In Reply: Although we assessed sources of funding in our sample, there are a number of reasons that we did not perform statistical analyses comparing the level of spin according to funding source, as suggested by Drs Allison and Cope. First, this analysis was not prespecified in our protocol. Second, the funding source was not reported in 37.5% of the reports, and it is highly likely that the reporting of funding source is not random. Third, the comparison could be performed on different variables because funding could be categorized into 3 classes (for-profit; nonprofit or no funding; not reported) or in 2 classes (considering data not reported as associated with for-profit funding or with nonprofit funding). Fourth, there were multiple possible outcomes (presence of spin in the article, presence of spin in the conclusion, extent of spin, or other). We did not want to run several different analyses and choose the results according to our preconceptions.
Boutron I, Ravaud P, Altman DG. Randomized Controlled Trials With Statistically Nonsignificant Results—Reply. JAMA. 2010;304(9):965. doi:10.1001/jama.2010.1246