Effects of anecdotes in belief change processes: A dynamic decision making study

When making daily decisions, people have been relying less on statistical and more on anecdotal evidence. Anecdotes are oftentimes the first information we hear about medical issues, climate crises, financial decisions; even news reports use anecdotes to engage the reader. While much research has investigated people's preference for anecdotal versus statistical evidence, how people combine these different types of evidence is not well understood. In a dynamic decision making task, we investigated how people combine vivid and nonvivid anecdotes with either congruent or incongruent statistical evidence. We asked participants to play in a horse race betting game in which we presented them with anecdotes about prior race results to form a prior. Then, we sequentially displayed specific race results (statistical evidence). On each trial participants were asked to make a guess about a horse after evaluating some evidence. First, they read an anecdote about the past race performance of another horse belonging to the same breed. Then, they also received the ranking of same breed horses from three separate races (i.e. statistical information). Participants either made a decision immediately after or asked for additional race results. Unexpectedly, we found that people relied more on statistical cues than anecdotal evidence. However, anecdotes were not completely disregarded because when anecdote and statistical cues were incongruent, participants requested additional race results. We also noted that a majority of decisions were congruent with the valence of the last cue participants requested, suggesting that they may be accumulating evidence to pass a decision threshold.