When I first began studying what I later dubbed “impersonal influence,” I approached it from the perspective of someone attempting to shed light on a pernicious phenomenon. As a graduate student interested in media's impact on the political process, media's role in portraying the opinions of large collectives seemed obviously detrimental to the democratic system. I sympathized with those who sought to curtail the publication of poll results and those who wanted to abandon the sequential primary system because of its dynamic component. Surely these practices detracted from the independence and rationality of people's political decisions. The emergence of the sociotropic voting studies only confirmed my concerns, in this case about media portrayals of mass collectives. If people ignored their own personal experiences and instead voted on the basis of media-derived perceptions of collective experience, surely the accountability of public officials was suffering as a consequence.

It was not until I decided to combine my various studies into a book and sat down to write the conclusion chapter that I realized just how far my thinking had evolved. In fact, close readers may notice that although I draw on some data from my previously published studies, in many cases the conclusions I draw here are quite different from the conclusions that I drew in my journal articles on these same topics. There is a very simple explanation for these discrepancies: I changed my mind.