Since the 1970s, when the prevalence of childhood obesity began its steep climb, much research has sought to understand what drives this epidemic. During this period, observers have argued that declines in food prices, increases in mothers’ labor force participation and changes in the built environment are all likely contributors to the growth in childhood obesity. Another significant societal change occurring in the same period has received much less attention as a potential factor: the decline in the average number of children per family.
Between 1965 and 2008, the number of children per family (hereafter, family size) decreased from 2.44 in 1965 to 1.86 in 2008, according to data from the U.S. Census Bureau. However, the relationship between family size and childhood obesity in the United States has been generally unexplored. Previous studies have skirted around the question. Research has examined the link between family size and educational, labor market and some health outcomes. Only a few studies have looked at the link between family size and child body weight. These studies, however, have been mostly correlational and have therefore been unable to say whether larger family size affects a child’s body weight or whether family size and body weight are determined by third factors (e.g. family socioeconomic factors, geographic location, tastes and preferences) that aren’t always observable. Moreover, many are based on non-U.S. data, which may not be generalizeable to the U.S. context. Furthermore, no study has clarified the mechanisms linking family size and childhood obesity.
My recent paper in Social Science and Medicine aimed to fill these gaps. Using data from a nationally representative cohort of children in the United States and employing several complementary approaches (e.g. inclusion of a rich set of controls, estimating school fixed-effects and child fixed-effects models, and tons of sensitivity analyses), I was able to examine – (1) whether family size can influence child body mass index (BMI, which is defined as Weight in Kg/Height in m2) and likelihood of being obese, and (2) the mechanisms that link family size with child BMI and obesity, including physical activity, maternal work, family meal times, and dietary behaviors among others.
My findings suggest that having more siblings is associated with lower BMI and decreased likelihood of obesity. For every additional sibling, there is a 2.6 percentage point decline in the likelihood of obesity in early adolescence. Based on these estimates, a back-of-the-envelope calculation suggests that the decline in average family size in U.S. households between the late 1960s and 2008 (from 2.44 to 1.86; Source: U.S. Census Bureau) may account for up to 13% of the increase in childhood obesity during the same time period (from 5% to 17%). But why exactly?
The results suggest that having siblings is associated with better diet quality and less television watching, both of which are well-known risk factors for childhood obesity. A factor associated with better diet for children with more siblings might be the higher likelihood of eating meals at home together as a family in larger families. While we cannot tell from the data the nutritional value of different types of meals, existing research suggests that eating at home is associated with better nutritional outcomes than eating out. Because there are economies of scale in home food preparation, larger families may have a greater incentive to prepare and eat meals at home, which means better meal quality for children with more siblings. In addition, children in larger families are more likely to be supervised by adults, which also reduces the likelihood of choosing unhealthy foods and sedentary activities. At the same time, having siblings is also associated with less vigorous physical activity, but the data’s lack of measures on moderate or other physical activity means we cannot establish the link between number of siblings and children’s overall physical activity.
While my study finds a robust link between more siblings and lower BMI and obesity, it is after all an observational study too, and therefore cannot establish causality. However, there is enough evidence to suggest that children in small families (i.e. fewer siblings) in the United States may very well be at higher risk for obesity. Given that identification of children in small families is straightforward, this could be an easy target for obesity-prevention programs.