CSE 2012 Survey on Research Misconduct Allegations

Five years ago, the Council of Science Editors conducted a survey regarding editors’ responses to allegations of research misconduct in manuscripts and publications. The survey, repeated in 2012, was divided among issues involving manuscripts and those involving publications. The purposes of the surveys have been to prompt editors to consider such issues before an incident occurs, to offer benchmarking for editors, and to provide a vehicle by which editors can share their experiences in handling such matters. Unfortunately, since the initial 2007 survey, more editors have had experience with allegations of misconduct. Although many of the editorial responses remained consistent, a few reflected editors’ greater awareness of the tools for addressing problematic articles or manuscripts or the refinement of survey options that allowed more subtle answers. For example, relative to the 2007 survey, a much larger number of editors indicated that they would use an expression of concern if an institutional official indicated that an article should be withdrawn but authors did not (8%), if an author submitted a revised figure after admitting that a published figure was incorrect (8%), if an author expressed concern about the integrity of data (27%), or if an institution made a misconduct finding but the relevant government agency did not (31%).

Fewer editors would report an incident to an institution for investigation, preferring to resolve issues by communicating with authors. For example, in the 2007 survey, if a researcher wrote to an editor and asserted that he or she should have been included as an author, 32% of editors indicated that they would write to the institution to investigate the claim. In the 2012 survey, only 11% of editors would use that approach; instead, most editors would write to the corresponding author (48%) or all the coauthors (34%) to ask whether the researcher should be included. However, if an author indicated that he or she wanted to be removed as an author of a published paper because of integrity concerns, many editors would immediately contact the institution for investigation (16%), and some would ask the authors to address those concerns and then provide the information to the institution for investigation (23%). Nonetheless, a large fraction of editors would simply work with the authors on the issues and not involve the institution (34%).

Like the 2007 survey, the 2012 survey indicated that few editors would report an incident to government authorities. Only 12% indicated that they would contact a government agency if they were aware that a prior allegation of misconduct had been made against an author. The lack of involvement of a government agency is not surprising in countries in which the relevant government body has little experience. However, the failure to involve federal agencies in the United States is somewhat surprising in that those agencies would have reports of prior findings of misconduct that editors could consider in assessing a new allegation.

In the 2007 survey, 69% of editors indicated that they would retract a paper if the authors indicated that they could not locate the primary data and wanted to retract the paper. In the 2012 survey, only 25% of editors would retract an article on such a basis. Rather, 25% of editors indicated that they would retract only if the article were less than 6 years old, 24% would retract only if a concern had been raised about data fabrication or falsification, and 19% would publish the letter from the authors but not retract the article.

About 30% of editors indicated that they would ban a researcher from submitting articles if the researcher’s institution found the researcher guilty of misconduct, and an additional 30% indicated that they would ban the researcher if the institution found misconduct and a pattern of misconduct. Some 16% of editors would not ban a researcher found guilty of misconduct, 8% would ban a researcher only if he or she refused to correct the literature, and 18% would ban a researcher only if a government agency found him or her guilty.

Although 41% of editors would not impose a sanction against an author of a multi-author paper if another author were found guilty of misconduct, 14% would sanction the author if he or she did not help with correction of the literature, 28% if the institution found the author remiss in his or her obligations as a coauthor, and 16% if the researcher did not advise the journal as soon as possible that the published research was flawed.

During the session about the survey at the 2012 CSE annual meeting in Seattle, a number of questions were posed comparing the survey responses with the approaches suggested by the Committee on Publication Ethics (COPE) flowcharts. Some characteristics of the environment in the United States contributed to approaches different from those recommended by COPE, for example, the existence of institutional and government infrastructure and the proclivity toward litigation threats. Accordingly, questions were posed about the best options to mitigate against the risk of being sued. Editors are encouraged to review the complete survey results ( and to contribute to this growing body of collective knowledge.

DEBRA M PARRISH is a partner with Parrish Law Offices, Pittsburgh, Pennsylvania.