Checking whether a paper you cite has been retracted
A retracted paper is marked as retracted in some of the places a reader meets it and not in others, and a writer who cites a paper can search the record to find out whether it has been retracted.
Where a retraction is marked
In 2014 a retraction was easy to miss in PubMed, which used no watermarks, bold or pallid, to show that an article had been retracted. A PubMed search for one retracted author's articles in one journal returned an unremarkable list, in which the links to the retraction notices were easy to miss for a reader not looking for them. Selecting one of the articles gave a link to the retraction notice just above the abstract, which a reader would probably not miss, but a brighter mark would probably be better, and highlighting the retraction in red might work.
The publisher's website marked the same article more clearly, and a reader could not miss it. It prefaced the article's title with "THIS ARTICLE HAS BEEN RETRACTED", and its PDF link led to a copy of the article bearing a transparent watermark.
In a case reported in 2016, a clinical chemistry journal retracted a paper it had accepted on two fraudulent reviews. Its retraction notice is freely available, and the retracted article can still be accessed but carries an unmissable translucent "RETRACTED" watermark.
In 2012 a retraction or correction was noted on PubMed and on the publisher's website, but the first articles a Google search returned might not show it. CrossMark, then a new initiative from CrossRef, was meant to communicate information about an article to its readers, and to be particularly useful for corrections and retractions. Any document with a digital object identifier (DOI), which is often assigned on acceptance of a manuscript, could carry a CrossMark, including online early articles, PDFs, HTML versions and abstracts. Clicking the CrossMark logo opened a box giving the status of the article, with details such as the publisher, the publication date and the DOI, and the status also showed updates such as a correction. Another tab showed metadata such as whether the article was peer reviewed, its publication history, its copyright holder and its funding disclosures. In 2012 its pilot could be viewed online, and CrossMark was expected to make retracted articles easier to identify.
Finding out whether a cited paper has been retracted
In 2012 MEDLINE and the Web of Science could be searched for retractions, but no single database of scientific article retractions existed.
A writer should not cite a retracted article. Retracted articles can be found in PubMed with an advanced search that selects "retracted publication" as the publication type and combines it with a general keyword related to the manuscript, to check that none of the manuscript's cited articles appears in the results. A writer should also look on the publisher's website for any retraction notice.
In 2012 the best way to keep abreast of retractions was to visit the Retraction Watch blog or sign up for its alerts. Adam Marcus and Ivan Oransky set it up in August 2010. The blog reports new retractions and the fate of the authors whose articles are retracted, and it discusses general issues around retraction. It can be useful reading for a writer faced with a retraction. One writer describes it as dedicated to exposing researchers who break the rules to get ahead, with disturbing yet instructive cases of ghostwriting, plagiarism, data fabrication and image manipulation, and recommends its catalogue of ruined careers and damaged lives to anyone looking for a disincentive to cheat. A 2011 conference reporter called it a must-read for everyone invested in the scientific community.
In 2015 Oransky and a partner were developing a new retraction database, with support from the John Arnold Foundation and the MacArthur Foundation. They aimed to have it running by the end of 2016, which should make it easier to discover whether a specific article has been retracted.