D1 All values are x SD. There were significant effects of television viewing vegetables (P 0.0001) in the third-grade sample and for percentages of energy fifth-grade sample. There were significant effects of day (ie, weekday or weekend and for percentages of energy from fat (P 0.05), fast food (P 0.01), fruit ( P significantinteractionsbetweentelevisionviewinganddayforpercentageofenergyfromfruit from soda (P 0.0001) in the fifth-grade sample.
snacks were consumed more frequently during television viewing than were any of the meals, and children may have more autonomy or choice in their snacks than in their meals. Accordingly, children may not choose vegetables for snacks during television viewing.
Although advertising has been shown to influence children’s food choices (12–14), our results do not support the hypothesis that children consume more highly advertised foods while watching television. In the diverse third-grade sample, significantly less soda and fast food were consumed on weekdays during television viewing than during other times of the day, and in the mostly Latino fifth-grade sample, significantly less soda and significantly fewer sweets and snacks were consumed on weekenddaysduringtelevisionviewingthanduringothertimesofthe day. These results are counter to those of previous research that showed that more soda, snack foods, and pizza were consumed by children in households in which the television was on for 2 meals/d (7), and this discrepancy in results highlights the differences between the ecologic study design and our direct correlation design. for percentages of energy from soda (P 0.01), fast food (P 0.001), and from sweets and snacks (P 0.01) and vegetables (P 0.0001) in the day) for percentage of energy from soda (P 0.01) in the third-grade sample
0.01), and vegetables (P 0.05) in the fifth-grade sample. There were (P0.05) inthethird-gradesampleandforpercentageofenergy
In the third-grade sample, the type, but not the amount or frequency, of foods consumed during television viewing was associatedwiththechildren’scurrentweightstatus.Thechildren who consumed more energy from fat while watching television had higher BMI than did their peers who consumed fewer highfat foods with the television on. A post hoc analysis showed no significant correlations between the number of servings from each food category consumed during television viewing on weekdays and the children’s BMI, but on weekends, soda consumption during television viewing was significantly associated with BMI (r 0.26, P 0.01).
There are some limitations to the methods used in these studies. First, although 24-h dietary recalls are recognized as the best method to estimate dietary intake in children (15–22, 29) and are the most appropriate method to determine what children eat whilewatchingtelevision,theintraindividualvariationindietary intakes makes multiple recalls necessary to accurately estimate energy intake (30). However, prior to our primary analysis, we detected weekend-weekday differences in food consumption with the television on or off and, therefore, analyzed weekend