Even the sciences are vulnerable to dishonesty at precisely the point where one least expects it – quantitative analysis. For eleven years John Bailar (chair of the department of epidemiology and biostatistics at McGill University) served as statistics consultant to The New English Journal of Medicine, during which time he reviewed nearly four thousand articles. He chronicles how scientists practice deliberate deception through the selective reporting of data – their version of half-truths. They accomplish this, for example, by failing to inform readers of the weak spots in their data, selecting data in ways biased to their own interests, failing to give credit to earlier work or placing reliable data in a context that causes readers to draw misleading (usually optimistic) conclusions about the success and significance of the project being reported.W. Jay Wood
Epistemology: Becoming Intellectually Virtuous.