Close
OPRU Briefing Paper
Executive Summary
Suggestions have been made by food and advertising industries that the effects of screen advertising on children’s dietary intake are small. As part of the Obesity Policy Research Unit (OPRU) work programme, a review of the evidence of the quantitative effect of screen advertising on children’s dietary intake and obesity was undertaken, a secondary aim of which was to identify the source of this suggestion, specifically that advertising accounts for 2% of the variance of dietary intake in children.
After conducting systematic searches and tracking-back of various references of the 2% figure, we concluded that the most likely source is a frequently cited paper ‘Modeling the impact of television food advertising on children’s diets’ [1]. This paper reports that television food advertising accounts for 2% of the variance in children’s snacking frequency.
The article by Bolton (1983) used complex statistical modelling techniques to examine the impact of television food advertising on children’s diet. The paper sought to address the effects of screen advertising, parental influence and child demographic factors on children’s diets by using a framework which incorporated environmental, social and behavioural influences.
The model estimated that parental behaviour was the most important direct influence of children’s behaviour and accounted for 29% of the variance in children’s snacking frequency, compared to advert exposure which accounted for 2%; however, TV adverts were reported to reduce children’s nutrient ‘efficiency’, implying children may substitute high nutrient low calorie foods for low nutrient high calorie foods.
We feel that there are major limitations to this work:
The model was constructed using data from a small cross-sectional survey conducted in Cleveland, Ohio. The sample (n=262) comprised children aged 2-11, the majority of which
- were white and of high socioeconomic status (SES); children of lower SES in developed countries are more likely than their wealthier counterparts to consume food of lower nutritional value but children from low SES groups were not represented in this study.
- The survey was conducted in 1977, since which time there have been substantial changes both in terms of marketing effectiveness and strategy, and the global food market.
- This was a highly statistical analysis from limited data; the amount of television watched by the group was low and it’s likely that the level of obesity in the group was also low. The way the 2% estimate was calculated is extremely unclear and there is insufficient information given to judge the validity of this approach. The indirect effects of television advertising are not accounted for and the paper later reports the proportion of variance explained for children’s TV advert exposure as 23% of the variance - very different to 2%.
- There were limitations with the model - exposure to advertising was estimated by self-reported television viewing time which is likely to be subject to inaccuracies. Without an accurate measure of advert exposure, deriving conclusions regarding its impact on dietary intake is unlikely to be reliable.
We conclude that there are major limitations to this work, not least in the lack of clarity in how analyses were conducted and apparently different potential conclusions, meaning that caution is required when interpreting the findings. This casts doubt on assertions that the effects of screen advertising on children’s dietary intake are necessarily small.
