The Computing Scale Co, Burnaby Village Museum by Kenny Louie
Title: TV Viewing and BMI by Race/Ethnicity and Socio-Economic Status
Author(s): Kerem Shuval, Kelley Pettee Gabriel, Tammy Leonard
Publisher: PloS ONE
Year: 2013
URL: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0063579
Topics: obesity, health outcomes, television viewing, socioeconomic status and race
Methods: Regression analysis of the Health Information National Trends Survey (HINTS), a nationally representative survey of people over aged 18 from across the United States about their communications and knowledge about healthcare and cancer.
What it says: This study uses data from HINTS to understand the association between obesity and television among adults when considering socioeconomic status and race/ethnicity. The authors use BMI as the primary dependent variable and TV viewing, in average number of hours watched per day, as the primary independent variable. They controlled for race/ethnicity and socioeconomic status using variables for race and whether or not respondents completed a college degree and whether or not they had health insurance. They also controlled for age, gender, marital status, number of children, and variables related to respondents’ level of physical health.
They found that the odds of being overweight increased as respondents entered the third and fourth quartiles of television watching across races and socioeconomic statuses, but that the strength of the effect varied with race and socioeconomic status. For example, while they find an increased tendency toward obesity with more television viewing among non-Hispanic whites, the employed, and those with insurance, the effect is not statistically significant in the case of Hispanic and black respondents, unemployed respondents, and those without health insurance. Both college graduates and non-graduates were at increased risk for obesity in the fourth quartile of TV viewing.
What I think about it: The authors note a few important limitations to interpreting the results. The sample size for racial subpopulations is fairly small, which might disguise a real effect due to small sample size. They note that their findings are cross-sectional and do not account for how respondents’ relationship to watching television might vary over time. Additionally, they note that they do not consider factors other than television that might indicate a level of sedentary tendencies.
In terms of the significance of the findings, I think that we need more evidence to understand how television viewing and obesity relate to socioeconomic status and race. Additionally, I wonder if a more helpful statistical approach would be to examine how the increased likelihood of watching television among a particular group might make that group more prone to obesity, instead of looking at how the same amount of television viewing might lead to an increased likelihood of obesity.
What it all means: Increased television viewing leads to a higher likelihood of obesity, perhaps because of an increased tendency toward sedentary behavior. While the results of the study suggest that increased television viewing is not necessarily associated with increased obesity for all racial and socioeconomic groups, the fact that people in certain subgroups watch more television might make them more likely to become obese if we think that the relationship is causal.