J. Exp. Psychol. Appl. 23, 386–402 (2017)

Because climate change projections are based on past climate data and future assumptions, they are not certainties. Rather, projections are used to derive probabilities of climate change outcomes. The IPCC provides specific guidelines for communicating outcome likelihood, mapping verbal probability terms (“unlikely”) to specific numerical likelihoods (0–33%). However, these guidelines do not account for biases in the way people use and understand probability information.

Marie Juanchich and Miroslav Sirota from the University of Essex examined how people form climate change predictions based on frequency distributions. When asked to make predictions about “unlikely” or “possible” sea-level rise by 2100 based on a distribution of 100 projections, participants generally selected extreme maximum outcomes. Participants also rated the outcome they chose as 20–35% more likely than that outcome’s actual probability. Predictions were less extreme when verbal probability terms were replaced with or accompanied by numerical probability values, but this did not reduce the gap between prediction frequency and perception of its probability. These results suggest that people have a preference for predicting extreme outcomes from climate change and overestimating their occurrence.