It sounds like a good idea when Mr. Bing Crosby croons it, but it’s not good for science. Negative studies are every bit as important as positive ones, but they are much less likely to be published. There are many reasons for this, but one of them (positive outcome bias by manuscript reviewers) is carefully examined in an interesting study found in the November 22 issue of the Archives of Internal Medicine (Emerson GB, et al. Arch Intern Med 2010;170:1934).
The investigators fabricated two versions of a manuscript (a “positive” version and a “no-difference” version), purposefully placing minor errors in each. They randomly sent one version or the other to over 200 peer reviewers. Their findings? Not only was the positive version more likely to be recommended for publication, but reviewers were less likely to detect the errors in the positive version, and more likely to score the methods section of the positive version higher (despite the fact that the methods sections were identical).
So listen up, authors, reviewers and journal editors: stop accentuating the positive and eliminating the negative! Submit those high-quality negative studies, review them fairly, and get them out there for all to see.
One thing Bing Crosby and I can agree on, though...you still shouldn’t mess with Mr. In-Between.
Tidak ada komentar:
Posting Komentar