Prolonged sedentary behavior (too much sitting, as distinct from too little exercise) may pose risks to health, which are additional to those associated with physical inactivity (24,26,40). Television (TV) viewing is the largest contributor to adults’ leisure sitting time (5,35,40) and may be a marker for a broader pattern of an inactive lifestyle (36). Prolonged TV viewing, independent of leisure time physical activity levels, increases risk of overweight and obesity (12,35), metabolic syndrome (11,13), cardiometabolic risks (39), type 2 diabetes and undiagnosed abnormal glucose metabolism (10,19,20), poor mental health outcomes (16), and all-cause and cardiovascular mortality (9).
Health promotion strategies should thus consider reducing TV viewing time. Unfortunately, the average TV viewing time has been increasing in Australia and the United States, and adults also tend to increase their TV time as they get older (1,3). Increasing TV viewing time is therefore a common trend, with a potential of increasing health risk. Although cross-sectional studies in Australia, the United States, and Canada have identified correlates of prolonged TV viewing, which include lower educational attainment (2,6,33,34), unemployment (2,6,33,34), lower income (2,34), and higher body mass index (2,33,34), little is known about attributes associated with change in TV viewing over time. Evidence from longitudinal studies is needed to further understand characteristics of those who may be at risk.
A recent article proposed an ecological model of sedentary behaviors as a framework for research, highlighting multiple levels of influences on specific sedentary behaviors (27). For TV viewing, individual (e.g., demographic attributes, behaviors), social (e.g., social interactions, neighborhood socioeconomic status (SES)), and environmental (objective and perceived attributes of neighborhoods) factors may be relevant. However, the concurrent influences of such variables have not been examined empirically (27,40).
The present longitudinal study examined associations of baseline individual, social, and environmental characteristics with TV viewing time at follow-up, adjusted for baseline TV time; that is, it examined correlates of changes in TV viewing time.
Sample and Procedures
Data were from PLACE (Physical Activity in Localities and Community Environments), a study designed to examine associations between neighborhood environments and adult physical activity based on ecological models of health behavior (8,25). Self-completed surveys were administered at baseline in 2003–2004 and 4 yr later during the same months of 2007–2008 in Adelaide, Australia. All study procedures were approved by the Behavioural and Social Sciences Ethics Community of the University of Queensland.
A detailed description of the PLACE study is provided elsewhere (8,14,25). Briefly, a stratified multistage cluster sampling strategy was used to recruit participants from 32 neighborhoods chosen using walkability and neighborhood SES as selection criteria. Each neighborhood consisted of three to nine adjacent urban Census Collection Districts (CCDs). In each selected CCD, simple random sampling without replacement was used to select adult residents of private dwellings. Eligible individuals (age 20–65 yr, English-speaking, and able to walk without assistance) who agreed to participate and provided written consent were sent a questionnaire. A total of 2650 respondents from 154 CCDs returned the baseline questionnaire, accounting for 11.5% of the residential addresses initially identified and 74.2% of those who agreed to participate. The low response rate was partly due to using households as sampling units: many of the households identified were without eligible participants. Of those who completed the baseline surveys, 1098 (41.4% of the baseline sample) returned the 4-yr follow-up survey forms. The final sample for current analyses included 897 respondents with complete data on all variables. This sample (n = 897) did not differ from the total baseline sample (n = 2650) in demographic characteristics, physical activity, or TV viewing, except for the final follow-up sample being older at baseline (P < 0.01).
Self-report TV viewing time was measured at both baseline and follow-up using one item from a previously published scale assessing leisure time sedentary behaviors. This measure has shown good reliability and validity (5,33). Respondents reported the number of days they watched TV or videos as the main activity in the past 7 d and the average amount of time spent watching TV or videos on those days. Total TV viewing (min·wk−1) was calculated by multiplying the average amount by the number of days; daily TV viewing time (min·d−1) was then calculated by dividing the total by 7 (37).
Independent variables were selected based on the ecological model of four domains of sedentary behavior (27).
Participants reported their age, gender, educational attainment, employment status, annual household income, living situation, and the number of children (≤18 yr) in the household. Self-reported height and weight were used to calculate body mass index. The International Physical Activity Questionnaire long version was used to measure physical activity (7). The total amount of occupational, transport, leisure time, and domestic physical activity were calculated following the International Physical Activity Questionnaire protocols based on questions regarding frequency and duration of each type of activity in the last 7 d (21).
Social engagement-related variables, including sense of community, social interactions, and social cohesion, were assessed using validated instruments (28). Neighborhood SES was calculated from the median household income of the constituent CCDs (according to 2001 census data); the top and bottom quartiles of SES were then used to categorize neighborhoods as high or low SES.
Objective environmental attributes.
A walkability index derived from Geographic Information System (GIS) databases was used to quantify neighborhood features that facilitate walking (23). The index was calculated at the CCD level, as a composite measure of four environmental attributes (land use mix, dwelling density, net retail area ratio, and street connectivity) (23,25,29), and then was aggregated to the neighborhood level. Neighborhoods in the top and bottom quartiles of walkability were identified as highly walkable and lowly walkable, respectively.
Perceived environmental attributes.
Neighborhood pedestrian infrastructures, aesthetics, traffic-related safety, and crime-related safety were assessed using relevant subscales from the Australian version of the Neighborhood Environment Walkability Scale (4,30).
Change in TV viewing time was examined by modeling follow-up TV viewing time adjusted for baseline TV viewing time. This approach is equivalent to modeling change in TV viewing time and controls for regression to the mean. TV viewing time followed a gamma distribution; therefore, generalized linear models were used, specifying a gamma distribution and using a log link. Results from these models are reported as antilogarithms of the regression coefficients (and their respective 95% confidence intervals (CI)), indicating the expected proportional increases (for values > 1) or decreases (for values < 1) in TV viewing time associated with a one-unit difference in the correlates.
Initial models included baseline TV viewing time, all demographic variables (age, gender, educational attainment, working status, household income, living situation, and having children in the household), body mass index, domain-specific physical activities, social attributes, and all objective and perceived environmental characteristics. Backward elimination was used to remove variables that were not associated with the outcome. Baseline TV viewing time, age, and gender were left in the model regardless of significance level; other variables were removed if their significance exceeded 0.20 (38) and their removal did not affect the overall model fit (likelihood ratio test, P > 0.20). These liberal criteria were used to avoid type 2 errors in model building (18). Interactions of each independent variable with age, gender, and working status were then estimated, adjusted for all variables that were retained in the backward elimination.
The final model contained all variables retained in the backward elimination and all interactions significant at P < 0.05. Examination of the residuals in the final model revealed good model fit. There was no evidence of colinearity on examining correlation among independent variables and variance inflation factor.
Although the study had a multistage sampling design, there was negligible clustering in TV viewing time (intraclass correlation coefficient = 0.007, design effect = 1.11); therefore, analyses did not correct for clustering. Analyses were conducted using Stata 11 (Stata Corp., College Station, TX).
Baseline characteristics of respondents are shown in Table 1. A large proportion of respondents were women (61%) and some 45% were older than 50 yr. Slightly fewer than half (47%) had a tertiary education, and more than two-thirds (69%) were working. Some 23% of the respondents were living alone and 31% had at least one child living in the household. The overall baseline survey sample from which our participants were drawn underrepresented men, younger people, and those who were not working, compared with the adult population in Adelaide (25). At baseline, respondents reported a mean TV viewing time of 112 ± 92 min·d−1 and a median of 90 min·d−1 (interquartile range = 45.0–154.3 min·d−1).
At follow-up, the mean TV viewing time was 116 ± 90 min·d−1, which was increased by 4 min·d−1 from the baseline. Results from generalized linear models provided no evidence for associations of household income, living situation, children in the household, body mass index, or leisure time physical activity with follow-up TV viewing time (independent of baseline TV viewing). Most of the environmental variables examined were not significantly associated with follow-up TV time, including neighborhood aesthetics, traffic-related safety, crime-related safety, pedestrian infrastructures, social interaction, and social cohesion. Thus, these variables were not retained in the final model.
The final model in Table 2 shows that each additional hour of TV viewing at baseline was associated with a 45% higher TV time at follow-up. Adjusted for baseline TV viewing, follow-up TV time was 13% lower among those with a tertiary education, relative to those with no tertiary education. Each additional hour of baseline occupational and transport physical activity per day was associated with a 2% and a 7% lower TV viewing at follow-up.
Baseline domestic physical activity was also associated with TV viewing, but the association was different for men and women (P for interaction = 0.031). In men, each additional hour per day of domestic physical activity at baseline was associated with a 7% higher TV viewing time at follow-up; in women, there was no association between domestic physical activity and TV viewing time at follow-up.
A significant interaction existed between working status and neighborhood walkability (P = 0.035). Among those who were not working, living in a high rather than in a low-walkable area was associated with a 23% lower TV viewing time, whereas among those who were working, TV viewing time at follow-up did not differ between those living in high-and low-walkable neighborhoods. Alternatively viewed, among those who lived in low-walkable neighborhoods, TV viewing time at follow-up among those who were working was 17% lower than those not working (exp(b) = 0.83, 95% CI = 0.70–0.96). However, among those who lived in high-walkable neighborhoods, TV viewing at follow-up did not differ by working status (exp(b) = 1.04, 95% CI = 0.88–1.20).
For those with the same amount of domestic physical activity, there were no apparent gender differences in follow-up TV viewing time. Differences in TV viewing according to area SES were small (7%) and not statistically significant (P = 0.177). Older participants had higher follow-up TV viewing with a 10% higher TV viewing in those age 51–65 yr compared with those age 20–35 yr, but the difference was not statistically significant (P = 0.064). Sense of community was not significantly associated with follow-up TV viewing (P = 0.081); however, the effect size was not negligible (7% lower for each additional point on a five-point scale) and the 95% CI (0.87–1.01) could not rule out some sizeable effects.
This is the first longitudinal study to examine individual, social, and environmental correlates of change in adults’ TV viewing time, guided by an ecological model of sedentary behavior (27). For individual-level variables, higher education and working were associated with less increase in TV viewing during 4 yr. For behavioral variables, greater baseline occupational and transport physical activity were associated with a less increase in TV viewing time. However, domestic physical activity was associated with an increase in TV time (among men only). No social variables reached the statistical significance of 0.05, but higher sense of community had a marginal association with lower TV viewing time at follow-up. For environmental variables, those living in high-walkable neighborhoods had less relative increase in TV viewing time than those living in low-walkable neighborhoods, but only among those who were not working.
Lower educational attainment and not being employed have been identified consistently as correlates of prolonged TV viewing in cross-sectional studies in several countries (2,6,33,34,40). The current study also found that those with these attributes were more likely to increase TV viewing time during a 4-yr period. Those with lower SES may watch more TV because it is a cheap recreational activity. However, neither household nor neighborhood income had independent associations with TV viewing; thus, the underlying mechanisms are unlikely to be purely financial. For example, educational differences could reflect the various preferences and actions acquired through the experiences of daily life. Possibly, lack of a job may lead to more time for adults to engage in leisure time sedentary behaviors such as TV viewing (6). The modification of the association by neighborhood walkability is suggestive that behavior undertaken during the additional leisure time might depend on the surrounding environment.
Several previous cross-sectional studies have examined the associations of TV viewing time with total moderate-to-vigorous physical activity (6,17) or leisure time physical activity (22,36) and found no evidence for a consistent relationship. This study extended previous findings by examining associations of TV time with domains of lifestyle physical activity. More transport and occupational physical activity were associated with less increase in TV time, more domestic physical activity was associated with more increase in TV time among men, and there was no association between leisure time physical activity and TV time. It is possible that those who engaged in more occupational and transport physical activity spent less time at home and were therefore less likely to watch TV for long hours. It is also possible that those who were more active in their occupations and in transportation were more likely to have characteristics (such as better health or fitness) that enable them to be more physically active and less sedentary. The location where activity takes place may be important because greater baseline domestic physical activity was associated with a relative increase in TV time, at least in men. Domestic physical activity may serve as a marker of total domestic time, such that those who spend more time in the home might do more chores and yard work as well as watching more TV. However, it is unknown why such an association was only observed in men. One possible explanation is that men who do more chores at home might also spend more time at home and thus have more opportunities for TV viewing.
There are established relationships of neighborhood environmental attributes with physical activity (15,31,32). However, the relationship of the built environment with sedentary behavior including TV viewing is not well understood. The findings reported here are consistent with those of previous cross-sectional analyses of this study where TV viewing time was inversely associated with neighborhood walkability at baseline (37). A slight difference was that the association of walkability with change in TV time was evident only in those who were not working and thus may be more exposed to neighborhood environments. The finding may imply the importance of competing choices for leisure time activities and leisure time constraints. Walkable neighborhoods featuring mixed utilities offer more destinations (e.g., shops, cafes) and opportunities (e.g., community events) for activities, which may prompt individuals to spend more of their available leisure time outside the home and less time in TV viewing. Possibly the effect was more salient in those with more available leisure time (i.e., those who are not working).
Strengths of this study include the longitudinal design, use of a wide range of potential correlates, use of an ecological model as the guiding conceptual framework, objectively assessed neighborhood walkability, domain-specific physical activity, and the examination of potential moderators. One limitation is that low response rate and loss to follow-up not completely at random may incur selection bias that negatively affects both internal and external validities of current findings.
Future studies should further examine the individual, social, and environmental determinants of TV viewing time. The strength of the relevant evidence may be improved by using longitudinal study designs or natural experiments. Both objective and self-reported measures of physical activity and sedentary behavior should be included; objective measures are less prone to reporting biases, but reports can quantify a specific type or domain of behavior (e.g., transport walking). To advance understanding of correlates of sedentary behavior, more effort should be focused on identifying the mechanisms through which sociodemographic and environmental attributes can act to influence TV viewing time and other sedentary behaviors.
Ding, Owen, and Sugiyama were supported by National Health and Medical Research Council of Australia Program grant 569940, by a research infrastructure grant from Queensland Health, and by fellowship no. 1003960 (Owen).
There is no conflict of interest to report.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Australian Bureau of Statistics. Time Use on Recreation and Leisure Activities. Canberra (Australia): Australian Bureau of Statistics; 2006. p. 5.
2. Bowman SA. Television-viewing characteristics of adults: correlations to eating practices and overweight and health status. Prev Chronic Dis. 2006; 3: A38.
3. Brownson RC, Boehmer TK, Luke DA. Declining rates of physical activity
in the United States: what are the contributors? Annu Rev Public Health. 2005; 26: 421–43.
4. Cerin E, Leslie E, Owen N, Bauman A. An Australian version of the Neighbourhood Environment
Walkability Scale: validity evidence. Meas Phys Educ Exerc Sci. 2008; 12: 31–51.
5. Clark BK, Sugiyama T, Healy GN, Salmon J, Dunstan DW, Owen N. Validity and reliability of measures of television viewing time and other non-occupational sedentary behaviour of adults: a review. Obes Rev. 2009; 10: 7–16.
6. Clark BK, Sugiyama T, Healy GN, et al.. Socio-demographic correlates of prolonged television viewing time in Australian men and women: the AusDiab study. J Phys Act Health. 2010; 7: 595–601.
7. Craig CL, Marshall AL, Sjostrom M, et al.. International physical activity
questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003; 35 (8): 1381–95.
9. Dunstan DW, Barr EL, Healy GN, et al.. Television viewing time and mortality: the Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Circulation. 2010; 121: 384–91.
10. Dunstan DW, Salmon J, Owen N, et al.. Physical activity
and television viewing in relation to risk of undiagnosed abnormal glucose metabolism in adults. Diabetes Care. 2004; 27: 2603–9.
11. Dunstan DW, Salmon J, Owen N, et al.. Associations of TV viewing and physical activity
with the metabolic syndrome in Australian adults. Diabetologia. 2005; 48: 2254–61.
12. Dunton GF, Berrigan D, Ballard-Barbash R, Graubard B, Atienza AA. Joint associations of physical activity
and sedentary behaviors with body mass index: results from a time use survey of US adults. Int J Obes (Lond). 2009; 33: 1427–36.
13. Ford ES, Kohl HW 3rd, Mokdad AH, Ajani UA. Sedentary behavior
, physical activity
, and the metabolic syndrome among U.S. adults. Obes Res. 2005; 13: 608–14.
14. Gebel K, Bauman A, Owen N. Correlates of non-concordance between perceived and objective measures of walkability. Ann Behav Med. 2009; 37: 228–38.
15. Gebel K, Bauman AE, Petticrew M. The physical environment
and physical activity
: a critical appraisal of review articles. Am J Prev Med. 2007; 32: 361–9.
16. Hamer M, Stamatakis E, Mishra GD. Television- and screen-based activity and mental well-being in adults. Am J Prev Med. 2010; 38: 375–80.
17. Healy GN, Wijndaele K, Dunstan DW, et al.. Objectively measured sedentary time, physical activity
, and metabolic risk: the Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Diabetes Care. 2008; 31: 369–71.
18. Hosmer DW, Lemeshow S. Applied Logistic Regression. 2nd ed. New York (NY): Wiley-Interscience Publication; 2002. p. 65.
19. Hu FB, Leitzmann MF, Stampfer MJ, Colditz GA, Willett WC, Rimm EB. Physical activity
and television watching in relation to risk for type 2 diabetes mellitus in men. Arch Intern Med. 2001; 161: 1542–8.
20. Hu FB, Li TY, Colditz GA, Willett WC, Manson JE. Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA. 2003; 289: 1785–91.
22. King AC, Goldberg JH, Salmon J, et al.. Identifying subgroups of U.S. adults at risk for prolonged television viewing to inform program development. Am J Prev Med. 2010; 38: 17–26.
23. Leslie E, Coffee N, Frank L, Owen N, Bauman A, Hugo G. Walkability of local communities: using geographic information systems to objectively assess relevant environmental attributes. Health Place. 2007; 13: 111–22.
24. Owen N, Bauman A, Brown W. Too much sitting: a novel and important predictor of chronic disease risk? Br J Sports Med. 2009; 43: 81–3.
25. Owen N, Cerin E, Leslie E, et al.. Neighborhood walkability and the walking behavior of Australian adults. Am J Prev Med. 2007; 33: 387–95.
26. Owen N, Healy GN, Matthews CE, Dunstan DW. Too much sitting: the population health science of sedentary behavior
. Exerc Sport Sci Rev. 2010; 38 (3): 105–13.
27. Owen N, Sugiyama T, Eakin EE, Gardiner PA, Tremblay MS, Sallis JF. Adults’ sedentary behavior
: determinants and interventions. Am J Prev Med. 2011; 41: 189–96.
28. Parker EA, Lichtenstein RL, Schulz AJ, et al.. Disentangling measures of individual perceptions of community social dynamics: results of a community survey. Health Educ Behav. 2001; 28: 462–86.
29. Saelens BE, Handy SL. Built environment
correlates of walking: a review. Med Sci Sports Exerc. 2008; 40 (Suppl 7): S550–66.
30. Saelens BE, Sallis JF, Black JB, Chen D. Neighborhood-based differences in physical activity
: an environment
scale evaluation. Am J Prev Med. 2003; 93: 1552–8.
31. Sallis JF, Adams MA, Ding D. Physical activity
and the built environment
. In: Cawley J, editor. The Oxford Handbook of the Social Science of Obesity. New York (NY): Oxford University Press; 2011. p. 433–51.
32. Sallis JF, Saelens BE, Frank LD, et al.. Neighborhood built environment
and income: examining multiple health outcomes. Soc Sci Med. 2009; 68: 1285–93.
33. Salmon J, Bauman A, Crawford D, Timperio A, Owen N. The association between television viewing and overweight among Australian adults participating in varying levels of leisure-time physical activity
. Int J Obes Relat Metab Disord. 2000; 24: 600–6.
34. Shields M, Tremblay MS. Screen time among Canadian adults: a profile. Health Rep. 2008; 19: 31–43.
35. Shields M, Tremblay MS. Sedentary behaviour and obesity. Health Rep. 2008; 19: 19–30.
36. Sugiyama T, Healy GN, Dunstan DW, Salmon J, Owen N. Is television viewing time a marker of a broader pattern of sedentary behavior
? Ann Behav Med. 2008; 35: 245–50.
37. Sugiyama T, Salmon J, Dunstan DW, Bauman AE, Owen N. Neighborhood walkability and TV viewing time among Australian adults. Am J Prev Med. 2007; 33: 444–9.
38. Tabachnick BG, Fidell LS. Using Multivariate Statistics. 5th ed. Boston (MA): Allyn and Bacon; 2007. p. 328.
39. Thorp AA, Healy GN, Owen N, et al.. Deleterious associations of sitting time and television viewing time with cardiometabolic risk biomarkers: Australian Diabetes, Obesity and Lifestyle (AusDiab) study 2004–2005. Diabetes Care. 2010; 33: 327–34.
40. Tremblay MS, Colley RC, Saunders TJ, Healy GN, Owen N. Physiological and health implications of a sedentary lifestyle. Appl Physiol Nutr Metab. 2010; 35: 725–40.