Stewardship of Integrity in Scientific Communication

Abstract

Integrity in the pursuit of discovery through application of the scientific method and reporting the results is an obligation for each of us as scientists. We cannot let the value of science be diminished because discovering knowledge is vital to understand ourselves and our impacts on the earth. We support the value of science by our stewardship of integrity in the conduct, training, reporting, and proposing of scientific investigation. The players who have these responsibilities are authors, reviewers, editors, and readers. Each role has to be played with vigilance for ethical behavior, including compliance with regulations for protections of study subjects, use of select agents and biohazards, regulations of use of stem cells, resource sharing, posting datasets to public repositories, and so on. The positive take-home message is that the scientific community is taking Steps in behavior to protect the integrity of science.

INTRODUCTION

Integrity in the pursuit of discovery through application of the scientific method is an obligation for each of us as scientists because, without integrity, the value of science is diminished. We cannot let the value of science be diminished because of the importance of discovering knowledge to understanding ourselves, the world in which we live, and our impacts on the earth and its water, both salt and freshwater, and the air above us. Therefore, one of our responsibilities as scientists is stewardship of integrity in the conduct, training, reporting, and proposing of scientific investigation. For the purpose of the session “An afternoon with journal editors: an open discussion with senior journal editors,” this article considers the coin of the realm (Smith, 1816) in science: peer-reviewed publications. More specifically, this article focuses on integrity in scientific communication through peer-review.

Peer-reviewed publications are the academic coin of the realm because peer-review validates scholarship. Without peer-review, studies and their results are not part of the scientific record. By convention, the scientific record (publication) is available and permanent, regardless of the type of scholarship, be it discovery, integration, application, or teaching (Boyer, 1990). To paraphrase, without peer-reviewed publications, scientific activity is invisible (Glassick, 2000).

Why do we publish? Publication provides opportunity to tell our scholarship story (Figure 1). Why tell a story? In a simplistic view, we tell our story for two principal reasons: (1) to advance our field and (2) to advance our career. One’s field is advanced through report of the solution of an impactful discovery, integration, application, or teaching problem. That is the altruistic goal of publishing one’s science. But, a secondary reason is to advance one’s career, including products of intellectual property (Olds, 2004). Sometimes, however, the secondary reason may collide or compete with the primary reason. Why? My experience is that bias, or ambition, and/or pressure to meet numeric expectations for faculty advancement, award of tenure, pay increase, allocation of laboratory space, development of intellectual property, honors and awards, or administrative advancement are examples that may contribute to collision or competition between advancing one’s career versus one’s field. Another of my experiences is that the digital era ushered in a relatively easy mechanism to add or remove pixels, or rotate or invert digital images or pieces of them, or otherwise change attributes of digital images. Any of these actions may compromise the integrity of science through scientific misconduct. Charles Babbage (1791–1871), a British mathematician, coined scientific misconduct synonyms of “cooking” and “trimming” data to agree with a researcher’s hypothesis (Babbage, 1830; Miller and Hersen, 1992). To paraphrase Babbage’s writings, such a researcher (“cook”) might enjoy short-term gain (“procure temporary reputation”) for long-term pain (“at the expense of his [“/her”; my addition] permanent fame”), to which I further add erosion of the integrity of science. Avoiding scientific misconduct is through ethical conduct, which is the topic of this article.

Fig. 1.

The attention-grabbing storyline of this original painting for the cover illustration of The Anatomical Record volume 292(9), 2009 is that a lowly triceratops gleefully prances into the sunset after dethroning a Tyrannosaurus rex.

ETHICS

The word ethics is derived from the Greek word ethos, which translated to English is “character.” In the scientific enterprise, ethics is an agreed upon set of rules and standards governing the conduct (character) of us, you and me, as scientists. Journals published in the United States are required to follow the Federal Government’s (Office of Research Integrity) definition of scientific misconduct. The definition is behavior that includes “fabrication, falsification, or plagiarism in proposing, performing, or reviewing research or in reporting research” (US Department of Health & Human Services, 2017).

A point raised by two contemporary papers is that science, in the broad context, has become hypercompetitive, especially in the review of manuscripts and grant applications (McDowell et al., 2014; Montgomery et al., 2014). Troubling descriptors provided by scientists who responded to the surveys included “harsh, tough, brutal business.” Also troubling is the finding that 58% of 970 survey respondents were either aware of scientists feeling tempted or pressured to compromise integrity for getting supportive results (Montgomery et al., 2014). These revelations, through surveys that may have inherent weaknesses, are worrisome because compromising integrity in the design, conduct, and/or reporting of research erodes the ethical foundation of science and the public’s confidence in science and scientists.

Players in the Peer-Review Process

From my perspective, four protagonists are involved, in multiple, and potentially conflicting, capacities. The protagonists are authors, reviewers, editors, and readers (Figure 2). These roles create potential for conflict of interest because a scientist may be many or all of the protagonists at the same time. Therefore, each role has to be played with vigilance for ethical behavior to protect the integrity of science.

Fig. 2.

Protagonists in publishing. For publishing to advance a field, all protagonists need to act ethically at all times. Photographs courtesy of Gary C. Schoenwolf.

Authors have the right to publish their scholarship and to develop their scientific story. Expectations also are important. Authors expect fair review in the peer-review process. Authors’ responsibilities are to be honest in scholarship and writing, place their story in context of the field, to write without bias, and to consider limitations of their study. As Editor-in-Chief of a scientific journal, I sometimes field statements from authors that the reviewers are harsh, or are not qualified to review the author’s field of research, or that the author knows who the bad reviewer is and that reviewer is the author’s enemy. Such statements undermine the purpose of peer-review, which is to improve the quality and clarity of a submission, as well as transparency of reporting methods and results. I suggest to such authors that they put aside the reviews and their submission for a week or two to distance themselves from personalization of the reviews. My experience is that returning to a critique after a time-out period reveals that many of the concerns raised by the reviewers are attributable to unclear, imprecise, biased, and/or contradictory writing. The player whose responsibility is to fix these concerns is the author. Authors also have the right to disagree with reviewers’ concerns. However, the disagreement should be focused at the level of the science. A respectfully articulated disagreement, backed up by evidence, may become persuasive to the particular reviewer(s) and editor.

Reviewers use their experience and expertise, which is why they are invited to review. Reviewers are asked to express their opinion about the science, its impact on the field, and quality of composition, including English grammar and punctuation. A point that I make to inexperienced reviewers is they should focus their criticism on the science, not the scientist, just as I urge authors. Further advice that I provide to reviewers is to focus their critique on the question that drove the study, adequacy of the methods and potential limitations, why the answer is important, and consider how the answer advances the field. Expectations of reviewers are that they will avoid bias or conflict of interest, review in a timely period, and clarify or improve scholarship. By so doing, reviewers fulfil an important contribution to peer review: to transform that which is not clear to clearer. Sometimes, reviewers request additional experiments. But such requests may raise concern by authors that a reviewer is delaying acceptance. A policy that The Anatomical Record applies is that recommendations for new experiments require justification by the reviewer, which is the Journal’s effort to provide check-and-balance. The flip-side of what sometimes arrives at an editors’ desk is reviewers expect reciprocal fairness by authors. For example, reviewers may identify argumentative and/or unresponsive authors, based on the authors’ detailed response letter to a review and/or lack of adjustment of the revised submission. Argumentative or unresponsive authors chip away at the intent and spirit of reciprocal fairness in the peer-review process. A consequence of an argumentative or unresponsive author may be dampened enthusiasm by reviewers, which editors are obliged to consider for decision. My approach is to encourage collegial dialogue between reviewers and authors regarding the science, its story, and its impact on a field. In a way, this feedback is a type of reward to reviewers for volunteering their intellect and time to improve clarity and impact of scientific communication. Readers are referred to an interesting consideration of this voluntary service (Ioannidis, 2014).

Reviewers also have the opportunity to mentor graduate students, postdoctoral fellows, clinical fellows, and early-stage investigators in the ethics, art, and communication of peer-review. This is an important opportunity for trainees and early-stage investigators to get feedback and strategy for participating fairly in the peer-review process. When this mentoring activity happens, the reviewer/mentor should inform the editor.

Editors’ efforts include recruitment of the best scholarship, overseeing published scholarship that is consistent with the mission of the journal, and creatively strengthening the journal’s esteem. Editors are expected to be vigilant stewards and guardians for the journal, fair peer review, and scientific integrity. Editors also are responsible to oversee timely review of every submission, which from time to time requires balancing differing opinions. Balancing conflicting recommendations by reviewers requires equipoise when one reviewer recommends acceptance, another reviewer recommends rejection, and a third reviewer recommends revision, be it major or minor. I find that sharing the rationale for the balanced decision with the reviewers keeps them engaged, without a sense that their recommendation is ignored.

Readers critically evaluate publications. Readers are free to draw their own conclusions. Readers also use results of published papers to drive their scholarship. These rights also carry responsibilities. Among the responsibilities for readers are to correctly represent scholarship by others and to give credit for scholarship by others. Thoughtful criticism and balanced incorporation into a reader’s manuscript, grant application, etc., is what authors of the reported science expect.

When these protagonists participate for the purpose of having a clear, impactful paper published as the product of peer-review, the process of peer review is successful.

Compliance

Journals require explicit statement in the Methods section that protection of human subjects or vertebrate animals for research was prospectively approved by the appropriate institutional oversight committees. For compliance for protection of human subjects, Title 45 U.S. Code of Federal Regulations, Part 46 applies (Title 45 U.S. Code of Federal Regulations, 2009), which stems in part from atrocities committed by Nazi Germany and Japan during World War II. Other important compliance documents are the WMA Declaration of Helsinki – Ethical Principles for Medical Research Involving Human Subjects (The World Medical Association, 2017), the Belmont Report (The Belmont Report, 1979), and the International Ethical Guidelines for Health-related Research Involving Humans report (The Council for International Organizations of Medical Sciences, 2017). Compliance is the purview of Institutional Review Boards (IRBs) in the United States or equivalent in other countries. For compliance for protection of vertebrate animals used for research, the guiding principles in the United States are provided by The National Research Council (The National Research Council, 2011), The American Physiological Society (The American Physiological Society, 2018), and The American Veterinary Medical Association (The American Veterinary Medical Association, 2013). Compliance is the domain of Institutional Animal Care and Use Committees (IACUCs) in the United States or equivalent in other countries.

Several other compliance topics are necessary to address at the time of submission. Among the topics are precautions for select agents and biohazards, compliance to federal regulations of use of stem cells, resource sharing, and posting datasets to public repositories. Authors are recommend to stay current with these regulatory topics.

Text similarity and figure manipulation.

Another guardianship role for editors is text similarity and figure manipulation. Guardianship for text similarity is related to the US Government’s definition of scientific misconduct being behavior that includes “fabrication, falsification, or plagiarism in proposing, performing, or reviewing research or in reporting research” (emphasis added because of the subject of this article). A guiding principle is that authors may not copy-and-paste published writing, whether from other authors’ published scholarship or their own published scholarship. When published work is described or quoted, citation, with full and accurate reference information, is required. Software is available to editors to check for similar text from indexed journals and books, as well as registered web pages. One such software is iThenticate (iThenticate, 2018). The vast majority of indexed journals have this software available. Each report shows the verbatim text that is copied, which is color-coded to the complete reference (source). The report may be provided to authors to revise the similar text in original words. One of the challenges using such software is description of methods because word choices should be precise for technical terms, reagent and equipment names, animal species names, protocol steps, and so on. Thus, methods descriptions will be similar from paper to paper by the same research group. This is unavoidable. Consequently, The Anatomical Record excludes the methods section from evaluation for similar text, as well as the Title Page, Literature Cited, and Acknowledgement. The sections of a manuscript that are assessed are the Abstract, Introduction, Results, Discussion, and Figure Legends.

Figures also are a potential source for manipulation. Briefly, the concerns are figure and pixel manipulation. Approaches for discovery that are used by journals range from required submission of original digital files against which the figures in the manuscript are compared, to using software applications that evaluate pixels. An interesting paper on this topic is available (Newman, 2013).

Committee on publication ethics (COPE).

Assistance to editors to navigate the day-to-day responsibilities of journal and science guardianship is available through the Committee On Publication Ethics (Committee on Publication Ethics (COPE), 2018). This is an online forum for discussion about ethical issues in the conduct and reporting of scientific studies. COPE provides a worldwide platform for discussions among editors. Common examples of problem topics are conflict of interest, authorship, compliance, and fabrication or falsification of data and images. The forum helps editors identify such problems, develop approaches to gather information, and steps to take to provide the information to a university or other research entity, which is responsible for investigation and action. Editors may use the outcome of investigations and actions by universities or other research entities to retract a published paper or have it withdrawn by the authors.

Is scientific misconduct worse today than in the past? This question is difficult to answer (Kakuk, 2009). Evidence suggests that awareness of scientific misconduct and inquiry about originality of writing, authenticity of data, or duplication of publication is more acute today than historically (Cokol et al., 2008; Steen, 2011; Van Noorden, 2011; Fang et al., 2012; Steen et al., 2013; Fanelli et al., 2015). A behavioral shift is theorized to reflect recognition and effort by the scientific community to correct or remove flawed published papers. For example, corrections happen in shorter time (shorter time-to-retraction) today and include adjustments for writing offenses, such as plagiarism. At the same time, the number of publications overall is increasing, along with increasing number of journals, especially predatory journals. Therefore, the denominator continues to increase, thus complicating interpretation of assessments. Another confounder is weakness in study design. Studies to date are mainly surveys of stakeholders in the scientific community (Ranstam et al., 2000; Martinson et al., 2005; Titus et al., 2008; Baerlocher et al., 2010; John et al., 2012; McDowell et al., 2014; Montgomery et al., 2014) or retrospective inquiries of databases (Fanelli, 2009; Steen et al., 2013; Fanelli et al., 2015), both designs of which have limitations, as identified in these papers. Nonetheless, the positive take-home message is that the scientific community is taking steps in behavior to protect the integrity of science.

My recommendation is simple: behave ethically.

SESSION DISCUSSION

Bruce Alberts (session moderator): I think we will take a few short questions after each because there are specific issues we do not want to forget. I want to take the first question. This issue of plagiarism to me, it can get in the way, because when I was a scientist I put all the methods in over and over in every paper at the back in the methods section. Now my sense is people do not feel that it is ok to use the same methods, repeat them. And so instead what I find in journals, it is the same as the methods we used in these papers, they give us some citations. I teach from these papers, so I with the students, we try to figure out actually what has been done. It is almost impossible often, and I wonder to what extent the idea of reproducing your methods section, which will show up in this plagiarism, is that a problem?

Kurt Albertine: How do we at The Anatomical Record manage that mechanically? Our editorial board established a policy that because methods can be described in limited ways, we exclude the Methods section from the iThenticate evaluation. We evaluate the Abstract, Introduction, Results, and Discussion sections, as well as the Figure Legends. We also exclude the Title page and Literature Cited section. Our feedback regarding plagiarism includes guidance for adjustment and we work with authors to assist them.

Bruce Alberts: Is this discussed at COPE (Committee on Publication Ethics)? I mean this general issue?

Kurt Albertine: It is discussed within COPE.

Bruce Alberts: Is it a general feeling?

Kurt Albertine: I do not know if it is a general feeling but the journal editors with whom I talk are in agreement. I think, from a philosophical standpoint, that the pressure on journals today to have more papers that are shorter in an issue to increase impact factor has come at a huge cost of reproducibility of science. If a reviewer or reader cannot find informative description of methods, then reviewers and readers have to use their time and effort to discover the methods, which sometimes requires stepping down a list of references to find the informative reference, if found at all. This harms reproducibility of science.

Question 1: I wanted to develop that question on plagiarism a little bit. So you said either rewrite or withdraw the paper, and I guess the same will apply for your image checking when you find things with images. One problem of course in the rare case where you do have fraud and that is a vast minority of cases, of course one has to be careful that the journal does not aid and abet basically the person just faking it better. At the journal that I am affiliated with, we have tried to develop a multi-step process to be quite sensible and not to say just mutate the text a little bit more so people or iThenticate do not detect plagiarism anymore. And as a whole, there is a whole infrastructure of companies that do these services for people. My question is what do journals do? So are we the ones that have to investigate the intent of the plagiarism or what the image manipulations are?

Kurt Albertine: One of the responsibilities of journal editors is to safeguard against plagiarism. However, editors are not responsible for determining intent.

Question 2: Ok, which is difficult of course?

Kurt Albertine: Yes, safeguarding is difficult; it takes time. But that is the expectation placed upon an editor if a call of authenticity of data or conduct is raised. An investigation to gather facts is warranted. In this regard, editors become gatherers: they gather information and present it to the relevant parties at an institution.

Question 3: And would you then say this is done independently of the institutions?

Kurt Albertine: That is how the process starts, at least in my experience, because a concern is brought to my attention about a paper published in the Journal or being peer-reviewed by the Journal. My responsibility is to gather information, the pros and cons, the original data, and present them to the home institution at which the concern arose. I have talked with college deans and university presidents at institutions where scientific misconduct has been alleged. But my role as an editor is not to determine intent. I follow the guidance within COPE. I assemble a report that leads to telephone conversations with the involved parties and, if requested, the report is sent to appropriate individuals at the home institution. This is not a comfortable process but it is an editor’s responsibility to the journal, the involved parties, and the scientific community. With respect to concerns about digital images, we request the original tiff files to compare with the submitted digital figures.

Bruce Alberts: Does anybody know of an effective course for teaching young scientists how to write scientific papers? If so, how long does it take? Does it work? I noticed that Science Magazine, I would estimate about half of the accepted papers, the abstracts I could not understand even in Biology, because they had so many three letter words and assumed vocabulary. One of the problems is once you understand something deeply it is very hard, I guess, to realize that people do not have the same background of knowledge that you have, but that is a very important part of clear writing.

Kurt Albertine: I recommend one book that I use. It is by Mimi Zeiger from the University of California, San Francisco (retired; I have no financial interest in the book). The title is “Essentials in Writing Biomedical Research Papers” (Second edition, McGraw-Hill, Inc.2000; ISBN 0–07-134544–2). The book teaches by example. You do not read it from cover to cover. Instead, you locate the section of a paper that is causing puzzlement to find original writing and rewritten writing, with wide margins that are used to explain why the original was unclear and how the rewritten version is clearer.