Age at first birth, education and late life self-rated health

Maja Djundeva, Erasmus University Rotterdam

This study relates the self-rated health of older European women ages 50 through 80 to parenthood, age at first birth and education, using data from the second and third wave of the Survey of Health and Retirement Europe (SHARE and SHARELIFE) collected in 2006/07 and 2010 (N=11585). Using a Heckman selection model I first account for selection into parenthood based on combined measure of childhood socioeconomic and health status. A joint Heckman selection model in a structural equation framework is used to test two scenarios: a) age of first birth mediates the effect of education on health but also has an independent effect on health; or that b) age of first birth only mediates the effect of education on health, and does not have an independent effect on health. Results from the pooled sample over countries show that show that the indirect effect of education on health (ß=.0018 SE=.0005 CI=(.0008; .0027) is very small, whereas the total effect of education and AFB on health is substantively bigger. ß= .0296 SE=.0028 CI=(.0241;.0350). It seems that AFB has an independent main effect following a parabolic association on self-rated health and that AFB does mediate the relationship between education and health.