摘要：Objectives. We evaluated the association between television viewing and pedometer-determined physical activity among predominantly racial/ethnic minority residents of low-income housing in metropolitan Boston in 2005. Methods. We used mixed models to analyze the association between reported hours of television viewing and pedometer-determined steps per day among 486 adults. We also examined whether television viewing was associated with the achievement of 10000 steps per day. Results. There was a mean 3.6 hours of average daily television watching. In multivariable analyses, each hour of television viewing on an average day was associated with 144 (95% confidence interval [CI]= −276, −12) fewer steps per day and a decreased likelihood of accumulating 10 000 steps per day (odds ratio [OR]=0.84; 95% CI=0.71, 0.99). Weekday and weekend television viewing were each also associated with fewer steps per day. Conclusions. Average daily television viewing was associated with reductions in total pedometer-determined physical activity levels (approximately 520 steps per day) in this lower-income sample. As part of a comprehensive physical activity promotion plan, recommendations to reduce television viewing should be made. Television viewing is a highly prevalent sedentary activity in the United States; on average Americans watch at least 4 hours of television each day, 1 and television viewing is the most time-consuming activity nationally—after work and sleep. 2 Watching television has been positively associated with excess body weight among both children and adults, 3 – 13 perhaps because the time that might otherwise be used to engage in physically active pursuits is used to watch television programs, and/or because media content promotes unhealthy dietary choices that negatively influence the foods that television viewers eat. Surprisingly little research has systematically investigated the hypothesis that television viewing is inversely associated with physical activity among adults; there is, however, emerging evidence supporting the relation. 5 , 8 , 14 , 15 The literature investigating the association between television viewing and physical activity has exclusively utilized self-reported physical activity measures. Although commonly administered, the validity of self-reported physical activity estimates may be compromised by recall bias. This may be a particular problem among lower-income populations, who derive a comparatively higher proportion of their daily physical activity through participation in routine, non–leisure-time activities (e.g., domestic, occupational, and transportation) 16 , 17 that are more challenging to recall, compared with less incidental, higher-intensity leisure-time activities. With 1 exception, 8 previous studies have investigated the impact of television viewing on leisure-time physical activity, thus limiting the ability to derive estimates of how television viewing affects total physical activity. We sought to evaluate the association between hours spent watching television and pedometer-determined physical activity in a sample of predominantly racial/ethnic minority residents of low-income housing. This population is of key concern given their high rates of television viewing, 9 , 18 , 19 physical inactivity, 20 , 21 and related health consequences.
