This paper considers whether survey respondents' views regarding the likelihood of stock index returns exceeding specific thresholds are comparable to market views indicated by index options with strikes at analogous thresholds. It is motivated by the observation thatthe wording used to elicit subjective beliefs in surveys about expected future returns resembles the question a purchaser of a call option might ask. Building on this association, we document a similarity between the views of survey respondents and those of financial market participants as measured through call options, although the association is not one-for-one. We find a closer association for those demonstrating a better understanding of the laws of probability, suggesting that numeracy affects the accuracy of an elicited response.

Expectations about stock market movements are an important factor in models of investments and savings. To better understand consumers' financial behavior, economic surveys such as the Health and Retirement Study (HRS) ask participants to report expectations about the stock market. However, readability statistics suggest that the HRS' stock market expectations questions use relatively complex wording, which may contribute to their relatively high rates of missing responses. Here, we build on survey design research to improve the readability of these questions. In two experiments using national online panels, we test whether revising stock market expectation questions to reduce their difficulty affects respondents' 1) survey experience, as measured by percent of missing answers and ratings of question clarity and difficulty, and 2) response validity, as assessed by respondents' confidence in their answer and comparisons between expectations and stock market outcomes. In both studies, our revisions improve survey experience. Unfortunately, revisions also decreased the perceived (Study 1) and actual (Study 2) validity of responses. We discuss implications of question revisions for the design of economic surveys.