Title: Do the Psychosocial Risks Associated with Television Viewing Increase Mortality? Evidence from the 2008 General Social Survey – National Death Index dataset
Author(s): Peter Muenning, Zach Rosen, Gretchen Johnson
Publisher: Annals of Epidemiology
Topics: television viewing habits, mortality risk
Methods: regression analysis of GSS and National Death Index datasets, discrete time hazard modeling
What it says: Previous research has established that increased hours spent watching television is associated with a higher rate of poor health indicators, even among those who exercise regularly. Scholars have suggested that a number of factors related to television viewing could be related to these poor health outcomes, such as a sedentary lifestyle or the ability of television programming or advertising to induce particular emotional states. Additionally, people who watch more television participate in fewer social activities and have less social capital than people who watch more television, factors that are also linked to poorer health.
The authors use a discrete time hazard modeling approach to calculate the mortality risk of watching television. Their sample included respondents who reported being in good health to the General Social Survey from 1978-2002, with follow up on death through 2008 using data matched to the National Death Index. In an effort to explore the potential of factors like heightened optimism or a positive outlook on life that might compensate for the psychological effects of watching television, the authors included mediating variables from the GSS, including questions such as how strongly participants agreed with the statement “people can be trusted” or “I have little control over the bad things in my life.” Only respondents who reported being healthy at the time of the survey were included in the sample so that the results wouldn’t be skewed by people who watch TV because they are in poor health.
The results indicate that increased television viewing is associated with increased mortality risk when controlling for age, gender, race, graduation from high school, income, work status, and year of interview. The results, if causal, indicate that each hour of television watched on average per day decreases life expectancy by about 4%, or 1.2 years of life expectancy in the United States. The authors thus conclude that “just halving television viewing time in the US could conceivably increase national life expectancy by over half of a year.” None of the psychological factors were significant at mediating health risks individually, but were significant when considered all together.
What I think about it: The key question here is whether the strong association between television viewing and shorter lives is caused by watching television or by something else. The authors make a robust effort to control as many factors as the dataset permits, leaving no obvious alternative explanations for the results reported in the study. Nevertheless, it is possible that some important variables that would help explain the association are left out of the GSS and, by extension, this analysis. Additionally, the authors’ finding that certain psychological factors are correlated with both higher mortality rates and television viewing may undermine the case for causality somewhat – in other words, people’s pessimism might cause them both to watch more television and live shorter lives. Given the importance of causality to the policy relevance of the findings, the paper would have been strengthened by a more robust discussion on that topic; it mentions potential limitations but neglects to make an affirmative case for assuming a causal relationship despite the authors’ seeming confidence that one exists.
It seems like health as a component of wellbeing tends to be associated with television habits across a number of studies, but other psychological factors are less strongly associated with negative impacts of watching television. That all of the psychological factors taken together had a statistically significant impact on mortality is interesting and suggests that a strongly positive outlook on life may mitigate the negative effects of television. However, given the relatively weak associations it’s unclear how much weight to put on this aspect of the analysis, especially since the GSS does not measure exercise or eating habits directly.
What it all means: Television is associated with an increased risk of mortality in the United States, with a possibility that some psychological factors may mediate the overall increased mortality risk. If we’re thinking about whether or not television is good or bad for society, these findings would seem to put television pretty firmly in the “bad” camp.
That said, the finding that psychological factors, when considered all together, mediate some of the poor health effects of television might support the idea that people experience different effects from watching television, some healthy and some unhealthy. If some frequent television watchers are able to maintain a generally positive outlook without a negative impact on health, it might be because they because they are choosing to watch television and find some value in the experience that doesn’t negatively affect their health. It may also indicate that attitudinal factors in general are more important for determining health outcomes than the behavioral practice of sitting in front of a TV.
Without a clear understanding of whether we should interpret these results as causal, it is difficult to know whether this evidence in the “bad” category is sufficient to outweigh evidence that television is a healthy activity for some. Despite limiting the sample to those who reported being in good health, if psychological factors and optimism have an impact on mortality risk when holding hours of watching television constant, the direction of the causality and the main takeaway from these findings are not completely clear.