Spin vs. Science: Fast Food Toys

Parsons over at Change.org has been cheerleading a San Francisco measure to ban giving away toys with fast food. The hope is that doing so will reduce obesity both currently (by reducing demand for fast food) and in the future (toys, it is hypothesized, make fast food more addictive and habit-forming). Since then, the San Francisco's Board of Supervisors' Land Use Committee has approved the measure, which will now go before the full board for final approval.

Parsons spins a few things in her favor in the process. She misquotes an op-ed which itself distorts the science it cites (Chou, Rashad, and Grossman, 2008, Journal of Economics and Law). The study compares children in 1979 and in 1997 (NLSY) using variation in the time spent watching fast food commercials to account for obesity. The data on fast food ads excludes national advertising (no variation), relying instead on metropolitan area codes; data on 1979 3-11 year old television watching comes from mothers while data on 1997 adolescent television watching is self-reported. Pairing those together provides the estimate of the amount of time spent watching fast food commercials. When BMI is the dependent variable, the estimate of advertising's effect is always positive, but is statistically significant only 2/12 times. The results on the probability of being overweight are somewhat stronger statistically, but this explains only 10% of the variation.  It seems rather strong to claim -- as Parsons and to a lesser extent the authors -- that we could reduce child obesity by 18% by banning something that explains 10% of the variation.

There is no simple direct obvious causation statement that can or ought be made from the research, though Parsons has. CRG cite, for example, the Institute of Medicine arguing that "the final link that would definitively prove that children had become fatter by watching food commercials aimed at them cannot be made." CRG admit that "Clearly, our estimate of the probability that a given child saw a certain message is subject to error" and that "As indicated in Section 1, there is conflicting evidence on trends in television and commercial viewing by children and youths since 1980. Hence, it would be premature to point to our findings as a partial explanation of the upward trend in obesity."

The conclusion of the study puts more conditionalities into the discussion:

Clearly, we have not provided enough information to fully evaluate the two policies just discussed. Indeed, we have not addressed the larger issue of whether the government should intervene in the food purchase decisions of its citizens. In the case of children, one justification for government intervention is that society as a whole may reap substantial current and future production and consumption benefits from improvements in children’s health. The case is strengthened because overweight children are extremely likely to become obese adults and because children are less likely to have information about the consequences of their actions or to heavily discount these consequences. The case is weakened because parents may more easily and immediately affect the choices made by their children than can the government.

In addition, one would need to consider the degree of government involvement that is merited and the costs of alternative policies if some intervention appears to be worthwhile. Hence, more research is required to evaluate the effectiveness of these policies and others. Our study should be viewed as one of many inputs in this process.

On top of all this, none of it deals with the toys themselves. How much would nixing the toys change the effectiveness of the ads?

There are a lot of big leaps being made from the carefully worded, well done research and the spin being used to promote the change. That's part of the difficulty in doing nutrition policy advocacy: how overwhelming does the evidence have to be before we're willing to stake policy on it?