If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

# Is obesity contagious?

## Problem

A recent study found that obesity tends to spread like a “contagion” through a social network. In other words, when a person experiences weight gain, close friends in the same networks tend to gain weight as well. The investigators conducted a detailed analysis of a mass network of 12,067 people who had been closely followed over 32 years, from 1971 to 2003. In the study, 5,124 people were used as key subjects, or “egos”, whose behavior was analyzed. Any persons linked to the egos serve as “alters”—those who may influence ego behavior. The researchers examined several aspects of obesity spread, such as clustering of obese persons within the network, association of weight gain among an individual’s social contacts, degree of dependence of association of social ties, and influence of gender or geographical distance. The researchers found that there were discernible clusters of obese persons (BMIis greater than or equal to30) in the network at all time points.
Figure 1 shows some results from the study. The extent of interpersonal association in obesity was evaluated with regression analysis. Homophily was taken into account by including a measurement of the alter’s obesity. The researchers evaluated the possible role of unobserved contemporaneous events by separately analyzing models of subgroups of the data involving ego-alter pairings. In particular, three types of “directional” friendships are defined: 1) an “ego-perceived friend” in which the ego identifies the alter as a friend; 2) an “alter-perceived friend” in which the alter identifies the ego as a friend; 3) a “mutual friend” in which the identification is reciprocal. Familial ties (parents, siblings) and marital ties (spouses) are treated as reciprocal. “Immediate neighbor” denotes the geographical distance between an alter and an ego. For example, from the results, we can see that if an ego stated that an alter was a friend, the ego’s chances of becoming obese appears to increase by 57, percent (“risk of obesity”).
Figure 1: Percentage increase in obesity risk for an ego based upon his/her relationship with an alter. The dependent variable in each model is the obesity of the ego. Independent variables include a time-lagged measurement of the ego’s obesity, the obesity of the alter, a time-lagged measurement of the alter’s obesity, the ego’s sex, age, and education. Mean effect sizes (solid black dot) and 95, percent confidence intervals (line) are shown.
In later studies on the relationship between social networks and health behaviors, one of the researchers further found that existing social ties (especially close friendships) are more likely to dissolve between people who have health traits that are dissimilar, including health traits that are immutable such as height and personality, and traits that are mutable such as BMI, blood pressure, etc. In particular, those with similar BMIs are less likely to dissolve existing ties and more likely to form ties. Another study demonstrated that food choices also were made in accordance to social networks. In particular, spouses showed the strongest influence in food consumption behaviors, controlling for social contextual factors. Across all peers (spouses, siblings, friends), eating patterns that were most likely to be shared were “alcohol and snacks”.
Source: Adapted from Christakis, N. A., & Fowler, J.H. (2007). The Spread of Obesity in a Large Social Network over 32 Years. The New England Journal of Medicine. 357(4), 370-379. O'Malley, J., & Christakis, N. A. (2011). Longitudinal analysis of large social networks: Estimating the effect of health traits on changes in friendship ties. Statistics in Medicine. M. A. 30(9), 950-964. Pachucki, M.A., Jacques, P.F., & Christakis, N.A. (2011). Social Network Concordance in Food Choice Among Spouses, Friends, and Siblings.American Journal of Public Health, 101(11), 2170-2177.
Which conclusion is best supported by the findings in Figure 1?