I wrote a while ago about drug companies as whistle blowers. Evidently, the pharmaceutical industry wastes more and more resources unsuccessfully trying to replicate published research studies.
The amount of irreproducible published research surprised me. If there is such a trend, questions accumulate. Are researchers becoming increasingly careless, or even fraudulent? Are researchers acting under too heavy pressure to publish positive results? Do many researchers lack sufficient skills in statistics?
Or has research in the life sciences entered such complex terrain that it has become virtually inhuman to survey all factors that may influence the results?
I’m not competent to answer these questions and welcome helpful comments.
A way to handle at least part of the problem has been suggested: set up a centralized retractions database. Such a resource would help scientists and the industry to exclude at least some of the most unsuitable candidates for replication.
An ambitious study of retractions used secondary sources when the journals’ own reasons for the retractions seemed incomplete or vague. According to this study, fraud or suspected fraud caused 43% of the retractions. Duplication accounted for 14% of the retractions, and plagiarism for 10%; only 21% of the retractions had to do with error.
If you want to read more about the study on retractions, it is summarized in this recent Nature News article.