“There is a 1 in 10 chance of catching Ebola”: the 1-in-X bias in probability judgment is independent of outcome severity

Abstract

ABSTRACT Possibilities presented as a 1-in-X ratio (e.g. “1 in 100” chance of contracting a disease) are judged more likely than an equally probable N-in-X*N ratio (e.g. 10 in 1000). This is known as the 1-in-X bias. It is also known that high severity outcomes (e.g. catching Ebola) are judged more likely to occur than low severity outcomes (e.g. catching Lyme disease). This is known as the severity bias. What is not known, is how these two effects behave in combination. This is important as the 1-in-X ratio can be used to communicate the likelihood of both high and low severity outcomes. Two preregistered experiments (N = 1,318) addressed this question by manipulating ratio format (1-in-X vs. N-in-X*N) and severity (Low vs. High) in a fully crossed design. The effect size of the 1-in-X bias on probability judgements did not differ as a function of severity. Instead, the two biases were independent and cumulative, with greatest overestimation of probability occurring when a high severity outcome was communicated using the 1-in-X format. The 1-in-X bias in likelihood judgement did not extend to binary decision-making (e.g. to hypothetically accept or refuse treatment) adding to a literature of small and inconsistent effects in this domain.