I forgot my wordpress login, and I can’t find the time to dig it out, so I’m posting this comment here, rather than on the target blog. David Crotty has a good post discussing the tension between null results and reproducibility. Mu comment on this is that I like to think of this in terms of Bayseian approach to understanding the world, and in that context I feel that the getting over the file-drawer problem could help. In this view the position one takes is not to take a blanket stance on trusting unpublished work, nor a blanket stance on not trusting published work, but rather to move the null results into the public domain and thereby provide more overall evidence on the trustworthiness or otherwise of a specific piece of knowledge. The more overall, weighted evidence, that we have, the more trustworthy our conclusions can become.
The trick may be to find a way to encourage people to part with these null results with low cost to their reputations, and low transactional costs to getting them out there. Celebrating null results could help with the former, sites like figshare and zenodo are helping with the latter.