What did they say? Respondent identity, question framing and the measurement of employment
Rosa Abraham, Nishat Anjum, Rahul Lahoti, Hema Swaminathan
Abstract:
In developing countries, a precise approach to measuring women’s employment remains elusive. Emerging evidence underscores the pivotal role of survey methodology, encompassing respondent selection and question framing, in shaping the assessment of women’s employment. Drawing from two labor market experiments in rural India, this study offers insights on the influence of survey design on the measurement of women’s employment. The first experiment contrasts self-reported women’s and men’s employment figures with proxy-reported data from spouses. Women’s self-reported workforce participation surpasses proxy-reported estimates by six percentage points, while men’s estimates exhibit negligible differences. There are significant differences in the type of employment activities reported by self and proxy for both women and men. These divergences emanate from asymmetric measurement errors, stemming from gender-based norm disparities between spouses, and divergent interpretations of employment. Additionally, information asymmetry between spouses concerning women’s marginal activities and disparities in spousal characteristics contribute to these self-proxy differences. The second experiment investigates if framing of questions and recall period has an impact on reporting of labor market outcomes. We find that employing multiple questions to capture weekly employment status yields a 10-percentage-point increase in reported women’s workforce participation, but men’s participation rate decreases by six percentage points. Furthermore, when a distinct employment query is directed at each day of the preceding week as opposed to a single query for the entire week, reported women’s workforce participation increases by seven percentage points, and men’s by four percentage points.