Famine Exposure and Adult Height: Disentangling Stunting from Survival Selection

Cheng Huang, Emory University

We estimated the stunting effect of famine exposure on the adult height of female survivors of the Chinese famine of 1959-61. To disentangle the stunting from the selection effect, we adopted a quantile regression approach and adjusted for the birth size of babies born to these women. We found that estimates at the 90th percentile of adult height always exhibited a larger stunting effect than those at the 10th percentile and that adjustment for offspring birth size magnified the stunting effect estimates at the 10th percentile but not at the 90th percentile, which suggested that attenuation bias caused by selection for survival occurred mainly at the lower quantile of adult height; therefore, estimating the impact on the upper quantile of adult height with quantile regression may serve as a better approach to capture the potential stunting effect than conventional mean regression.