Dr Duncan raises the issue of scientific evidence for a social phenomenon and also raises the issue of bibliographic citation. Unfortunately, science cannot address itself to complex, multifaceted social phenomena. A reductionist (scientific) approach in an open system cannot be rigorous. Instead, the tools of social science apply, with some help from epidemiologic principles. Here, common sense leads one to a probable cause-and-effect relationship. Listed are references1,2 that Dr Duncan may find useful. Although they do not rigorously establish the matter, I believe there are few who would disagree that advertising of formula has a negative influence on breast-feeding.
Finberg L. The Effect of Lay Advertising on Breast-feeding Prevalence-Reply. JAMA. 1990;264(16):2074. doi:10.1001/jama.1990.03450160042018