On Second Thought

Science magazine's Retraction Watch blog lists the top 10 scientific study and paper retractions of 2015.

Retractions can happen for a number of honest reasons, such as the discovery of a previously overlooked error in research. But it can also happen for dishonest reasons, and some of the papers on the list committed one of those sins. The reason such actions earn retractions -- essentially decisions on the part of the author that the papers no longer officially exist -- is that deliberate fudging of data or the creation of false peer reviewers means that the conclusions of the paper may have been influenced by something other than the plain facts of whatever study or experiment is its subject. Other scientists in the field might be able to run the same experiments in order to verify the result, but many of the rest of us can't. If the study becomes the basis for public policy, health recommendations, treatment protocols and the like, the people who use it have flawed information that may wind up producing the exact opposite of the desired result. A second group of researchers were unable to get the same results as the original experimenters in item no. 1, for example, which led to the discovery of falsified survey data.

My favorite is no. 5 -- a paper on detecting plagiarism had to be pulled because the author plagiarized part of the paper, and the folks at Retraction Watch found the notice of retraction on April Fool's Day.

The frequency of retractions and the difficulty in creating reproducible results is one reason that Real Clear Science writer Ross Pomeroy at the site's Newton Blog has decided to ignore nutrition studies. Pomeroy doesn't mention the Retraction Watch article directly, but that's a whole other kettle of fish. Battered and deep fried, as a matter of fact.