Nature Medicine | Spoonful of Medicine

Your cheatin’ heart

Over a year ago, we published a special report about fraud in science. One of the articles in the special quoted Darrell Smith, from Vanderbilt University, as saying “Graduate students would do anything to please their principal investigator”.

Dr. Smith then wrote to us to clarify that “To suggest that a graduate student or postdoctoral fellow would do anything to please their principal investigator indicates a willingness to engage in unethical, illegal or immoral conduct”, something that, he felt, was unthinkable for most graduate students and postdocs, who “are committed to high ethical standards in their research and would not engage in misconduct to please an advisor”.

This weekend, I got a letter from someone called Chris Muller, who disagrees. He writes:

“A bunch of us were discussing Dr. Smith’s commentary on how graduate students and postdocs have their own strict codes of conduct and ethics that makes them think of science first, and their PIs and their careers only afterwards. How noble! How pure! […] I personally know graduate students, postdocs and technicians who HAVE TO think of whether their PI will be happy with a given result, and if it doesn’t match the “desired” outcome, well, then they do the experiment till it does. If they didn’t do this, their survival, their degree, their ability to earn a livelihood based on their chosen profession would be threatened. It’s that plain and simple. What would you tell these poor sods — to destroy their own ambition and lives to please some abstract, non-existent, popular notion of selfless and ethics-driven professional suicide to appease even ruthless careerists who spend most of their time traveling for PR, writing grants and schmoozing their way into committees to influence policy-making and, in the process, earn THEIR livelihoods and vacations to exotic, tropical and skiing locations?"

The last question is somewhat confusing, but he seems to doubt that there is such a thing as a strict code of ethics worth upholding because, if you were to stick to it, you would be committing professional suicide. Quite a radical view, although I must say that, at Nature Medicine, we have seen cases of students or postdocs who fudge data, and we have seen our fair share of referee reports that say “it seems that these people will do anything to get the right result”.

To say that in science there will always be all kinds of people — those who stick to their code of ethics no matter what and those who fudge data to please their PI — is a platitude. What one would like to know is if Dr. Smith is right and we are looking at just a few rotten apples or at a terminal patient with metastasis in every organ. My own opinion, I regret to say, is closer to the metastatic view.

And yours?

Comments

  1. Report this comment

    Mr. Gunn said:

    If you’re going to make an accusation of widespread dishonesty among graduate researchers, you’d better put the data where your mouth is.

    Release, in anonymized fashion, the aggregate numbers of fudging incidents vs. total submissions.

    The biggest offenses I ever see are people trying to support a big conclusion with a too small sample size, and I think that’s inexperience rather than dishonesty.

    Seriously, if you’re going to make such a claim, you’d better substantiate it somehow or retract it with an apology.

  2. Report this comment

    Chris Muller said:

    If one says there is dishonesty everywhere, does one expect people to come up and swear to tell the truth and declare that they have done this, that and the other dishonest thing? I have actually worked with graduate students, postdocs and technicians who have all had to face extreme pressure from advisers who have told them that a particular experiment didn’t work well because the result was not consistent with their hypotheses and that they were hoping for better work from these poor hapless souls at this stage of their lives. Of course, the ones who were too honest or pigheaded to actually do anything about it and simply stood their ground saying “the data is the data” had a lot of difficulty graduating and publishing and moving on. The ones who understood the fact that the boss is the boss were the ones who simply repeated the experiment with different controls, changed the way they plotted the data and arrived at conclusions that the bosses wanted or at least said that “..in light of some of these current observations, further studies are warranted that would investigate the other underlying causes for these observations and their effects on our hypthesis.” In other words, fit the facts to the hypothesis. Now mind you, I won’t say all scientists are like this. Many of us have lived with the dull and stabbing disappointment that our hypothesis didn’t pan out and had advisers who were malleable (and basically nice enough) to ultimately accept this. But nothing in life can be generalized and certainly not all the things we encounter in life are to our liking. Nobody is going to be substantiating things to please you, Mr. Gunn, no apologies will be tendered nor statements retracted. Live with it.

  3. Report this comment

    Mr. Gunn said:

    It’s not you I’d like to hear some validation from, Chris. It’s the guy with the view that academic dishonesty is extremely widespread and says “we have seen cases of students or postdocs who fudge data”.

    Well, how many cases?

    What I’m saying is that if the number of cases where “it seems that these people will do anything to get the right result” divided by the total number of submissions is large enough to justify a claim that dishonesty is widespread, then tell us what that number is.

    And if that number is well below 0.01, then doesn’t that mean that you’re the one fudging things a little in an effort to be sensationalist?

    Feel free to argue that you’re a science writer, not a scientist, and therefore are expected to take such liberties.

  4. Report this comment

    Juan Carlos Lopez said:

    Gentlemen, no need to lose our cool.

    The only hard data I’ve seen comes from The Journal of Cell Biology. They check every paper they publish and as many as 10% have figures that have been manipulated. An overwhelmingly large number of them (1% of the 10%) seem to be just cases of “beautification” (trying to make the data look nicer) and not “fudging”, but who is to say?

    Now, this is certainly an underestimate, as there are a lot of figures in papers in which one doesn’t have access to the raw data.

    Again, very hard to know the real magnitude of the problem, but my hunch (and it’s just a hunch) is still that the problem is more widespread than these figures.

  5. Report this comment

    Chris Muller said:

    I agree, no need to lose our cool. Whether we get excited about this or not, things don’t really change. But this is an interesting topic so I can’t resist adding some things. I started by refuting what Dr. Smith had said that graduate students think of ethics first and careers afterwards. I disagreed with this notion as did many of my colleagues. Upon re-reading JCL’s understanding of my message, I didn’t agree with what he was saying.

    It’s not so much that academic honesty and ethics are not existent or worth upholding. There are many good publications and scientists who are very solid in their approach. If the hypothesis doesn’t work out, then think about it and see why and change that, not the facts. But what I have seen is that this is a personal difference. Ethics really cannot be taught in graduate school. If you have succeeded in entering graduate school and are a complete crook by this time, brochures, grandiose lectures and threats won’t deter you. If you have worked hard and are an ethical person before grad school, seeing others cheat will disillusion you but will (probably) not turn you into a complete crook.

    But to say anything in favor or against what is really seen in academia is just discussion. And to say that it’s all the graduate students’ or postdocs’ fault is very very wrong. It was Dr. Smith’s contention that I was refuting. So in other words, I do agree with JCL’s supposition that academic dishonesty has metastased a great deal. You have raised an interesting question though, Mr. Gunn, surveys and questionnaires try to look at so many things but is it possible to measure dishonesty although we know this is a ubiquitous phenomenon? The very existence of dishonesty precludes its detection or measurement. I too have come across many papers who made exactly the mistake you have described—they try to extrapolate the conclusions from a very small sample size to make huge sweeping statements.

    Now, I believe that mistakes made in ignorance are one thing although the whole aim of a researcher should be to dispel the ignorance which exists at the start of any undertaking. But reading, thinking about one’s project & interacting with others should help one improve in study design, interpretation of data, troubleshooting etc. So why then unless a paper is the VERY FIRST ONE IN ITS FIELD AND OF ITS TYPE does one see the same mistake being repeated? I have heard this from others (again no numbers provided) so here’s my guess—precedence. If the other guy in Scripps was able to publish this with so many replicates and do this experiment and conclude thus, why can’t I? And so it grows. But you do strike a low blow, Mr. Gunn, why should a scientific writer be excused from the same accountability that a scientist is subjected to and why call JCL a scientific writer as if that is an insult? Anyway, that’s a topic for another day…