Vandewater, E, Shim, M., Caplovitz, A (2004)  'Linking obesity and activity level with children's television and video game use', in Journal of Adolescence, Vol 27: 71 - 85.

There is a popular view that children's use of television and video games leads to obesity, but the 'empirical evidence for this link is mixed at best' (71), despite the existence of a number of studies. Public opinion seems in favour of this link, and suggests that it works in several ways:  (a) 'time spent with these media displaces physical activity'; (b) children increase their intake of foods as they view, sometimes as a result of food advertising -- but video games do not as yet contain adverts for snacks; (c) 'Television viewing actually decreases metabolic rates, more so than simply resting or sleeping'-- again, there is no suggestion that this might apply to playing video games (72).

If any of these or other mechanisms were in operation, there should be a positive correlation between television and video game use and increased weight or obesity  [an argument suggesting that correlations can at least falsify causal hypotheses, which is correct]. Some studies, cited in detail on page 73, show complex results -- weak associations and an absence of general effects. This disappointing finding could arise because of weak measures, especially of media use and other activities. If respondents are merely asked to estimate the amount of time they spend watching or playing, there can be considerable errors introduced.

This study draws on a 'nationally representative sample of children aged 1-12 using well-established measures of obesity (Body Mass Index)' (73), and also relies on time use diaries which more accurately represent actual behaviour.

A number of demographic characteristics were to be treated as covariates --  'child agenda... child race... age of child ... education of household head... family income to needs ratio  [to measure disposable income]. Children's time use was then assessed across two days, one weekday and one weekend. The 'primary caregiver' seems to have reported time use here. Children's media use was also recorded, broken down into television use, electronic game use computer use and print use. Children's activity participation was recorded via the diaries, and data were gathered for both sedentary and active leisure pursuits. Children's weight status was assessed according to the Body Mass Index (BMI), with parents collecting the data. The data were organised to create percentile rates within different categories of age and gender. Multiple regression analysis was then undertaken with these variables, with BMI as the dependent variable. Both linear and curvilinear relationships were examined  [more details on page 77].

The overall model  'showed no relation between children's weight status and television viewing. However,... there was a positive linear relationship between electronic game use and weight status... [and]... also a significant curvilinear relationship between electronic game use and BMI' (79). The latter curvilinear relationship showed that children with low BMI and children with high BMI played very little or a lot of electronic games -- it tended to be those with higher weight status who played moderate amounts of electronic games. There were complex gender affects too -- for boys,  'a linear component was nonsignificant, while the curvilinear component was significant in a negative direction'. For girls both components were significant, but in positive and negative directions respectively  (79). Relationships also varied according to age.

Similarly complex relationships -- negative linear and positive curvilinear components -- emerged for the other sorts of activities involved, including non game computer use and print use. Surprisingly,  'No relationships were found between children's weight status and either highly active or moderately active activities... [although]... a positive linear relationship between the time children spend in sedentary activities and their weight status... [was indicated]' (81).  [Whatever you make of these findings in detail, I hope is clear that there certainly is no straight forward simple relationship between playing video games and watching television and obesity].

Overall, television does not seem to be associated with obesity, but video game use is more strongly associated. However, these relationships are complex, being both linear and curvilinear, and they seem to interact with age and gender. It seems that the 'children with the higher weight status spent moderate amounts of time playing electronic games... while children with lower weight status spend either little or a lot of time playing electronic games' (81), but this does not seem to apply once children reached the age of nine. As most American children play moderate amounts of games,  'video-game use... holds some place in the story behind the increase in the prevalence of paediatric obesity in the United States' (81).  [This may make sense in statistical terms, but no sense at all in theoretical terms -- why should only moderate levels of video game use seem to have the greater effects? What possible theory might explain these curvilinear relationships?].

Finally,  'the data used... are correlational in nature and thus no causal inferences can be made' (81). In particular the positive relationship between playing games and being obese among girls might easily be explained by the view  'that girls who are overweight or obese might turn to electronic media because there are socially isolated... Indeed, it seems likely that in some cases, there may well be a causal direction from electronic game use to obesity... [as well as the other way about]' (81). Further research is clearly needed.

The findings also suggested some curvilinear relationship between use of print media and computers and obesity. This may simply show up the effects of sedentary activity. However, if these activities lead to 'information seeking and educational purposes... both parents and researchers [would approve]' (82).

Another finding indicates that despite a general link between sedentary activities and higher weight status, there was no converse link between highly active activities and low weight status. This is  'somewhat puzzling', and the other literature cited on page 82 is inconclusive as well. There is even some evidence that children who are overweight actually participate in more activities, perhaps as part of an effort to lose weight. However, 'our findings can be added to others indicating that there is indeed a link between physical inactivity and obesity among children' (82) [although any simple theories seem to be utterly ruined by these findings].

Television seems to be absolved as a factor in obesity, especially when the effects of socio-economic status and ethnicity were controlled. Again, this is  'somewhat puzzling' (83). It may be that only extreme television use has an effect, although again this maybe because  'obese youth may turn to television as the result of social isolation because of obesity' (83). There is a stronger relationship between playing electronic games and obesity, especially in children under the age of eight , although again other researchers do not agree. It would seem to be required logically that video game usage is replacing more physically demanding activities -- but watching television and playing video games seem to produce so different a set of results. Does video game usage displace more physically demanding pursuits than television usage? Is it a matter of affecting different age groups?  [And why is there no association between active pursuits and weight status?].

More research is needed. There are no quick and easy solutions to the problem of obesity, no magic bullets, but instead,  'a complex and interrelated pattern of factors' (83).

back to key concepts page