2176 PURPOSE: Self-reports of television watching behaviors are frequently used as an indicator of inactivity, particularly in large health surveys. It is possible however to be very active, yet still frequently watch television, suggesting that television watching may be an incomplete indicator. METHODS: We examined data from the 1999–2000 National Health and Nutrition Examination Survey (NHANES) to determine the prevalence of reported television watching behaviors by reported physical activity, using SUDAAN to account for the complex survey sample design. RESULTS: Among 5,448 respondents (n = 2,898 women, 2,550 men) ages 18–85, 44.1% of women and 43.9% of men reported watching 3 or more hours each day of television. When crossclassified by physical activity (inactive, irregularly active, or active at CDC/ACSM recommended physical activity levels), the prevalence of watching television 3+ hours a day was 50.0% among inactive respondents, 42.8% among those who were irregularly active and 37.3% among those active at recommended levels (p = 0.0001). When stratified by body mass index, the prevalence of reported television watching 3 or more hours each day was: BMI < 25 kg/m2: 39.5%; BMI 25–29.9 kg/m2: 42.2%; BMI 30+ kg/m2: 50.9% (p = 0.0001). Although 44% of respondents reported watching television more than 3 or more hours each day, a direct association between television watching and body mass index was observed when the data were stratified by physical activity levels. Among respondents who were active at CDC/ACSM recommended levels and had BMI < 25 kg/kg/m2, 31.5% watched 3+ hours of TV, compared to 36.8% of overweight respondents and 49.5% of obese respondents who were active at the recommended level. CONCLUSIONS: Television watching behavior appears to be a useful, if incomplete, indicator of inactivity.