Are female employment statistics more sensitive than male ones to questionnaire design? Evidence from Cameroon, Mali and Senegal

This paper investigates the effect of several survey questionnaire characteristics on employment statistics. It also assess the differences in sensitivity to survey design across gender and living area. Indeed, as suggested in the literature, women, especially those living in rural areas, are expected to be more sensitive than men to survey design, due to both the nature of the work (seasonal, occasional, temporary, informal, unpaid family work) and social norms. In many African countries, labor force surveys are not available on a regular basis and the way existing household surveys and census measure employment differs greatly, both over time and between countries. This makes it difficult to properly study labor market dynamics and to draw meaningful policy recommendations. Using about fifty surveys and censuses collected in Cameroon, Mali and Senegal between 1976 and 2012, we first review the diversity of survey instruments used and highlight the key questionnaire characteristics that are likely to affect employment statistics. Exploiting within-survey variations of the wording of questions, the detail of the labor module and the length of the reference period, we then assess the effect of these features on labor statistics. Empirical results shows significant effects of each questionnaire feature and suggest that women are not systematically more sensitive than men to survey design, nor is it the case for rural individuals compared to urban ones.