together. As anticipated given the research methods and the use of television viewing as a measure of exposure to advertising, the quality of the measures was medium or low, never high, and, with the one experiment as an exception, causal inference validity was never high and most often was low. On the other hand, with two exceptions, ecological validity was never low and was most often high because the research methods and measures permitted assessment of everyday life phenomena. Also as anticipated, ecological validity tended to be higher when causal inference validity was lower (see Table 5-18). The results for the relationship of marketing to diet-related health do support generalization to everyday life, but do not make it easy to determine cause and effect.
The preponderance of results (51 of 74, 69 percent) reported a significant association between marketing and diet-related health. Examination of the data in Tables 5-16, 5-17, and 5-18 indicated that there is no apparent difference in the characteristics of the research producing significant and nonsignificant results. Further exploration of possible differences based on breakdowns using multiple variables did not change this finding. Moreover, the two relevance ratings (see Table 5-17) indicated that the significant results were from research that was at least as relevant as the research that produced the nonsignificant results: For causal inference validity, 28 percent of the significant results were rated medium (one was high) compared to 22 percent of the nonsignificant results, and for ecological validity, 86 percent of the significant results were rated high compared to 83 percent of the nonsignificant results. This general pattern did not change when causal inference validity and ecological validity were considered together (see Table 5-18).
As Table 5-16 shows, available evidence about the relationship of marketing and diet-related health is almost entirely evidence about the relationship of television advertising exposure and adiposity among children and teens. In contrast to the evidence about the relationship of marketing to precursors of diet, or to diet itself, in which two or three different types of outcomes have been studied, there is only one outcome that has been studied to any degree: adiposity. Also, only one type of marketing has been studied: exposure to television advertising. To determine what can be concluded about the relationship of television advertising exposure to adiposity in children and teens, one result (Wong et al., 1992) that was not about adiposity was omitted from further consideration. Of the 73 results remaining in the evidence table, the infant/toddler sample in the Vandewater et al. (2004) study and the cardiovascular fitness outcome in the Guillaume et al. (1997) result were both removed from further consideration.