Michael R. Eber

JOB MARKET PAPER

Stated preferences and numerical ability: Implications for public policy analysis

Abstract: Stated preference surveys that assess the value of goods and services inform regulation and broader policy implementation related to public health, the environment, and many other domains. This paper examines how estimates of willingness to pay (WTP) derived from stated preference surveys are influenced by respondents’ numerical ability. Using experimental data from over 10,000 people, I show that WTP estimates are often insensitive to the magnitude of benefits and that this insensitivity is largely explained by respondents' numerical ability. Difficulty understanding numerical information leads to context-specific biases, at times overstating WTP by more than 40%. Bias is generally larger among socioeconomic groups with low education and income levels. These results indicate that WTP elicitations often misrepresent people’s preferences and can confound preference estimation across socioeconomic groups. I provide a set of assumptions and methods that practitioners can use to statistically correct for bias to improve the consistency and representativeness of WTP estimates.