The Use of Narrative Evidence and Explicit Likelihood by Decision Makers Varying in Numeracy

Decision makers are often presented with explicit likelihood assessments (e.g., there is a 10% chance that an attack will occur over the next 3 months) and supporting narrative evidence in forecasting and risk communication domains. Decision makers are thought to rely on both numerical and narrative information to the extent that they perceive the information to be diagnostic, accurate, and trustworthy. In two studies, we explored how lay decision makers varying in numeracy evaluated and used likelihood assessments and narrative evidence in forecasts. Overall, the less numerate reported higher risk and likelihood perceptions. In simple probabilistic forecasts without narrative evidence, decision makers at all levels of numeracy were able to use the stated likelihood information, although risk perceptions of the less numerate were more affected by likelihood format. When a forecast includes narrative evidence, decision makers were better able to use stated likelihood in a percentage as compared to frequency or verbal formats. The more numerate used stated likelihood more in their evaluations whereas the less numerate focused more on the narrative evidence. These results have important implications for risk analysts and forecasters who need to report the results of their analyses to decision makers. Decision makers varying in numerical ability may evaluate forecasts in different ways depending on the types of information they find easiest to evaluate.