Back in 1995, I thought it would be a fairly simple job to draft a set of guidelines for the responsible communication of scientific results. The Society for Neuroscience (SfN) wanted such guidelines, and I chaired the committee that drafted them. At first, we assumed that a small group of people would easily prepare a simple document of a few pages. The reality proved to be quite different: The process spanned more than three years, during which our text evolved from a few paragraphs written by two people to about 5,000 words written by a committee of seven, and finally to 13,500 words composed by 13 individuals with experience in academia, industry, publishing and law.

In the end, our guidelines didn’t just give instructions; they explained the reasons that researchers should behave in certain ways. And in generating those explanations, we sometimes decided to modify the very rules we had set out to justify. The SfN continues to refine its guidelines today, and the process has stimulated much reflection on the role of professional societies in establishing ethical guidelines—and keeping them alive, relevant and effective.

Getting Specific

Professional societies have a long and honorable history, tracing back at least to the 15th century. But only very recently have scientific societies begun to establish guidelines on research ethics for their members. Today, more than 50 societies have written guidelines, and this is as it should be: Professional organizations are in a unique position to promote the responsible conduct of research.

Responsibility for overseeing research ethics has typically fallen to research institutions rather than societies. But institutions can set standards only for the most basic and universal matters, such as plagiarism and fabrication of data. Other aspects of research, including authorship, data management and the sharing of reagents, can be too specific to a given field to be regulated at the institutional level.

True, professional societies are generally ill-equipped to investigate claims of misconduct. They also have limited powers of enforcement and few penalties to impose on those who misbehave. But societies are often in the best position to understand and set standards of conduct for the specific segment of science that they represent. Moreover, although there are some 2,500 colleges and universities in the United States alone, there are far fewer scientific societies—each of which can efficiently communicate with a large number of scientists. Finally, societies are increasingly involved in publishing research journals and organizing conferences at which scientists present results. In such venues, these organizations have a unique opportunity—and an obligation—to educate their members about responsible conduct.

Integrity in Neuroscience

My own work developing guidelines for professional conduct is almost solely with the SfN. But my experiences may illustrate how other societies could begin such an endeavor, what they might do once guidelines are adopted, and what problems they might encounter. The SfN example also illustrates how a society can go beyond guidelines alone to promote research integrity in other dynamic ways.

The SfN is a relatively young society. The word neuroscience itself did not even appear in the literature until the 1960s, and the SfN was created in 1969. Since then, both the field and the society have expanded rapidly. Indeed, during the past five years, a quarter of a million research articles on aspects of the nervous system were published. That amounts to more than 100 papers every day! And the size of SfN membership has expanded dramatically since its inception, growing from a few hundred to more than 40,000 members.

Shortly after its inception, the SfN began to take on ethical issues. In the early 1980s, public concern arose over the treatment of monkeys at the Institute for Behavioral Research in Silver Spring, Maryland. In response, the SfN organized a symposium on the use of animals in research, established a standing committee on the treatment of laboratory animals and human subjects, and developed a formal policy on those issues. The society also established a general policy on research ethics and an initial statement about fabrication, falsification and plagiarism, which was replaced in 1999 with a comprehensive set of guidelines for responsible conduct in scientific communication.

Also in the early 1980s, the SfN began to offer a Social Issues Roundtable at the society’s annual meetings. Sessions have included “Neuroscience in Developed and Developing Countries: Partnership or Exploitation,” “Perspectives on Gender in Neuroscience” and an overview of several other ethical issues. The series continues today and adapts to encompass timely issues. Shortly after the destruction of the World Trade Center towers, for example, the program committee abandoned the intended topic in favor of a special session on the impact and treatment of trauma.

As another addition to the SfN meetings, my colleague Beth Fischer and I began in 1997 to hold annual workshops on professional skills such as writing research articles and making oral presentations. We firmly believe that research ethics are best taught in the context of other skills rather than in isolation, and our workshops adhere to this principle. Skill-centered events tend to attract more students than those that address integrity alone. And where better to discuss authorship criteria and plagiarism than in a workshop on writing research articles? We use the lunch hour for case discussions led by neuroscience faculty members, thereby emphasizing that ethical concerns are worth the time of working scientists. And we establish a supportive culture by acknowledging that students and researchers do need guidance in professional skills that they might not learn in traditional training programs. Melissa Anderson, a social scientist at the University of Minnesota, and her colleagues have shown that such a culture fosters research integrity. To use the words of Addeane Caelleigh, former editor of Academic Medicine, we provide a “hidden curriculum” by presenting a “message of actions rather than formal statements.”

The twin objectives of educating and promoting a culture of responsible conduct also led the SfN to establish in 2003 the ongoing annual Neuroethics Lectures. At the most recent lecture, Henry Greely spoke of the many ethical issues raised by discoveries in neuroscience. Among the difficult questions he addressed were: If we develop cognitive enhancers, should they be available to anyone or only to those with special needs? And should we try to detect neurological diseases before symptoms appear even when there are no treatments? At the same meeting, the SfN held a special session called “The Brain on Trial,” in which neurologists testified before an actual judge about whether a fictitious individual should be convicted of murder despite evidence in his brain scans.

In 2006, growing interest in the intersection of neuroscience and ethics spawned a whole new organization, now named the International Neuroethics Society. It coordinates its meetings with those of the SfN and hosts symposia on topics such as global health, predictive biomarkers for disease, and neuroscience and national security.

Starting from Scratch

It was in the context of these expanding ethics initiatives that the SfN decided to develop guidelines on responsible conduct in communicating scientific results. The stimulus for this committee, which I chaired, was a concern among some society members that a few journals were fast tracking papers—that is, soliciting and then prioritizing certain papers rather than evaluating manuscripts in the order in which they were submitted. But the scope of our committee quickly expanded.

It is often said that one should not reinvent the wheel, and scientists often (but not always) examine what other people have written about a topic before initiating research in that area. Of course, learning too much about what has already been done can close one’s mind to new possibilities. At least that is the committee's excuse for the fact that when we started working on our guidelines, we were not hampered by knowledge! (For those embarking on the task anew, I would now recommend doing a little homework. This might include examining existing guidelines and reading “Eighteen Rules for Writing a Code of Professional Ethics” by Michael Davis, a philosopher at the Illinois Institute of Technology. Caelleigh’s “Roles for Scientific Societies in Promoting Integrity in Publication Ethics” is another excellent resource.) We did eventually make considerable use of the pioneering efforts of the American Chemical Society, which kindly gave us permission to adapt and extend its guidelines for our purposes.

We might also have begun by considering our objective: How did we want our guidelines to be used? Mark Frankel, a senior staff member of the American Association for the Advancement of Science, has been a major force in encouraging professional societies to promote research integrity. He has noted that codes of ethics come in three flavors: aspirational (what would we like everyone to do?), regulatory (what must we do?) and educational (why should we do this ?). Education researchers Felice Levine and Joyce Iutcovich have made a related distinction, noting that such guidelines generally focus on “(1) general education and professional development, (2) prevention and advisement, and (3) complaint handling and enforcement of codes of ethics.” Levine and Iutcovich add that societies also vary in the level of effort they devote to each objective.

But our committee did not begin by contemplating either set of distinctions—or how much effort we would expend. We started out developing what would have been, in retrospect, a brief set of aspirational guidelines. But over three years, we transitioned to a more regulatory and educational approach, so that our final document included extensive discussion and reasoning for each guideline. I believe that this approach made the document not only more valuable but also more honest. By forcing ourselves to justify each directive, we found instances in which we could not do so. And rather than use the “because we said so” explanation, we modified the guidelines.

For some topics, our guidelines were at odds with the codes of other organizations, and we were explicit about this in our document. A case in point involves the matter of placing pre-publication copies of manuscripts on a website. Although this practice is common in some fields, such as physics, some journals do not wish to consider a manuscript that has already appeared on the Web or been circulated widely among colleagues. They may consider such papers to have been previously published or to have lost their novelty. But our committee felt that both practices were not only acceptable but desirable because they helped promote communication among scientists. We did, however, feel obliged to warn authors that, if they followed our recommendations, they might risk having certain journals refuse their manuscripts.

Conundrums

The question of sharing manuscripts evoked little debate in our committee or in the governing council of the SfN. Some issues were even more straightforward: One cannot imagine, for example, a serious argument in favor of falsifying data. But few aspects of responsible conduct are so simple. Like most sets of guidelines, our document dealt with many subtler issues, some of which have led to a good deal of discussion. Among them were authorship, dual publication, plagiarism and sharing reagents.

Authorship: Should the head of a research group automatically be listed as an author of every publication that derives from his or her lab? Does a technician who spent many hours collecting valuable data qualify as an author? The SfN guidelines say no in both cases, stating clearly that intellectual contribution is an essential criterion for authorship. But it took a lot of discussion for our committee to reach this position, and it is still not universally accepted within our field.

Dual publication: Although we encouraged informal sharing of manuscripts, our guidelines say that publishing the same material in two primary research journals is always wrong. But what if an author works in a country where English, the language of nearly all internationally known journals, is not well understood by those who would benefit from reading the paper? In such cases, a translation could be beneficial, and some have argued that the “no dual publication” rule hampers the distribution of knowledge.

Plagiarism: The prohibition against publishing someone else’s text or data without permission is clear enough, but what about the use of someone else’s ideas? If, for instance, a 1993 publication by Jones states: “We believe that maternal stress often leads to a marked change in the endocrine response to stress in offspring,” then the use of those very words would require quotation marks and a specific reference to Jones, 1993. But what if a later author writes: “We think that stressing a mother can alter the hormonal reaction to a stressor in neonates”? Quotation marks are no longer appropriate, but surely the minor change in wording still warrants a reference to Jones. How many words must one change to absolve oneself of providing a citation? Often this is a judgment call, but it is always better to err on the side of giving Jones some credit.

Sharing research reagents: The SfN guidelines are quite emphatic about this, stating that “unique … materials used in studies being reported must be made available to qualified scientists for bona fide research purposes.” This directive is consistent with the regulations of most biomedical journals and funding agencies. But what of the student who spent years developing a reagent in order to conduct specific experiments and now wishes to reap the benefits? Surely, in evaluating what is best for science, one should also consider what is best for the careers of scientists, especially junior ones. This tension became all too clear to me at an event that I wrote about in 2003 in Science and Engineering Ethics. I quote from that report:

A year after the publication of the [SfN] guidelines I decided to help promote an awareness of the document by organizing a workshop at a small conference of neuroscientists. My focus was on the stipulation that authors be prepared to share with other investigators any materials developed in a lab and described in a peer-reviewed publication—a requirement that … had been considerably strengthened as a result of input to the guideline committee. I invited several “opinion leaders” to help with this task, including a member of the editorial board of a prominent neuroscience journal and the director of a federal agency that supports research in neuroscience. To bring a little bit of levity to an otherwise very serious discussion, party hats were provided for each of the discussants. For example, the editor was given the shade of a copy editor and the director of the funding agency received Uncle Sam’s top hat. Then I distributed an ethics case:

“Dr. Michelle Tyson is happy. She has just completed three years of hard work as a postdoc and has a great deal to show for it: She’s developed a knockout mouse (Syko) that is a model for schizophrenia, published a paper on that mouse in a prestigious journal, and has secured an assistant professorship at State University. She knows that the new mouse and the paper were critical in getting her the job and she’s ready to show the search committee that they did the right thing: She will immediately set up her new lab and then begin to reap the benefits of her hard work by exploring the neurobiology of this mouse. However, no sooner had she arrived at State University, than she received an email message from Dr. Max Megalab asking her to provide a dozen mice from her Syko colony. It is clear from the email that Dr. Megalab understands the potential of the mouse line and will have no trouble figuring out and completing the very experiments that Dr. Tyson had plans to pursue. Should Dr. Tyson provide the mouse?”

Imagine my surprise when no one on the panel felt that Dr. Tyson should provide the mouse and thereby follow the guidelines for sharing! Instead, concerns were raised about the fairness of asking a hardworking junior researcher to turn over the fruits of her labor to Dr. Megalab, the possibility that people would rather postpone publication than share a unique resource that was critical to their ongoing experiments, the absence of funds to facilitate the distribution of those materials, or the means by which the guidelines would be enforced. (For comments on this case, see [http://skillsandethics.org].) I was still trying to deal with this unexpected outcome the next morning when I returned to the location of the workshop only to discover that my own key reagents—the party hats—had apparently been stolen! I have not tried to repeat this experiment.

In a recent review of the original guidelines, the SfN reaffirmed its mandate that reagents must be shared if they are not otherwise available—a practice that reduces the waste of funds for making duplicate reagents and also promotes attempts to replicate published results. But just as in the case of honorary authorship, the regulation is not universally followed.

The Big Thaw

If some senior scientists refuse to follow an ethical guideline even in a mere case study, perhaps it is fair to ask: Do guidelines for responsible conduct matter? Do they make a difference? Unfortunately, we have little information with which to evaluate this essential question. But then, we actually have little information about the impact of any ongoing effort to reduce scientific misconduct—behaviors that are almost certainly very rare but are nonetheless of great significance to science. And we must ask, as we do of the proverbial tree falling unheard in the forest, does a set of guidelines that remains unread make a sound?

In an earlier essay in this series (March–April 2011), John Ahearne quoted the Roman poet Juvenal, who observed that “honesty is praised and then left to freeze.” And Caelleigh has written:

Anyone who works to change human behavior engages in “magical thinking” at some point—such as the irresistible hope that small, simple changes can produce large complex results in behavior… Magical thinking also underlies the situation when a scientific society passes a resolution that its members are committed to the highest standards in all aspects of research and then, based on the resolution alone, expects members to meet the standards.

As the chair of the initial SfN committee on guidelines for communicating research findings, I worried a great deal about the fate of our document. How might we avoid letting our guidelines “freeze,” sitting in the archives only to be thawed when a potential ethical breach was discovered and procedural guidelines were required? How would we avoid “magical thinking”—something that should, after all, be anathema to any group of scientists? As a researcher in the field of brain disorders, I am interested in preventing diseases as well as treating them, and I believe that guidelines should serve a similar pair of functions. Thus, my colleagues and I have sought ways to make our code a more effective educational tool, not just a reactive mechanism to deal with misconduct. And I believe that we are beginning to be successful.

To accomplish this, the SfN has added to its lively array of lectures and workshops on ethics, providing several specific programs that complement the guidelines. This has been particularly evident since 2010 when the society issued revised guidelines for responsible communication and established an e-mail address through which individuals can make comments or raise questions about those guidelines. The SfN has also recently hosted two international symposia on responsible conduct in communicating scientific results, along with a two-day workshop on the subject. And the society has commissioned a manual that will contain its guidelines on research communication along with related ethics case studies, discussion notes and a bibliography for further reading. That manual will be ready for distribution to SfN members and other interested individuals later this year. We anticipate that it, too, will help the guidelines serve the educational purpose for which they were designed and keep them from being relegated to the archives. I urge other professional societies to take similar approaches if they have not already done so. Let us do away with magical thinking and rewrite Juvenal’s quote to state: “honesty is praised and then helps us change the world around us!”