Understanding panel conditioning: an examination of social desirability bias in self-reported height and weight in panel surveys using experimental data

Abstract

Typically reliant on self-reports from panel data, a growing body of literature suggests that relative body weight can have negative effects on labour market outcomes.  Given the interest in the effects of relative weight in the social sciences, this paper addresses the question of whether repeated interviewing affects the quality of these data.  A theory that focuses on the sensitivity of the questions rather than the survey context is proposed.  Examining experimental panel data from Understanding Society using quantile-regression, the findings for women are consistent with the argument that conditioning reduces social desirability effects.  The ameliorative effects of panel conditioning on social desirability bias in self-reported height and bodyweight appear to strengthen the association between relative weight and employment for men, but not women, however.