"Scientist" or not?

[TenaciousBushLeaper]

Over the past year I've come to notice a trend in the lab I'm currently part of, and other well known successful labs in my field, though there are exceptions.  For a bit of context, to do even remotely decent research in my field it requires a good amount of computer programing knowledge and skill amongst other things, (mathematical maturity, excellent writing, social skills ).  Throughout undergrad I always thought of successful researchers as these amazing polymaths with tons of domain knowledge, a (relatively )deep understanding of mathematics and statistics, and computer science. Though what I've come to understand from my current position and experience is that this really isn't the case, at all. What seems to be common is, you have person A, who is a post doc or a graduate student, who has this great idea and has read a ton literature on the topic. Then you have person B, and maybe even person C. A, B, and C all communicate with one another in some capacity, A is usually the "leader" and B and C are the technicians. They do the math, write the code, do the analysis, replicate the results, make sure everything is square and good to go and then send the data over to A. Person A then has B and C explain all the steps, proceeds to publish and the rest is history. This is where I start to get uneasy about calling person A a scientist, for a few reasons:

 

1. If not for the technicians person A wouldn't have the results that he/she has, wouldn't know how to do the analysis, and not be capable of it save you stop time for(at least) a couple of years for the requisite knowledge and skill to be acquired.

2. B and C could, in theory just lie about the results with enough reasonable sounding BS talk and person A might not be able to discern solid research from...not so solid research.

3. Removing the possibility of any intentional foul play, what about the cases of innocent  hiccups and mishaps that may go overlooked, how is person A ever to identify these events?

 

Now, if I step back for a bit,  this setup  really does make sense, it seems ideal. Why? Because you can't possibly expect one person to know all of these things at once, and in the few cases that you do have this, it's because someone has come over from a BS/CS masters or BA/BS in math, and has along the way held an interest in the topic. If university admissions committees were to limit admissions only to those with this type of background, our graduate departments would soon cease to exist

 

So therein lies my little internal conflict with what people do in this field  that I've dedicated so much time to. 

 

Opinions on this matter? Is this common in your line of research? How am I looking at this in the wrong way? 

[TakeruK]

If I understand what you are describing correctly, then this is all very normal. In my field, collaborations are often with many people. Maybe not as big as labs, but the papers I've led are usually with around 10 other coauthors. Some of these coauthors contribute directly to the paper, either by providing data that they have analyzed and/or plots they have made. I trust them to have not lied to me. But in addition to the trust, it's the responsibility of the "leader" to ensure that they believe the results they have been given. So, I make sure I actually understand their work. I have my coauthors explain their analysis to me and when they do, they use citations to works that prove their method is correct. I ask questions until I am satisfied that I understand what they did and everything about their results make sense.

But you are right that I don't actually go and do the work myself. They might have said all of the correct explanations but made a mistake in the code, or in the worst case, actually just made up the results and didn't do the work at all. This is why we are all coauthors---when we collaborate, we all stand behind each other's work together. It is a necessity because one person cannot do all of the work independently. Science requires collaborations.

In this example, I am the "person A" in your analogy. I just published a paper as a "Person A" but in the last few months I have been a "Person B" or "Person C" in 3 or 4 other papers. Everyone else is in a similar situation too so no one wants to be a lying "Person B" because when they are a "Person A", they don't want their "Person B"s to be lying to them. This is why in academia, our reputation as honest researchers are so important. This is why academic honesty and integrity is so important and when you see people posting about ethically dubious things on these forums, other researchers are quick to shut it down.

The system is not perfect and there are many examples of paper retractions and careers being ruined because of bad decisions. But it does work most of the time. It's also important to choose collaborators wisely!

[TenaciousBushLeaper]

14 hours ago, TakeruK said:

If I understand what you are describing correctly, then this is all very normal. In my field, collaborations are often with many people. Maybe not as big as labs, but the papers I've led are usually with around 10 other coauthors. Some of these coauthors contribute directly to the paper, either by providing data that they have analyzed and/or plots they have made. I trust them to have not lied to me. But in addition to the trust, it's the responsibility of the "leader" to ensure that they believe the results they have been given. So, I make sure I actually understand their work. I have my coauthors explain their analysis to me and when they do, they use citations to works that prove their method is correct. I ask questions until I am satisfied that I understand what they did and everything about their results make sense.

But you are right that I don't actually go and do the work myself. They might have said all of the correct explanations but made a mistake in the code, or in the worst case, actually just made up the results and didn't do the work at all. This is why we are all coauthors---when we collaborate, we all stand behind each other's work together. It is a necessity because one person cannot do all of the work independently. Science requires collaborations.

In this example, I am the "person A" in your analogy. I just published a paper as a "Person A" but in the last few months I have been a "Person B" or "Person C" in 3 or 4 other papers. Everyone else is in a similar situation too so no one wants to be a lying "Person B" because when they are a "Person A", they don't want their "Person B"s to be lying to them. This is why in academia, our reputation as honest researchers are so important. This is why academic honesty and integrity is so important and when you see people posting about ethically dubious things on these forums, other researchers are quick to shut it down.

The system is not perfect and there are many examples of paper retractions and careers being ruined because of bad decisions. But it does work most of the time. It's also important to choose collaborators wisely!

I may be wrong but, from your statement of  everyone being person B and C at some point, this suggest that you could carry out the analysis, had that been your role in project X, where you are person A. The cases that I talk about are ones where person A would be incapable of doing it, they wouldn't be able of fulling the roles of persons B or C, if someone else was person A. It's that fact that sort of rubs me the wrong way.

[TakeruK]

7 hours ago, TenaciousBushLeaper said:

I may be wrong but, from your statement of  everyone being person B and C at some point, this suggest that you could carry out the analysis, had that been your role in project X, where you are person A. The cases that I talk about are ones where person A would be incapable of doing it, they wouldn't be able of fulling the roles of persons B or C, if someone else was person A. It's that fact that sort of rubs me the wrong way.

That's a good distinction, but I do mean it in both ways. I have had Persons B/C as coauthors who have done analysis that I could not do myself (although, sure, in theory, if I spent 4-6 months learning a specific skill then I am capable of learning it, but not capable of doing it now) and I've also have had Persons B/C as coauthors who do stuff that I know how to do, but since they already have a tool coded up, I just use their analysis instead of re-writing the code from scratch. And of course I've been in the flip side of things as well. In the latter case (A could do B/C but B/C already has a tool made), this would only be cool if the analysis is not the primary focus of the work. Whenever I am a Person A and my paper's main argument depends on an analysis, I do ensure I learn it myself and I have re-written code that others have provided me to ensure that I really truly do understand the critical analysis holding my whole argument together.

As for your concern (i.e. A not being able to do something B/C has done), I'm not sure why this is always a bad thing. Here's an example of why it must be this way: In my first paper, I was a Person B and there were 29 coauthors in total. The first author was the PI of the whole project, which was a balloon telescope. About 12-15 of the coauthors were involved in designing and building the telescope, and the other half were involved in various analyses. I would say that the PI was fully capable of doing the analysis of the second half of coauthors, but not on the spot. If you quizzed the PI on the fine details of the analysis that I made for example, he would have a hard time explaining it. He would be able to explain it in general terms, but he would probably forward the person to me if they had more questions about the finer details. However, the PI would not be able to replicate the contribution of the coauthors that built, designed and tested the telescope at all. This would be something completely outside of their expertise. 

I think this is okay. For big science projects, the leader is not really acting in a role of a "scientist" any more. They are really working as project managers. In fact, for the telescope in question, I would say that the main contribution of the PI was his leadership and project management. He coordinated a collaboration of 10+ schools across at least 4 countries with around 30 people in total. I think the PI did very little in terms of actual scientific analysis. 

And in my current department, there is a senior professor that does very little science nowadays. Their main role is to write grant proposals for their grad students, postdocs and staff scientists to carry out. For some projects, they are not even capable of doing the science any more because their skills are obsolete (e.g. they learned to program in the 70s and their coding skills are both rusty and maybe not applicable to the programming languages of today).

However, in both of these cases, I would still call both of these people "scientists". And the reason is because they were originally trained as scientists and I think their approach to solving problems in our field is still a scientific one. Their day-to-day role may be more managerial than scientific analysis, but they still work to enable science to happen. I don't think the term "scientist" should be only applicable to what they actually do, but instead, to the kind of contribution they make to the community.

Finally, just want to point out that not every senior professor goes on this path. Although it may depend on the field, many professors do end up more and more like managers as they age, there are profs in my department that have always had small groups (just 1 or 2 students and a postdoc) and have been writing papers and doing science ever since they were graduate students. I don't think one is any less than the other, they're just different ways they are supporting the department and our community.

[eternallyephemeral]

The definition of scientist here is something that others would probably disagree with you on.

 

There's scientist as the job you have, and then there is scientist as the type of work you do.

 

Would you say person B/C is not a scientist? Their position might not say scientist, but they are part of the scientific process.

 

As also mentioned above, there might be multiple reasons why the person A can not do the analysis/other steps. It is best to specialize because then each person can get really good at something, so in that case they could not do person Bs or Cs work. But perhaps their skills are outdated, someone else is better at those techniques, or they don't have the ideal background for this work.

 

I would agree with your basic premise that it's not good not to know what your collaborators are doing, especially if you are in charge (like PI or first author). I also agree that people who are not aware of what is going on with the work/collaboration are more susceptible to errors or falsification for multiple reasons.

[JQR]

First of all, this is truly an exercise in semantics. Is someone an 'athlete' if she is in high school sports? What about a rec soccer team? What about pick up soccer? I haven't played competitively in years, does that mean I am not one?

 

Or take business, for example. Which persons are "businesspersons"? The CEO? What if the only thing she does is wine and dine clients? What about the President who relies on the CFO for actual understanding of financial performance and the COO to execute her business 'ideas'? 

In the same way that all of these people are in some way 'businesspersons', I think that in the realm of science--especially given the sophistication and diversity of the field today--it can only be for the sake of vanity to decide who is "not" a scientist.