My advisor cited her colleague's papers in her own methodology even though she did not use those methods. Does this qualify as scientific misconduct?

Question

I'm currently working on my MSc thesis under a PhD candidate's supervision. She has very recently published a paper as first author in a well-known journal where she claims to have followed methodology used by her colleagues and co-authors (in their other papers). However, I know that this is not true and that the PhD candidate is fully aware that she's lying too. I know that she has lied because the thesis of MSc students who worked under her previously state something entirely different from what the PhD candidate has cited as source in her paper.

Question: What I don't know is if this act qualifies as scientific misconduct and/or how severe this is. If yes, then what is an act like this called?

Some of the differences between the methodology she has actually used, and the one she claims to have used are minor. The issue is that there are so many of these "minor differences" that they might end up having major influence on the results. e.g for how long samples were centrifuged, rpm of centrifuge, temperature settings of gas chromatography machine, etc.

The major (?) difference pertains to composition of nutrient broth (microbiology) which I'm confident she has never even prepared throughout the course of her PhD.

Unfortunately, it's clear to me that the PhD candidate's supervisors don't care about this since they seem to be working in tandem to increase each other's h-index. So, reporting to them won't make much of a difference. But for my own sanity, I would like to know just how big of a deal is this kind of act?

Answer 1

The issue here is not citation, it is correct reporting. If a method varies from the cited version, the appropriate thing is to say something like "Method from FrozzBozz et al.[73], modified by change of RPM from 3000 to 6000, growth media from LB to M9, and phase of the moon from new to full."

But did the method actually differ?

Now, there are definitely arguments to be made about the level of detail that is necessary and appropriate in protocol reporting. For example, the details about how a sample is shaken may or may not matter, but are often omitted because people are often simply not very systematic about this. Likewise, layout of samples on a 96-well plate is typically not reported, even though there are sometimes locality effects. Thus, many of the things that you are concerned about might be reported by one group but not by another.

The key distinction between reportable and non-reportable method details is whether a reasonable reader might expect them to have a significant impact on outcome. In many fields, this is more clearcut: for example, an algorithms paper reporting the number of iterations required to solve a problem would not declare the system the algorithm is run on, but one reporting the time of execution would. In biological sciences, however, it is often quite unclear and reasonable people have very different perspectives here. Thus, the Ph.D. candidate might not even believe that they have meaningfully changed the method at all!

My advice it to start by approaching it from this perspective, in a non-confrontational manner. Start by just asking for help in understanding the differences in method. If they do not explain in a way that makes sense, then it would be appropriate to escalate to the PI that ultimately supervises both of you, and ask for help in figuring out what details of method variation are appropriate to include. No need to get into accusations of lying at this stage, and you both might end up learning new things in the process.

Answer 2

Hanlon:

Never attribute to malice that which is adequately explained by stupidity.

This appears to be an error or laziness, rather than misconduct. These things often happen when a student does an experiment and the supervisor writes about it. The supervisor thinks they know what the student did, but they don't.

Usually it is not a big deal. However, if the error has life safety implications, you might still need to point it out.

Answer 3

There are a lot of experimental details in this area of science that often are really arbitrary in the end. Centrifuge settings are a very common example of that. In many cases it really doesn't matter if you keep stuff in the centrifuge for too long or at higher g than necessary. But on the other hand there is often no real benefit to trying to determine the actually necessary settings. For non-critical steps it is very, very common to use somwhat excessive settings that you know are generally reliable, but almost certainly are not optimal. In those cases, if someone would centrifige a bit harder or a bit shorter, it wouldn't matter at all.

In chemical reactions there is another case that happens very often. Reaction times up to 8 hours are kind of common, above that you'll quickly get to "over night". Performing a reaction for exactly 14 hours is kind of annoying if you also want to sleep, and in many cases keeping the reaction for a bit longer doesn't do any harm. So it makes sense to just try the reaction over night, and if that works without issues, you generally don't try to figure out if anything between 8 hours and over night would work, it simply wouldn't get you any benefit.

Knowing which parameters are critical, and which ones are arbitrary is a very important skill. It's often impossible to determine that from papers alone. Even the composition of the growth medium can be irrelevant, depending on what exactly you do with your bacteria.

Strictly speaking, you really should mention any deviations from the protocol in your paper. If you deviate, you didn't do it exactly as mentioned in your reference and you should mention that even if it doesn't matter at all. Personally I'd strongly prefer to list the methods explicitly in every paper anyway (at least for the major methods used), chasing down a series of "as performed in citation XYZ" is no fun at all. But in the area I worked, there are quite a few changes that might look significant that I probably would ignore entirely. Depending on the exact topic, there can be a lot of non-critical steps that don't affect the result in the end. They might affect the yield a bit or something like that, but not the conclusion of the paper. Of couse omitting these details isn't entirely correct, it would be better to be exact here, but it's also not really scientific misconduct.

Answering your question is impossible without the exact details and domain knowledge. It really depends on whether the parameters that differ are important or not. I personally wouldn't assume misconduct from these facts alone.