As a clinical pharmacologist (G.D.N.) and biostatistician (R.S.C.) involved in the design, analysis, and reporting of hundreds of controlled clinical trials, we found the 2 editorials on the publication of raw data for clinical studies to be quite thought provoking.1,2 Dr Wilhelmus proposes that making data sets available might decrease the potential for misconduct and allow for greater potential pooling for meta-analyses. Drs Beck and Ferris respond to these suggestions, stating that while actual computational errors and inappropriate analyses may be an issue, more often it is the actual protocol design or study conduct that is of greater concern. Regarding the use of these large data sets for manuscript review, they point out the problems with finding the large amounts of skilled person-power required to perform such analyses.
Novack GD, Crockett RS. Submission of Data Sets to Journals: What's the Real Issue?. Arch Ophthalmol. 2000;118(12):1706. doi: