Abstract:
We study the nature of rounding in reports of probabilistic expectations by analyzing response patterns across numerous expectations questions and waves of the Health and Retirement Study (HRS). We discover a systematic tendency by about half of the respondents to provide more refined responses in the tails of the 0-100 measurement scale than in its center. In contrast, only about five percent of the respondents provide more refined responses in the center than the tails. We also find that rounding practice varies somewhat across question domains, which range in the HRS from personal health to personal finances to macroeconomic events. Using this evidence, we develop a two-stage algorithm to characterize person-specific rounding. The first stage uses each respondent’s response pattern across questions and waves to infer the extent to which the respondent rounds responses in each question domain and scale segment. The second stage replaces each original point response with an interval, representing the range of possible values of the respondent’s true latent belief implied by the degree of rounding inferred in stage one. Next, we demonstrate how the interval data thus obtained can be employed as either an outcome variable or a covariate in prediction analyses of substantive interest.
Host: Edwin Leuven