvious error? Here, we provide evidence that this difficulty may often arise from statistical naïveté: Even when shown the optimal strategy of maximizing, many people fail to recognize that it will produce better payouts than other strategies. In 3 preregistered experiments (N = 907 Americans tested online), participants made 10 choices in a spinner game and estimated the payout for each of 3 strategies: probability matching, maximizing, and 50/50 guessing. The key finding across experiments is that while most maximizers recognize that maximizing results in higher payouts than matching, probability matchers predict similar payouts for each .