Gebrenegus Ghilagaber , Stockholm University
Studies on associations between education and health-outcome have traditionally used woman's education (alone or in combination with partner's education) as regressors in models relating covariates to the outcome. In this study, we adapt models developed in the social mobility litearture to examine effects of differences between couples' educational levels on women's health-related decisions (such as the propensity to deliver in health facilities). Both conventional modeling and Diagonal Reference Modeling (DRM) which accounts for origin (woman's education), destination (partner's education), and 'mobility' (differences between couples' educational levels) are applied on data from Demographic and Health Surveys (DHS). Results from conventional models reveal strong effects of educational differences on women's health-related decisions; but such strong effects disappear when data is analyzed using DRM. Recent works in the area (Fosse & Pfeffer, 2019) argue this may be an artifact because DRM tends to generate results that implicitly force the mobility effect to zero.
Presented in Session 5. Health & Mortality 1