Predicting the weather or the economy? How to make forecasts more trustworthy

"When a forecaster predicts that something has a high likelihood of happening, consumers infer that the forecaster is more confident in her prediction, that she is basing the prediction on more in-depth analysis, and that she is more trustworthy," write the authors of the study, Rajesh Bagchi (Virginia Tech) and Elise Chandon Ince (University of South Carolina). "The prediction is also judged to be more accurate."

The authors conducted eight studies in reaching those conclusions. The studies dealt with predictions such as which team would win a particular basketball game, whether a stock price would go up or down, and whether a book would get published. In the studies, participants were asked to react to high predictions--a 70% chance that something would happen--and low predictions--a 30% chance that something would happen.

The authors found that the participants made judgments about the forecaster and the prediction based on the prediction itself. When the prediction was high (70%) instead of low (30%), the forecaster was judged to be more confident and trustworthy; she was thought to have conducted more research; and the prediction itself was deemed more accurate. That is because participants focused on the likelihood of the target event occurring--say, Team A wins the basketball game--than on a complementary event occurring--Team B wins the game.

"These results have practical implications. It could be, for instance, that consumers prefer information outlets such as news programs and websites that make higher predictions. Thus, political pundits, meteorologists, and stock analysts may want to highlight their predictions when they offer a higher forecast for an event, as a way to help build their reputation," Bagchi and Ince write.