Is Being in School Better? Using School Starting Age To Identify the Impact of Schools on Children’s Obesity

Year: 2009

Research Center: Department of Nutrition at the University of California, Davis

Investigator: Anderson, Patricia, Elizabeth Cascio, Kristin F. Butcher, and Diane Schanzenbach

Institution: Wellesley College

Project Contact:
Kristin F. Butcher
Department of Economics
Wellesley College
Wellesley, MA 02481
Phone: 781-283-2179


In recent decades, the increase in childhood obesity has been stark, with rates tripling from 5 percent in the early 1970s to 15 percent by the early 2000s. This increase in childhood obesity raises many concerns about children’s current and future health and well-being. Although adult obesity has been rising over this same period, recent research finds that increases in parental obesity cannot fully explain the increase in children’s obesity, suggesting that the increase in children’s obesity does not merely reflect changes within the family. Child-specific environments, like schools, may play an important role in children’s health outcomes. Given that children spend a great deal of time in schools and consume about 30-50 percent of their daily calories in schools, it is important for policymakers to understand the role of schools in determining children’s obesity outcomes.

Research into the specific role that schools play in determining obesity rates tends to show that what happens in school matters. For example, an increase in availability of junk food leads to an increase in student body mass index (BMI). Research into school lunch consumption reveals that children who regularly eat the school lunch (as opposed to bringing lunch from home) are about 2 percentage points more likely to be obese.

These research findings suggest that the school food environment—eating the school lunch or having access to junk food—affects children’s weight outcomes. This study shows that some school environments are better than others. A second related, but different, question is whether being in school, as opposed to being out of school, is better or worse for children’s weight outcomes. This research adds to this second line of inquiry. Further, if being in school is better for children’s weight outcomes, is it better for the outcomes of all children? In particular, this research examines whether the impact of being in school is different for students who do and do not participate in various Government programs, where participants are typically poorer than nonparticipants.

The goal in this research is to examine the causal impact of being in school, as opposed to not being in school, on children’s obesity and BMI. The main methodological challenge is to be able to compare children who have and have not been exposed to additional schooling, but who are the same in other ways that affect their weight outcomes. We used data from the Early Childhood Longitudinal Study, Kindergarten Cohort (ECLS-K) and the National Longitudinal Survey of Youth, Mother-Child matched samples, to get information on children’s height and weight, years of schooling at a given age, and a large set of other individual and family characteristics.

In order to compare otherwise similar children who have different exposure to schooling at a given age (in other words, to get exogenous variation in exposure to schooling), we used a child’s exact birth date and the school starting-age laws in the child’s State to determine whether a child who is 6 years old, for example, should be expected to be in kindergarten or first grade. Consider a child in a State with a school starting-age cutoff of October 1: a child who is 5 on September 2 will start kindergarten and have 2 years of school exposure when he is observed in the data at the age of 6. A child with this birthday in a State with a September 1 cutoff will start kindergarten the following year and will only have 1 year of school exposure at the age of 6. By using the child’s birthday relative to his State’s school start-age cutoff date, we compared individuals who have more and less exposure to schooling at a given age, where this variation comes solely from birthdays and State laws, not from parental choices about when to start their children in school. In addition to using variation in State laws and birthdays, we controlled for a rich set of family and individual characteristics.

This study first demonstrates that a positive relationship between school exposure and obesity is spurious because of parents’ reluctance to send small-stature children to school “on time.” Turning to the estimates that compare only those students who are exposed to more and less school by virtue of their birthdays and their States’ starting-age cutoffs, the study finds that children exposed to an additional year of schooling have obesity outcomes that are not statistically different from students exposed to less schooling. However, the causal effect of school exposure varies significantly by subgroup. For example, an additional year of school exposure benefits weight outcomes for children who do not eat school lunch, who are not food stamp recipients, and who have not been on the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). By contrast, school exposure does not appear to be causally related to weight outcomes for children who do participate in these programs, and these children will typically be less well-off than children who do not participate in these programs.

For children participating in these programs (poorer children), additional school exposure is not related to weight outcomes, suggesting that there is little difference in opportunities for caloric intake and expenditure between their school and alternative (for example, home or daycare) environments. However, for students who do not participate (better off children), school and alternative environments offer different opportunities for caloric intake and expenditures, with the school environment leading to better weight outcomes.

Direct inquiries about this study to the Project Contact listed above.