Today at Inside Higher Education, an article identifies "The Real Science Ethics Issues". Which means, I suppose, you don't have to keep taking my word about what's an issue and what is not. The focus of the article is not on the flashier instances of fraud, but on more mundane stuff that may rot the scientific enterprise from the inside.
Nicholas H. Steneck, a University of Michigan history professor and a consultant at the Office of Research Integrity at the U.S. Department of Health and Human Services, said that most research fabrication is "not very subtle or clever." If the research is important, Steneck said on Friday at an American Association for the Advancement of Science panel in Washington, other scientists trying to replicate it will often expose the fraud.
When scientists don't report data that contradicts a conclusion, or take funding from places with agendas that extend beyond science, however, the practices usually go unchecked, and, in the long term, have the potential to erode public confidence in scientific research, he said.
A June 2005 paper in Nature compiled responses from 3,247 scientists funded by the National Institutes of Health. Over 15 percent admitted that they had changed a study design to please a funding source, and 10 percent admitted they had inappropriately put their own name or someone else's name -- a practice commonly know as "guest authorship" -- on a published paper.
Steneck said that, though practices like guest authorship don't damage research results, they do damage "the integrity of the practice."
"I would argue questionable practices are worse than [outright fraud]," he said. He mentioned other practices, like trying to pad a rÃ©sumÃ© by publishing the same paper in two different journals, which damages meta-analyses. "This stuff affects policy," Steneck added.
John Horgan, author of The End of Science, and director of the Center for Science Writings at the Stevens Institute of Technology, cautioned scientists about "taking money from people whose values make them uncomfortable." ...
If some of these "questionable practices" aren't addressed by the science community Steneck said, the government will have to take the lead. Steneck added that scientists have been very slow to adopt any procedures that would clamp down on questionable practices. "I know you don't want to hear this," Steneck said, "but if institutional and professional commitments do not improve," the government may have to regulate more to make sure funds aren't wasted and public confidence doesn't suffer. As research and economic interests become increasingly intertwined, the integrity of science research is more often at risk.
Felice Levine, executive director of the American Educational Research Association, said that the government is already sticking its hands way too far into the research community. "Political forces have sought to encroach on peer-review," she said. She noted that, for example, veteran scientists on review boards at institutes at the National Institutes of Health were replaced by people with less scientific background, and who were instructed to critically review human sexuality studies.
Levine said that, now more than ever, "the state of legitimate peer review is not very secure."
Longtime readers know that I have raised concerns about guest authorship. Indeed, I think there are parallels between dishonest representations of who is responsible for scientific work and plagiarism. (I'll have the longer plagiarism post up as soon as I slog through some grading that really needs doing.) As quoted above, Steneck seems a little vague about how exactly guest authorship and other questionable practices damage the integrity of scientific practice. Here's a suggestion: If it's common practice to lie, it's impossible to tell when people are actually telling the truth. And, in an activity that requires not only working with other scientists and the knowledge they produce, you need to be able to trust these other scientists.
Another issue the article flags is the effect of funders' agendas on the research scientists conduct. Perhaps study design gets tweaked to please the funders. In some cases, that might be problematic (e.g., if certain possibilities are being ignored because the funder would prefer that they be ignored), but in other cases this might be fine. What would be more worrying to me are instances when researchers altered the results they reported, the data included in their analyses, or the spin they put on their conclusions to suck up to the funder. The funder can surely specify the question to be studied, but cannot tell the researchers what the answer is -- finding the answer is the reason you bring in the scientists to do the research.
Finally, the article poses the question of how to improve the collective behavior of scientist who are now completely used to gaming the system, whether through minimal publishable units published in multiple journals, or guest authorship, or whatever. Can scientists figure out how to police themselves? Must the government be brought in to provide the oversight? (This government>? The one that seems not always to care for the knowledge scientists are building?)
This is a good question. It seems to me certain features of the institutional structures (both academic and corporate) in which science is practiced tend to lead to behaviors which really do undermine the integrity of the scientific community. While these behaviors may seem expedient (and minor) to the people displaying them, in the long run they really could undermine both the scientific community and the goodness of the knowledge it produces.
If the goal is really to build a good body of scientific knowledge, this should give scientists pause.
Can scientists overcome their inertia to change the scientific culture? Or are they comfortable leaving it to the government to make changes at some point down the road?
This article was very helpful in alot of ways.
Here's one part of the answer: ethics courses in undergrad and postgrad science education. I think the situation is a bit better in the US, but I studied in Australia and went through a BSc and then a PhD without one class on philosophy of science, research ethics, or anything related. I think that if working scientists were more accustomed to think about ethical issues and more familiar with ethical arguments/decision making, they would be less inclined to cheat.
Guest authorship is a tough one. I'm not sure I agree with all the criteria you cite from the Medical Journal (or whatever it was) in that previous article.
I'll give you a specific example. My much celebrated "Knop 2003" paper has a whole bunch of authors on it, not all of whom are qualified to talk about every detail in the paper. (In fact, I may be about the only one, and there are one or two details in the disucssion of systematic error that even I'm a little fuzzy on the actual implementation of.) I'll highlight two authors in particular, though: S. Garmond and K. Garton. Both of these are undergraduates, who worked with me (although K. Garton worked more with S. Burns, who is also on the paper) over a summer. Both did good work on what they did. What I had them doing was some detailed and annoying stuff that had to be done or checked, but was reasonably labor intensive. One was verifying that I wasn't hurting myself by being a little cavalier with the psf's I was using to fit photometry, the other did a lot of preliminary data reduction. Both made substantial contributions to the paper. Both, in my opinion, deserved to be co-authors. Neither, almost certainly, could discuss the cosmological fits, or really take responsibility for all of the analysis. (S. Garmond is now a graduate student at Texas; I don't know what he's doing, but it is possible that he knows a lot more now on this topic than I think.)
By strict definitions, both of them (and a lot of other authors) should have been left off of the paper. Yet, many or most of the authors on the paper made substantial contributions either to the development of the project, or to the aquisition and reduction of the data, that made the whole thing possible. They deserve some recognition and credit. Are they "guest authors"? I don't know.
By the way, I've asked that collaboration (which I'm no longer really a part of) to stop putting my name on papers by default. They used to do that, but there was one paper that I didn't like some things that were done in the analysis, and I realized that I didn't have time to pay enough attention to all of the papers to make sure that I liked them all. There was some resistance on this; papers will still come out to which I contributed substantially to the work way back when, and as such I "deserve" some credit. However, I also am not sure that having a large number of nth author papers is helpful; some people object to a low ratio of first-author to nth author papers, even if that ratio is *lower* than the average author list. (So much for fairness.) But that was my choice. I think that a lot of "guest authorship" really is legitimate, at least in Astronomy, but not all of it. Within the field you have to understand whether or not every single author really should be able to defened every bit of every paper published.
We are quite strict on co-authorship. We do not do it even within the lab if the collaboration is strictly discussion. Help pipetting, and you're on my paper. I'll help you with surgeries, and I'm on yours. Of course, we all talk about all of our work all the time.
Authorship becomes all the more fuzzy since (as coturnix implies) standards differ so much from field to field and country to country. In some places, for instance, the head of the lab is always, naturally added, in other places just as naturally not. Apparently it's normal in some places to have your name added even if you have never even seen the paper in question.
Republishing data is another slippery slope. Publishing the same paper twice is certainly dishonest. But allowing the use of any data once, and only once, is of course impossible. So what criteria should there be for when the work differs enough to warrant a new paper?
We do not do it even within the lab if the collaboration is strictly discussion. Help pipetting, and you're on my paper.
What if I'm a summer student, and all I do is load one gel with the samples you gave me in the order you told me? What if I'm a TA? (My take: summer student off the paper, TA on; but I've seen the reverse done.)
In some places, for instance, the head of the lab is always, naturally added
I think this is pretty much universal -- last author is senior, the PI in whose lab the work was done. The argument, with which I agree, is that without the PI there'd be no lab, so no paper. Even if he/she hasn't touched a test tube in ten years, the grants that bought those test tubes didn't write themselves.
We do not do it even within the lab if the collaboration is strictly discussion. Help pipetting, and you're on my paper.
It's interesting to note that in social science this is often quite different. If you help to add to the discussion and formulate ideas you might be asked to co-author, but if you aren't adding to the research in any way other than by following the steps told you by an upper level colleague, you may not be included. That's why we don't usually have the 8+ people listed as authors, and why we tend (I think) to have less publications coming out of grad school.
Summer student vs. TA - it all depends.
The person has to be involved pretty heavily both in the intelectual and the physical aspect of the project.
We used to have a technician who always ended up in Acknowledgments as he was not present at lab meetings. We had undergrads helping with the manual work and also ending up in Acknowledgments. But, one of them was interested enough to attend the meetings and contribute also to the intelectual part of the project and will be rewarded with co-authorship once I get off my ass and finish writing it up. Her work in our lab got her accepted in 7 out of 8 top-level grad schools she applied to, just on the strength of the PIs letter. Wherever she went they talked about her as "the girl with the capital-L Letter"!
Being a part of a very broadly diverse department, I noticed that molecular folks tend to have a lot of authors per paper with quite a lot of politics about the exact order, the physiology/behavior folks tend to have 3-4 authors and it is easy with the PI in the last slot, while the ecology grad students often publish single-author papers (but they also mostly get their own grants as well).
I agree with Megan. What Rob Knop describes sounds like the duties of a "Research Assistant" in Economics. And they get credit - an appearance in the initial footnote, being thanked for helping ("we thank Bla Bla for his excellent/outstanding/ able/adequate research assistance").
I don't know how this works in other disciplines.
I find this all very interesting. But I have another issue to address. What about the issue of a PI not publishing a completed paper just because she lost interest in the project (FYI, the science is very sound). Is it ethical to hold back the career of a future PI just because both you as the PI and the student who is now a Post-doc have moved on from the project? My former PI has now done this to three graduate students (me included) and two post-docs. I even confronted her before I graduated because I was afraid this might happen. It has been a year and a half since I left the lab. I have two fully revised (by both me and her) manuscripts waiting to be submitted. I have put gentle pressure on her to submit the papers and each time she responds, there is a different excuse. I fear if I put too much pressure on her, I may receive bad recommendation letters from her for the rest of my career.
(It is not lack of money which prevents the manuscript submissions but total lack of interest in the topic by the PI). At least five other papers from the lab have been published since I left, none of which were on the same topic.
Is this really ethical?
Unfortunately, in the academic world today, it is almost impossible to separate the interests of science from the interests of business. And since science, in many respects, is the main business venture of academic institutions, time and again, when the ethical considerations dictate the exposure of a fraudster, the business considerations, almost exclusively the domain of administrators, trump the ethical considerations. That is why I believe there is much more cover-up of scientific misconduct done by university administrations than is known to the public. Moreover, the success of a research scientist today depends largely on extramural funding. Without it one can forget about promotion and tenure. There are those who will sucumb to the pressure of receiving a grant and securing promotion and tenure, even if it means cheating, and there are those who will keep quiet, even when they are aware of such cheating by their colleagues, because they cannot afford to lose their job, which is paid by the grant of the cheater. You can read more about the cover-up of scientific misconduct in my book (see URL).
There's another aspect of authorship that is often overlooked - and that's of its role in encouraging students to pursue a career in the sciences. Several years ago, we were working on a survey [yea, the science is marginal] of differences in the care of patients with a particular cancer across the US. With over 1000 respondents, we had beaucoup data to crunch. An interested undergrad pre-med student went through thousands of handwritten data forms to verify our data. With most of the analyses pretty straightforward, the student was very familiar with the project which led to recommendations hopefully affecting the care of many patients. The PI sent the paper off to a journal that did not require signatures from all authors - so none of us saw the final version. Having to do some additional analyses as the result of the review, I saw that the student was not listed as a co-author and was not acknowledged. When I asked the PI why the student wasn't mentioned, he said that the student was a just a work-study student and that he paid him $5/hr for his efforts. At times there appears to be an unwritten rule that if you pay someone for their assistance, it doesn't count as "true collaboration" and thus the person isn't included as a co-author; that "hired guns" are just in it for the money - it's a job, not a career.
In my case, I asked the PI to include the student and got nowhere, so I just dropped the dept chair [an old friend] an email about the problem and sure enough, on the re-submitted version, the student was now a co-author. Plus I found that the PI had been instructed on the value of mentoring and of giving students research experience.
About a month after the paper was published, I ran into the student who had taken a copy of the paper home over the Christmas holidays to show his family and friends. You just can't imagine how excited they were at seeing their name in print for the first time - and also how proud it made their family. For the first time they were seen as a Real Scientist - and who knows, maybe in 10 years they will be.
Some of us who have been doing research for much longer than 10 years sometimes lose track of how important these little acknowledgements are - a pat on the back that costs us essentially nothing. If it encourages a young scientist who fully participated in a project [even if they were paid], I'd be liberal with authorship.
I love this it is very alsome.
The article was good......it explains alot of things that i would of never knew!
This article told me a lot of things that I did not know.
this was an amazing articl