PHENOMENA OF RETRACTION:
REASONS FOR RETRACTION AND
CITATIONS TO THE PUBLICATIONS
John M Budd and Mary Ellen Sievert
School of Information Science and Learning Technologies. University of Missouri-Columbia, 104 Stewart Hall, Columbia, MO 65211, USA
Objective: The study's purpose was to examine a set of publications identified as being retracted in the biomedical literature to ascertain why the publications were retracted (and by whom) and to what extent they continue to be incorporated in subsequent work through citation. Previous work on the latter has been inconclusive. We hypothesized that the citations to retracted publications continue and include positive citations.
Design: A search of MEDLINE identified a set of 214 retractions. The retraction notice was examined to determine the reasons for retraction, such as unavoidable error and misconduct. The retracted publications were classified according to the reasons for retraction and persons retracting. Following this classification, a sample of available citations to each retracted article was examined so as to determine how the article was cited (whether the problem leading to the retraction was noted, the article was simply mentioned in a literature review, or the article was used as substantive background in the citing paper). A chi-square test indicates whether the number of positive citations exceeds that which might otherwise be expected.
Results: Reasons for retraction were somewhat varied. In 80 papers some error was acknowledged; in 37 results could not be replicated; in 73 misconduct was evident; and in 24 no clear reason was given. A total of 173 papers were retracted by some or all of the authors; 41 were retracted by a person or organization other than the author(s). The 214 retracted articles were cited 1,765 times after the retraction notice.
Conclusions: It is evident from these data that there are serious problems with the majority of articles in biomedical science that are retracted. Error or misconduct affect the outcome of the published work. Even after retraction the articles continue to be cited in the literature. The majority of the citing works give no indication of awareness of the retractions and many of those citations indicate substantive use of the retracted work. The extent of continued citation of retracted articles is a cause of concern regarding the conduct of biomedical research.
Return to Session Information
