Themes

Retractions: correcting the scientific literature

What is this about?

Retraction is the process of withdrawal from publication of those articles that display seriously flawed or erroneous data. Retraction aims to correct the scholarly literature and alert readers of an article’s serious mistakes (1). The flawed data can be the result of honest error or from research misconduct. When unnoticed, retracted papers are still seen as valid and decrease the trustworthiness of science that follows.

Why is this important?

The integrity of the scientific record is important because published research serves as a basis for new research or application in practice. If the published report on research results is not correct, it may waste future research effort and, what is more dangerous, have direct adverse effects on the public. This is particularly relevant for health research because incorrect health research results may cause harm to patients or the general population. After publication, when it becomes apparent that the results and/or interpretations of an article are seriously flawed, an article can be retracted. Retraction differs from correction, where an article is corrected after publication. Retraction is more serious as a retracted paper should no longer be considered as a source of scientific knowledge. It is also a signal to alert other scholars of the errors. Main reasons for retraction are honest research errors, plagiarism, redundant publication, fabrication, falsification, experimental artefacts and unexplained irreproducibility (2,3). The COPE guidelines state that retractions are not to punish the authors. In addition, authors of a paper, as well as others, can call for a retraction upon discovering errors. COPE guidelines state that retracted papers should be labelled as retracted and be accessible (both offline and online). Most journals have their own retraction guidelines. Over the years an increase in the percentage of retracted papers is observed (3). The two possible explanations for this are 1) an increase in pressure to publish flawed papers or 2) an increase in detection of such flaws (4). Nonetheless, retractions have an impact on the scientific community. First, it is a waste of resources, both in financial terms, time and participants. Second, when unnoticed, authors implicitly or explicitly use retracted sources as valid scientific results leading to decreasing trustworthiness of science (5).

For whom is this important?

Journal publishers, Journal editors, Senior researchers

What are the best practices?

PubMed

For example, the U.S. National Library of Medicine, which produces PubMed, the largest bibliographical database in health research, describes different types of procedures for amendments to the published scientific record. Retraction of an article involves publishing a retraction notice that explains the reasons for retraction and who is making the retraction. This notice links to the retracted article, which is clearly marked in the indexing database.

The current list of all retractions in PubMed, regardless of the cause (error or misconduct) is available here.

Some journals want to differentiate between retractions due to misconduct and those that are due to an error that makes a research conclusion wrong, but can be corrected. Such corrections have been termed “retraction with republication” or “retraction with replacement” by the International Committee of Medical Journal Editors (ICMJE). This type of retraction is used when a serious error makes published results unreliable, but it was judged that the error was not intentional and the corrections are possible. Such retraction and replacement should be accompanied by full explanation, including a clear presentation of the extent of changes that were made.

The problem may occur when the retracted and replaced articles keep the same pagination and bibliographic identifiers (such as DOI – digital object identifier). Some bibliographical databases may not recognize this as a proper correction of literature. Databases usually require that the original and corrected/replaced publication are kept as separated publication items, with added notice that links them. Different understanding of how a retracted publication can be replaced with a corrected publication has caused differences in indexing of such publications in bibliographical databases, creating confusion for the users of published research articles.

A relevant tool to detect retracted papers and keep up to date on retraction scandals is Retraction Watch. Retraction watch is both a blog and a database of an estimated 17,000 retracted papers. The blog regularly updates on papers and authors that are retracted/about to be retracted. The database, found here (insert hyperlink http://retractiondatabase.org/), is a tool to find out whether a paper has been retracted. Journals, authors and reasons for retractions are given in the search results.

References

1. COPE. Committee on Publication Ethics (COPE): Guidelines for retracting articles. 2010.

2. Kleinert S. COPE's retraction guidelines. The Lancet 2009;374(9705):1876-1877.

3. Wager E, Barbour V, Yentis S, Kleinert S. Retractions: guidance from the Committee on Publication Ethics (COPE). Maturitas 2009;64(4):201-203.

4. Cokol M, Ozbay F, Rodriguez‐Esteban R. Retraction rates are on the rise. EMBO Rep 2008;9(1):2.

5. Budd JM, Sievert M, Schultz TR. Phenomena of retraction: reasons for retraction and citations to the publications. JAMA 1998;280(3):296-297.

Ana Marušić, Iris Lechner contributed to this theme.

Latest contribution was June 1, 2019