...scientific papers serve the needs of their authors above those of their readers. Why else would a journal devote five pages to a paper that reached this conclusion? “In this pilot study, the null hypothesis that both treatments will show equal results cannot be confirmed or rejected because of the small number of participants.” Michael O’Donnell, BMJ 2005;330:256 (29 January), doi:10.1136/bmj.330.7485.256-a. [Note: I would link to this, but BMJ now hides most of its content after the first week of publication and re-opens it after one year. The paper he is quoting from is Lamers HJ et al. Br J Gen Pract 1998;48: 1819-23.]