Just watched this excellent video from the inimitable Ben Goldacre:
It pretty much highlights some of the problems I discussed in the last post on the excess of positive results in academic research. Choice quote: “But this problem of negative results that go missing in action is still very prevalent. In fact, it’s so prevalent that it cuts to the core of evidence-based medicine”.
My expertise is not in medicine, but it would make sense if clinical trials were registered based on their methodology before any study took place. This would then provide a reference point for independent-confirmation of results by different laboratories. It doesn’t necessarily solve the withholding of negative results — a laboratory may choose not publish data if it has a vested interest in the drug under study (Goldacre highlights this point in his talk concerning the registration of trials). I guess change can only come from shifting the culture of doing science: that is, we place a renewed emphasis on devising methods and testing hypotheses (look, here is a drug that may or may not help fight cancer better than what’s currently available), rather than the pressure to publish positive results (look, my drug works really well, but I didn’t realise there were all these previous negative results and, in truth, it’s only better than a placebo).