“John, you know what causes most problems? Ambiguity. Not knowing which decision is the right one. You’re the best first officer I’ve ever served with and I consider you a friend. So there was no ambiguity. And no problem.”
Gideon in “Crusade”
I’ve written about the serious game “The Lab” by the Office of Research Integrity. In short, play it. From all four perspectives. Trust me on this, while a bit optimistic at times, it gives you some additional ideas on what is involved in research, not only regarding scientific misconduct and how to deal with it if you encounter it (PhD student). But also regarding work-life balance (Post-Doc) and what it means to actually advise students (PI’s perspective), or how to deal with complaints (RIO).
A few days ago, I had a look at the guide to “The Lab”, which provides some further questions to ask and also lists what happens if you chose the different options. And it answered on issue I had with the game: Whether it wouldn’t be fair to talk with the accused beforehand.
Suppose you work in a lab — no matter whether you lead it, are a post-doc, or “just” a PhD student — and a researcher is suspected of having committed scientific misconduct. You hear an accusation — which seems to have something to it based on the presented evidence. But let’s face it — you know the person who is accused, you like that person, you really can’t imagine this person committing such a blatant fraud. This person is your friend, or your protege. Shouldn’t you give this person the benefit of a doubt and talk with this person? Honestly, straight up, upfront, face to face. Let’s see whether there is something to this issue or whether it will be — most likely — due to some stupid misunderstanding?
When playing the game I thought “no”. The best thing you can do as supervisor (PI, or, actually, any other person in the lab) is to remove yourself as best you can from the situation by handing the investigation over to someone impartial. If there is the possibility of misconduct, anything that happens within the lab has a rather bad taste to it. Things that are handled internally rarely see the harsh light of the day. Even if you were capable of doing so, questions remain.
So, while playing the game, once there was credible evidence I turned to the Research Integrity Officer. No matter whether as PI or PhD. But some doubt lingered whether this was really right.
Until I read in the guide that whenever you confronted the person suspected of having committed fraud this person would then conceal the evidence. And yup, that’s something I actually had not considered. I thought that misconduct always leaves tracks (e.g., “rare” statistical occurrences in the data). But sure, while you cannot retroactively (re)do experiments, you can remove proof of misconduct. That does not leave the person standing there as innocent, but as not guilty. It will also make it impossible to find out the truth.
It might seem the civil thing to do — inform the person accused of scientific misconduct that there is an accusation. But the second this happens this person is also forewarned. Even if you trust that person, and no matter whether this person is guilty or not, just the fact that this person knew that an investigation is at work enables this person to conceal evidence. As a consequence, it becomes impossible to decide whether a lack of evidence for fraud means that the person innocent, or whether this person has just concealed any evidence. And that doubt is poison. In any investigation of scientific misconduct, there should not be doubt — at least, not of this magnitude.
As strange as it might sound, by talking to someone against whom a complaint was made, by trying to remove doubts, you can actually ensure that the doubts will never be removed. The worst betrayal you can do in such a situation is to inform the person who is accused before the data and material can be collected without forewarning. And my guess is that it will raise issues in the PI – Post-Doc relationship. But seriously, if a Post-Doc would not do the same thing if roles were reversed … that would be the bigger issue. If the person turns out to be innocent, point to the evidence, ask how the person would have reacted, and point to the fact that by not addressing the issue face to face you have ensured that the investigation can be trusted. Not taking the side of the accused, even if you totally trust this person and want to protect this person, is a virtue. The investigation must show that the person did nothing wrong, not the supervisor, colleague, or subordinate.
So, have a look at “The Lab” — and the corresponding guide on the ORI page. And in case you ever encounter misconduct, have a look at the really interesting article by Gunsalus.