Sociological interpretations of science, extending from Kuhn's (1962) Structure of Scientific Revolutions to the social constructivists of the 1980s, have enlivened the traditional debate about the foundations of science and the very idea of scientific progress. But there is perhaps no place to see the central flaw of such interpretations better than in academic medicine, for here the importance of reliable reporting of the data from empirical testing reveals clearly the distinction between genuine and sham science. In his paper 'Ghosts in the Machine: Publication Planning in the Medical Sciences', Sergio Sismondo (2009) exposes the larger background of marketing and promotional activities in the vast network of pharmaceu tical industry ghostwriting. By having infiltrated the ranks of marketers he brings a unique perspective to their activities, and his report on this expe rience is valuable. Outside of litigation one seldom sees the inner workings of this process. Sismondo, however, makes the serious mistake of claiming that pharmaceutical industry-sponsored research and ghostwriting produce genuine knowledge and science (albeit commercial science) not different from established medical science. He cites David Bloor, Karin Knorr Cetina, Andrew Pickering, and Harry Collins to support this claim and says that science is 'choice laden', implying in this case that it matters little whether it is produced by academic scientists or spun in marketing strategies of public relations firms. It is widely acknowledged by the leading philosophers of science that sci ence is 'choice laden' in the sense that as authors of scientific theories we cre ate hypotheses by which experience is interpreted; but by subjecting these hypotheses to rigorous tests, we discover which ones are falsified and which ones will serve as our best, tentative solutions to problems (Popper, 1959). It is this crucial role of genuine and rigorous testing that demarcates real sci ence from its impostors. Thus unlike Sismondo, I do not condone the activ ities of commercial science as merely part of a constructivist view of science, one that fails to distinguish between better and worse, genuine and sham.