Exploring selection bias and the WIC program: Confronting a methodological challenge to child health program evaluation

Fifty-five percent of infants born in the US participate in the Supplemental Food and Nutrition Program for Women, Infants and Children (WIC). Previous research finds that women participating in the program experience shorter average breastfeeding durations than women who do not participate in the program, even when controlling for potential confounders. Criticisms of this finding center on whether there is positive or negative selection bias into the program. Selection bias presents an important methodological challenge to studying this child health program. If women who are more likely to breastfeed are also more likely to enter the program, than the magnitude of previous associations is understated. If women who are least likely to breastfeed are the most likely to enter the program, then the magnitude is overstated. Existing literature uses observable demographic characteristics, such as race, education and income, to produce a profile of those least likely to breastfeed. This paper is the first in the literature to use prenatal attitudes towards breastfeeding and perceived social support as markers of prenatal proclivity to breastfeed. Data was obtained from the United States Centers for Disease Control and Prevention and the Food and Drug Administration. The Infant Feeding Practices Study II is a longitudinal study of over 5,000 women beginning in late pregnancy and continuing through the baby's first year. The prenatal attitudinal factors from this study were combined into a "proclivity score", and mothers were ranked based on their prenatal commitment to breastfeeding. Results show that prenatal commitment to breastfeeding negatively predicted selection into the program, but that this negative selection bias was salient only for second and third time mothers. A positive, but insignificant, relationship was found among first-time mothers. Limitations of this study, implications for future research and implications for public policy are discussed.