Recently, I had the pleasure of attending the Publication Bias Workshop in London, organized by the NC3Rs, a British scientific organization.
The meeting was organized around the question of whether preregistration can be a solution for the problem of publication bias in medical research. I believe that it can, and I've been writing about this for several years. I didn't speak at the workshop, but the cause of preregistration was amply represented by speakers including Chris Chambers of the journal Cortex and Trish Groves of the BMJ. However, in the formal debate which concluded the first day, our side lost support in the audience vote by a fair margin. Whereas before the debate, a show of hands showed the great majority of the attendees to be in favor of preregistration, the vote after the debate revealed that many had changed their minds. What happened? In my opinion, the problem was that the discussion was dominated by the image of ClinicalTrials.gov, the central preregistration system for clinical research. There seemed to be an assumption, on both sides of the debate, that something like ClinicalTrials.gov would be the model for the proposed future system of preregistration of preclinical (basic) research. The opponents then argued convincingly that such a centralized site (what I think of as a "PreClinicalTrials.gov") would be expensive, bureaucratic and unworkable, given the vast amount of research that it would have to deal with. Yet this isn't how it has to be. Preregistration wouldn't have to follow the ClinicalTrials.gov pattern. Speaking for myself, while clinical trials certainly served as a big inspiration when I began writing about these issues, it was never my only model. Three years ago I outlined four ways of implementing preregistration: a central registry like ClinicalTrials.gov; journal-based preregistration; funding agency based preregistration; and 'personal' preregistration, in which people use their own web-space or blogs. The centralized approach was just one option. And in the years since I wrote that, many pioneering experiments of preregistration in basic research have arisen, some semi-centralized (but still, very different to ClinicalTrials.gov) such as the Open Science Framework (OSF), and others using decentralized models, such as the journal-based approach (Registered Reports), or personal preregistration. So we shouldn't worry that we don't have the infrastructure or the resources to create a 'PreClinicalTrials.gov', because we don't have to do that. Preregistration is simply the idea that if you set out to investigate something, and you might want to publish the results, then you should publicly state what you plan to do at the outset, not just after the fact. Any system that makes that possible is a system of preregistration.
