Anna Folke Larsen, University of Copenhagen
"Misreporting Month of Birth: Implications for Research on Nutrition and Early Childhood in Developing Countries".
A large literature relies on accurate measurement of birth dates to estimate the impacts of events experienced in utero and early childhood on subsequent health and development outcomes, such as height-for-age Z scores (HAZ). Using 163 Demographic and Health Surveys with almost one million children from 62 countries, we document two systematic anomalies in HAZ relative to recorded ages. The most important is a gap in HAZ between December- and January-born children, such that January-born children tend to have markedly lower HAZ than those born in December. We also find a gap in HAZ between children with round ages and children just below a round age, such that a child aged two years and eleven months tends to have a lower HAZ than a three-year-old child. These observations have previously been interpreted as actual effects of health shocks, particularly seasonal shocks. We show that these patterns arise irrespective of agroclimatic conditions but are linked to calendar type, parental education and use of birth registration cards. Using a simulation model, we show that the average December-January gap in the DHS data of -0.32 standard deviations can be explained by assigning eleven percent of measured children with a purely random month of birth. With this model the observed December-January gap can be used to estimate the extent of age misreporting in any survey that collects information on both heights and ages.
Contact person: Jakob Roland Munch