You Can Be Too Thin (But Not Too Tall): Social Desirability Bias in Self-Reports of Weight and Height You Can Be Too Thin (But Not Too Tall): Social Desirability Bias in Self-Reports of Weight and Height

By Mary A. Burke and Katherine G. Carman

In the United States, as in many other countries, self-reported values of weight and height are sometimes used to estimate obesity rates. Despite the need to obtain accurate measures of body mass index (BMI) and obesity rates in order to guide public health policy, it is widely recognized that self-reported values of weight and height are often inaccurate. A leading explanation for what causes these errors contends that self-reported values are influenced by prevailing social norms for weight and height by gender. If so, this would represent an instance of "social desirability bias," a broader phenomenon in which survey-takers bias their answers in the direction of socially desirable responses, such as reporting that one voted in a recent election when in fact one did not. .. To date, however, evidence of social desirability bias in self-reported values of weight and height is not conclusive, in part due to the absence of a precise framework describing how social norms might influence self-reporting behavior. To fill the gap, this paper proposes a theoretical model in which an individual's self-reported body weight is a function of the person's true weight and an exogenously determined social norm; essentially, the respondent faces a tradeoff between reporting a truthful value and a socially desirable one. The model generates testable predictions that can be compared with predictions based on alternative explanations for self-reporting errors. These various predictions are tested using a sample from the National Health and Nutrition Examination Survey that records both self-reported values of body weight and height and examined values for the same individuals. The empirical analysis is also used to make inferences about social norms for BMI.

see more

up down Key Findings