Professor of Decision Sciences
Subjective risk perceptions are often encoded as responses to 0-1 questions in surveys or other qualitative risk scales. However, reference points for assessing an activity as risky are confounded by various characteristics of respondents. This paper uses a sample of workers for whom quantitative risk assessments as well as dichotomous risk perception responses are available. It is shown that, given a quantitative risk measure, the thresholds for assessing an activity as risky vary systematically, particularly by education. The differences in such thresholds across worker groups are estimated. The resulting implications of using qualitative risks variables for assessing wage-risk trade-offs are estimated, yielding results that are also relevant for many other areas involving similar qualitative variables.