You are honored to receive an appointment to your alma mater’s search committee to identify another full-time faculty member. Clinical research has continued to intrigue you since completion of your residency, and as serendipity has it, the short list includes a renowned candidate whose research interests closely parallel yours. In fact, you notice a striking similarity between the candidate’s most recently published article and a study that you had reviewed several years ago during a related literature search. Your concern is that the candidate and the author of the original article are 2 different persons. The experimental design, sample size, results, and conclusions are identical to those of the first publication. Even the subjects’ age distribution is remarkably similar. You begin to become suspicious that the 2 articles are inappropriately related.
The value of publication in the academic arena is a long-standing truth. “Publish or perish” is the classical admonition that summarizes one of the most conspicuous indicators of scientific productivity. Furthermore, retraction—a journal editor’s indication that a previously published article has serious flaws—is an indication that an investigation has been found it to be deficient. Retraction is therefore a necessary prerogative in the pursuit of authentic, evidence-based knowledge. The reasons for retraction vary, spanning errors, duplicate publication (publishing the same article in more than 1 journal), plagiarism, and fraud. Fraud includes fabrication of data or data falsification.
The alarming aspect of article retraction is that the frequency of retraction for plagiarism or fraud has increased significantly in recent years. In a review of 2047 retracted articles listed in PubMed, only 21.3% were retracted because of errors. Fraud or suspected fraud (43.4%), duplicate publication (14%), and plagiarism (10%) comprised the other reasons for retraction. A few other reasons for retraction completed the statistics. The frequency of retraction because of fraud or suspected fraud per total scientific publications was 10 times greater by 2012 than in 1975. Authors from the United States, Germany, Japan, and China were responsible for 75% of the retractions for fraud or suspected fraud. American involvement is not surprising because one fourth of all scientific publications originate in the United States. China and India combined to exceed more duplicate publications and plagiarized articles than the United States.
It is interesting that the frequency of retraction is highly correlated with a journal’s impact factor. The impact factor is a numeric figure that indicates the frequency with which a journal’s articles have been cited in subsequent publications within a 2-year period. The association between the impact factor and retraction frequency has been attributed to the presumably keen scrutiny of reviewers for these journals and their tendency to publish studies involving topics that address emerging concepts in their field.
What can be done to control the frequency of retraction for reasons of unethical behavior? Vigilant reviewer scrutiny and technology capable of identifying duplicate publications are possibilities. Increased education of researchers about bioethics might be helpful, but ethicists have long debated whether increased awareness of ethics leads to improved behavior. Ethics education, now mandatory in graduate research programs, is intended to increase an investigator’s understanding of the pressures that incite unethical behavior. Additional awareness of ethics may also encourage senior research personnel to mentor others in honest research and publication. The reason for retraction—either error or fraud—should be routinely publicized despite the potential impact on an investigator’s prestige and credibility. There is a Web site entitled “retractionwatch.com” that monitors and reports retraction events.
Should you advise the other search committee members of your suspicion of fraud regarding this applicant’s history? In any ethical decision, clarification of the facts is the first step to resolving a dilemma, and this case is no exception. You might find that the evidence is so clear that there is no dilemma at all.