Childhood overweight in the United States: A quantile regression approach David C. Stifel *, Susan L. Averett Department of Economics, Lafayette College, Easton, PA 18042, USA 1. Introduction Childhood overweight is associated with depression, lower cognitive ability, and a myriad of public health problems (Dietz, 1998; United States Office of the Surgeon General, 2001; Schwimmer et al., 2003; Daniels, 2006; Averett and Stifel, 2008). The strain that these conse- quences of childhood overweight will place on the health- care system highlights the need to understand the various causes of childhood overweight in order to inform policy makers as to the best way to reverse this trend. This paper adds to the growing literature on the causes of childhood overweight (Anderson and Butcher, 2006) by using quantile regression methods along with ordinary least squares (OLS) regression to estimate the correlates of child weight status in the United States. Previous research (Chou et al., 2004; Lakdawalla and Philipson, 2002; Anderson et al., 2003b; Cutler et al., 2003; Komlos and Baur, 2004) has used OLS to explain central tendencies or variations in weight outcomes for the average child. Since the average child in the United States is not considered overweight by standard measures, it is not clear that this econometric method fully addresses the issue of interest. Overweight children are generally at the upper tail of the weight distribution, not in the middle. Quantile regressions permit analysts to estimate the determinants of weight status at any percentile of the weight distribution. While probit methods do estimate these relationships at the upper tail of the distribution (Anderson et al., 2003b; Cutler et al., 2003), they discard important and available information, and require stronger assumptions about the error term than is necessary given that the ‘‘latent’’ variable is not latent, but is observed. 1 We use quantile regressions to illustrate that model estimates for blacks and whites by gender differ along the Economics and Human Biology 7 (2009) 387–397 ARTICLE INFO Article history: Received 24 July 2008 Received in revised form 13 May 2009 Accepted 15 May 2009 JEL classification: C13 I12 I18 J1 Keywords: Childhood overweight Obesity Underweight Quantile regression ABSTRACT The prevalence of overweight children in the United States has increased dramatically over the past two decades, and is creating well-known public health problems. Moreover, there is also evidence that children who are not overweight are becoming heavier. We use quantile regression models along with standard ordinary least squares (OLS) models to explore the correlates of childhood weight status and overweight as measured by the Body Mass Index (BMI). This approach allows the effects of covariates to vary depending on where in the BMI distribution a child is located. Our results indicate that OLS masks some of the important correlates of child BMI at the upper and lower tails of the weight distribution. For example, mother’s education has no effect on black children, but is associated with improvements in BMI for overweight white boys and underweight white girls. Conversely, mother’s cognitive aptitude has no effect on white boys, but is associated with BMI improvements for underweight black children and overweight white girls. Further, we find that underweight white children and black girls experience similar improvements in BMI as they get older, but that for black boys there is little if any association between age and BMI anywhere in the BMI distribution. ß 2009 Elsevier B.V. All rights reserved. * Corresponding author. Tel.: +1 610 330 5673; fax: +1 610 330 5715. E-mail address: stifeld@lafayette.edu (D.C. Stifel). 1 Ravallion (1996) makes this point when he criticizes the use of binary response models to estimate the determinants of poverty when unit data is available for incomes. Contents lists available at ScienceDirect Economics and Human Biology journal homepage: http://www.elsevier.com/locate/ehb 1570-677X/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.ehb.2009.05.005