Assistant Professors who cannot let go of the details are screwed
May 7, 2010
Candid Engineer has reached a critical turning point in the gradual maturation of a scientist from bench jockey to Principal Investigator.
I don’t know if I ever really realized that the day would come when I wouldn’t understand the details of experiments being done in my name. That at some point, my interns would be doing experiments that I can’t even pronounce, that they’d be teaching one another, without passing the information through me. That one would come to me and say, “I’ve been thinking about it, and I’m almost positive that it would be better to smash the Mango peel before adding it to the blender”. And that I’d look at him, and think “I have no fucking clue what he’s talking about because I don’t know how to operate the blender”, but somehow smile and say, “I trust you, you know what is best experimentally”
.
That day had better arrive or your are going to be in for a world of hurt as a lab head trying to survive and get some science accomplished.
Keep this in mind when it comes to junior PIs at the bench. Tattoo it on the back of your/their eyelids.
It is NOT your job to be a postdoc, grad student or tech. You are the lab head and you have other duties which require your attention.
qaz is demented:
As exhilarating as it is, I strongly recommend that you learn how to use the blender. Even once. You really don’t want to cede control of the details. Let the intern teach you.
There is a classic story about a postdoc who came into a new lab, asked about the control solution to calibrate an instrument which needed to be calibrated each time before use, and got told “we don’t need to do that anymore” by the tech. (Tech was incorrect BTW.) When the PI found out, the feathers flew because over two years of data was garbage.
There is nothing about this little vignette that requires the PI to know every detail..in most cases. The PIs job is to look for weird stuff in the data, ask about proper controls and validation, etc. To make the people closer to the science explain the science, if not the specific details until it is relevant. The PIs job is to be alert for when something is not adding up. Like the “garbage” data to which qaz refers.
YES, it is risky. There are going to be screwups, some of them quite painful and time consuming.
This is the cost of doing business.
The more successful PI concentrates her time on getting good people, training them properly and backstopping their interpretation of their results. NOT on micromanaging each technique and experiment.
That way lies disaster.
DrugMonkey 
May 7, 2010 at 12:41 pm
You’re absolutely right DrugMonkey, and I find this fact of academia exasperating.
I am one of those detail-oriented people who likes being in the trenches. I went into science because I like doing the actual hands-on part of science. I like spending time with the organisms. I like troubleshooting issues at the lab bench. And I’m good at it- I’ve never had any trouble landing a postdoctoral position.
Likewise, I’m not fond of meetings, or of handling the organizational aspects of science. I didn’t go into science to be an administrator. Yet that’s where the prospects for career advancement lie. In my field I max out at 35K/year doing what I’m good at and what I enjoy.
What do I get for being good at science? The choice between being “promoted” into a permanent PI job that I’d rather not have, or perpetual uncertainty in low-paying temporary positions.
LikeLike
May 7, 2010 at 12:46 pm
Yep, this is one model for running a laboratory, and the only one that can work if you want to have more than three or four people in your lab. However, it is possible to run a small lab–three or four people total–very effectively in which the PI operates in large part like another trainee: sitting at the bench and doing experiments and knowing everything about exactly what the other two or three people in the lab are doing.
LikeLike
May 7, 2010 at 12:56 pm
The more successful PI concentrates her time on getting good people, training them properly and backstopping their interpretation on their results. NOT on micromanaging each technique and experiment.
This is wrong. Job #1 for the successful academic PI is getting funding. Full stop. End of story. If you can get funding, given that you have already been qualified as a PI by getting the job in the first place, everything else will work itself out. If you can’t get the funding, no amount of “getting good people, training them properly and backstopping their interpretation on their results” will save you.
LikeLike
May 7, 2010 at 1:01 pm
yes, whimple, of course. I should have said “lab time” or “experimental time” or some such qualifier. I tend to assume my audience knows what I think the PI’s job is really about but I should have elaborated.
LikeLike
May 7, 2010 at 1:04 pm
DM – I never suggested micromanaging. All I suggested was checking in regularly, and making sure that you understand the techniques being used in your lab.
Absolutely, micromanaging is disaster. Trainees need room to explore and discover things on their own. But so also trust without verification is disaster. That two years of garbage could have been saved by the PI knowing what was going on in the lab.
More importantly, trainees are very unlikely to have as much depth of knowledge about a topic as you do. I’m pretty sure that CE knows more about Mango slicing than the trainee, and I’ll bet that if CE took a day to learn how to use the blender, CE would know whether mashing Mangos is the way to go.
LikeLike
May 7, 2010 at 1:04 pm
In the medium and long term, this is the only way to get and maintain funding.
LikeLike
May 7, 2010 at 1:06 pm
“it is possible to run a small lab–three or four people total–very effectively in which the PI operates in large part like another trainee”.
Yes CPP, I’ve seen that and it is actually exciting and very productive science to practice and teach. The alternative might work very well for people with a very strong call for business administration.
LikeLike
May 7, 2010 at 1:21 pm
Welcome to management. Management rule #1:
For all X: Your job isn’t to do X. Your job it to see to it that X gets done.
Doing X is probably your comfort zone. If you want to stay in your comfort zone, there are job titles that are appropriate. You don’t want those job titles? The learn to deal with the new one instead of shirking and doing your old job.
LikeLike
May 7, 2010 at 1:47 pm
I’m pretty sure that CE knows more about Mango slicing than the trainee, and I’ll bet that if CE took a day to learn how to use the blender, CE would know whether mashing Mangos is the way to go.
This may be true, but I still think that the more trainees you have, and the more non-standard experiments they begin doing, the more and more impractical this approach will become.
When I was a 3rd and 4th year grad student, I started using maybe 3 or 4 techniques that had never been done in my lab before. I would have thought my PI was an absolute lunatic if he had insisted on coming into the lab and learning each of the techniques.
At some point, I truly believe you need to let go of details that can be competently handled by the people who work with you. It’s not worth being a PI if you can’t trust your trainees with mundane stuff like blender operation.
LikeLike
May 7, 2010 at 2:06 pm
Merits aside, anon’s comment @8 makes me wonder — doesn’t this approach perpetuate the ponzi scheme aspect of the scientific enterprise?
LikeLike
May 7, 2010 at 2:23 pm
I should have said “lab time” or “experimental time” or some such qualifier. I tend to assume my audience knows what I think the PI’s job is really about but I should have elaborated.
Yes, you should have. It an an academic conceit to blather on about getting good people and doing good science, just like the three-legged stool of research-teaching-service is an academic conceit. It is demonstrably false, but nearly universally ignored. There should be an enormous banner over the door to the new employee orientation room at academic medical research centers that says:
ASSISTANT PROFESSOR IS A FUNDRAISING JOB
In the medium and long term, [getting good people, training them properly] is the only way to get and maintain funding.
I disagree. I know a stack of big-name famous labs where the PI has no role in training and the quality of the people is widely variable. Back in day I did a postdoc in a very famous (still) and productive (still) lab where management consisted of: “Here’s your bench. Here’s a huge pile of cash. Do some research (preferably but not necessarily on this particular topic) and write up your results.” The PI’s role in the process with respect to trainees was:
1) course-correction: making sure the staff stayed on topic and didn’t wander off into too many unproductive dark corners.
2) communicating the results of the as-yet unpublished literature (this is why the guy spent 20 weeks a year at meetings).
The per-person and per-dollar productivity of this lab (and others like it) has never been particularly high, but enough big results get generated by a few of the workers to keep the operation humming.
LikeLike
May 7, 2010 at 2:23 pm
As a new PI in a small non-profit research institution, I think CPP’s comments are right on about the PI being involved directly in a small lab. I have only one tech right now and I do consider it a good use of my time to be at the bench; the tech works full time at the bench on collecting preliminary data for an upcoming proposal submission while I can spend some time playing around with my own new projects. That said I probably spend less than 25% of my total time at the bench right now.
LikeLike
May 7, 2010 at 4:21 pm
This is exactly why PI-professors who have large enough groups that they can’t be directly involved in hands-on research, are NOT real scientists. They are managers of personnel and resources, and lab directors. Which is a totally different job function from being a real scientist. Let’s stop pretending they are still real scientists – stop giving them all the intellectual credit for the work that is done by their underlings and which translates to the PI becoming more and more ‘famous’ and respected as a scientist (which he has long ago ceased to be), while the real scientists who did the work remain pretty much ignored.
LikeLike
May 7, 2010 at 4:33 pm
As a postdoc, I’m finding it very difficult to let go of control in colloborating projects. I tend to do the work myself than delegate it, even though I could use the extra time that comes with delegating stuff. Looks like its a conscious choice one has to make as a PI.
LikeLike
May 7, 2010 at 4:35 pm
AHAHAHAHAHAHAHAHAHAHAH!!!!!!!!!!!!!!!!!!! What took so long?
LikeLike
May 7, 2010 at 4:47 pm
Wait, so my advisor isn’t a real scientist? Bullshit, you’re fucking retarded.
LikeLike
May 7, 2010 at 5:32 pm
Wait, if one of the responsibilities of the PI is training, how would that be accomplished if the PI has no idea of experimental details? Or do you mean “provide a training environment”. I think we should drop the charade that we “train” anyone in the strict sense of the word.
As a (not-so) new PI, I find myself often struggling with just this issue. Is my job to raise the money, recruit new people and chart new directions or to actually do some of the work myself? Of course, I still find that I can do many of things I assign to my students in half the time they take (I am a theorist, btw).
LikeLike
May 7, 2010 at 7:32 pm
Back to management theory: Where X is “subordinate acquires necessary competences,” the manager’s duty is to see that subordinates acquire the requisite competencies.
How that happens depends on circumstances. Is the manager qualified to teach the necessary skills? Are there other potential trainers? Does the manager have the bandwidth to do the training? If the manager is one of several potential trainers, which arrangement is on balance the most productive?
Technically, it’s not even the manager’s duty to make that decision, just to see that the decision gets made. In practice, “making the decision” is one of those things that the manager will “make sure happens” by simply doing the deciding.
LikeLike
May 7, 2010 at 8:05 pm
Of course they are not. When: 1) One doesn’t come up with the ideas, 2) doesn’t understand how they are realized, and 3) needs explaining WTF it all means – how can one claim being a scientist? Sure, one can be in the business of doing science but that’s not the same as be a scientist
LikeLike
May 7, 2010 at 8:13 pm
As a rule, from the moment you stop understanding how the shit is done there is a very short period of time to the moment when you have no clue why the shit is done. And why this shit and not the other. From that moment on, you are a charlatan, not a scientist. There are exceptions, of course, but they are exceptions.
LikeLike
May 7, 2010 at 8:15 pm
That rather depends on your definition of “scientist.” If you define “scientist” as someone who spends the day at a bench, then you win by definition. If you define it as “someone engaged in advancing our knowledge of the Universe” then you get a different answer.
All in all, I prefer a definition which doesn’t exclude Stephen Hawking.
By all means, however, enjoy your tautological victory this weekend — it’s looking to be a lovely one. Cheers!
LikeLike
May 7, 2010 at 9:17 pm
Sure, if Grad Student A has to do weird assay X to get a result, fine, I don’t need to know or want to know the details of how long he needs to incubate step 9.
BUT… what if highly productive and technically sophisticated Grad Student B, who has just finished setting up a very expensive piece of equipment to run highly specialized experiments upon which a significant part of your research will depend on, suddenly, well, graduates and does not train anybody on how to use the equipment or do the experiments. That’s pretty fucking scary. You better be sure that I, as the PI better learn to use this equipment and am able to perform and troubleshoot these experiments while this student is around.
LikeLike
May 7, 2010 at 9:18 pm
“someone engaged in advancing our knowledge of the Universe”
Our office accountants are all engaged in advancing our knowledge of the Universe.
LikeLike
May 7, 2010 at 9:22 pm
Dudes, I disagree with the idea that PIs in big labs who aren’t involved in the details aren’t real scientists. I worked for an academy member who, even tho he thought we could simply push the blender button and make tomato sauce out of thin air, quickly came up to speed with the details once explained. Some projects he handed people, others he supported from afar without much input until the writing stage. Plus he wrote enough grants to have 3-8 funded at any one time. He also paid well and was supportive of dealing with family issues. Smart guy, not a tool.
LikeLike
May 7, 2010 at 9:47 pm
This is a great post and those that disagree with it are idiots. Keeping up with the progress of the field, coming up with new ideas to explore, being involved in project design and data analysis, and contributing intellectually to the writing up of results…those are the things that define a scientist. The rest is just monkey-business.
LikeLike
May 7, 2010 at 10:19 pm
People, people. Is very simple- Only thee and me are scientists, and I’m not at all sure about thee!
LikeLike
May 7, 2010 at 11:44 pm
@”Chad”:Wait, so my advisor isn’t a real scientist? Bullshit, you’re fucking retarded.
Well I don’t f8cking know who your advisor is so I can’t tell if he’s a real scientist or not, obvious isn’t it. How hard is that to understand?
LikeLike
May 7, 2010 at 11:57 pm
If you want to stay at the bench, and have some security, get a job in a National Lab. I was quite happy there for 7 years or so post-PhD. The resources are amazing and there are many brilliant scientists to work with. It can be hard to keep your edge in that environment, though.
Once you have more ideas than you can successfully carry out with your own hands, you have three choices. Step away from the bench to grow your group, give up on pursuing many of your ideas, or give your ideas away to others who can run with them (just to know the answers). Which choice works for you is for you to decide. At my old lab, I saw all three in action. I chose #1, and went back to academia.
LikeLike
May 7, 2010 at 11:59 pm
I’m not sure if this is what qaz was saying but here’s where I stand:
when I’m a PI, if I spend, say, 10k of our grant money on a fancy-schmancy blender, you better believe I’m gonna make my trainees show me how it works. Not just because then I can make sure they aren’t doing anything too dumb, but also because it’s fun to learn how stuff works and (in some cases) makes you a better grant-writer to boot.
LikeLike
May 8, 2010 at 3:54 am
#21: “Of course they are not. When: 1) One doesn’t come up with the ideas, 2) doesn’t understand how they are realized, and 3) needs explaining WTF it all means – how can one claim being a scientist? Sure, one can be in the business of doing science but that’s not the same as be a scientist.”
Exactly. Yes so many are unable to grasp this simple concept.
If one actually doesn’t want to do bench work anymore, then how can you still call yourself a scientist? YOu want to be a manager, because there’s more power and influence and prestige in it. Being a manager of scientists isn’t the same thing as being a scientist. Why is it seen as an insult or some heresy to point this out? And someone here who earlier said something to the effect of “anyone who helps to advance knowledge of the universe is a scientist” (to paraphrase, in defense of claiming that a PI does not need to be doing actual research themselves anymore to still be labeled a scientist). OK, so are the department secretaries, equipment maintenence personnel, tech transfer office personnel, now scientists too?
LikeLike
May 8, 2010 at 10:50 am
In my field (and experience of academia), I don’t know any PIs who aren’t either the originator or co-originator of ideas. I also don’t know any PIs that are not involved in data analysis and interpretation. These people are claely scientists, even if they aren’t at the bench.
When I was a staff scientist, it was more common to see PIs that were only peripherally involved in the day to day science. But they were instrumental in keeping the lab well-funded, and often were heavily involved in dissemination of the group’s results and in putting data into the big picture. They also are clearly scientists.
I don’t know any PIs that “needed explaining WTF it all means.” If such people do exist, I would not consider them scientists.
LikeLike
May 8, 2010 at 12:16 pm
Those who think you have to sit at a bench all day to be a scientist are fooling themselves. If you operationally define science by sitting at a bench running the machinery, you don’t really understand what science is. Science is the scholarly pursuit of knowledge through systematic investigation and experimentation. Punching the shiny “GO” button on a fancy piece of equipment, or pipetting some shit onto a gel is not science, it’s manual labor. If someone tells you what experiment to run, and you run it, you are not a scientist, you are a day laborer. It’s the forethought, experimental design and interpretation that make you a scientist.
LikeLike
May 8, 2010 at 12:46 pm
Arguing with delusional “PIs are managers, not scientists” fuckbags is as useful as arguing with Obama “birthers”. The only people who make these claims are angry bitter losers whom life has passed by and left them on the side of the road with nothing.
LikeLike
May 8, 2010 at 3:31 pm
Not having gone the route of becoming a big time PI, I concentrated on making the fishes I studied well enough known and interesting enough that I could be second or third author on important papers about them. These papers use methods and techniques which, as they say, are beyond my pay grade. So I would agree one has to trust people who know things.
LikeLike
May 8, 2010 at 6:40 pm
Understanding something is one thing, concentrating on something is another. The PI absolutely MUST understand what is going on in the lab, down to the smallest detail, but in order to have time to devise new ideas, plan the research strategy and acquire funding he/she needs to delegate responsibilities for more mundane aspects of the group’s work.
LikeLike
May 8, 2010 at 10:29 pm
Suppose that the PI does NOT have to “understand what is going on in the lab, down to the smallest detail”. If this is the case, is the PI still *responsible* for what happens in the lab? How can you be responsible for something you don’t understand?
LikeLike
May 9, 2010 at 2:23 am
#33:”Arguing with delusional “PIs are managers, not scientists” fuckbags is as useful as arguing with Obama “birthers”. The only people who make these claims are angry bitter losers whom life has passed by and left them on the side of the road with nothing”
LOL – let’s see now: no intelligent response, just a random personal attack. Clearly touched a nerve, haven’t we! let’s see how much more predictable you can be. (eagerly awaiting more stupid incoherent ranting)
LikeLike
May 9, 2010 at 3:04 am
I largely agree with Dr. F @32 — bench labor does not a scientist make.
However, I have known more than a few older, internationally-recognized PIs who aren’t even managers, but are merely empty-suit politicians.
I think some of the complaints of the perpetually-disgruntled postdoc bloggers are based on encounters with these tools.
LikeLike
May 9, 2010 at 4:40 am
PIs are scientists who are able to practice also the science of lab management. However, there are almost always exceptions to the rules. And there are PIs who appear to be no scientists, no managers but money and public relations operators. These ones are very rarely at the lab, much less at the bench but have the opportunity to delegate both science and lab management as to have the time and creativity for scientific leadership by being in every and all committees everywhere.
Of course, the theory of “divisiveness” and “angry losers” is conveniently applied when issues at the heart of true science come up for public discussion and enlightenment.
LikeLike
May 9, 2010 at 8:30 pm
I think this argument comes down to Tarc’s assertion that you need to be at the bench to remain a real scientist. I dont think you do, but you should be versed enough to be able to watch over what people do, question them, and make sure that the processes in the lab are running appropriately. I know every step of every process in my lab. Are there people in my lab better than me at pipetting shit for qPCR? Sure. Do I still do some stuff in my lab? Sure. I run my HPLC. Do I process all the shit for my samples or make my mobile phases? No fucking way. If there is a problem with my mobile phase do I adjust it myself slightly? Sure, I am not a TOTAL asshole….. Do I do anything remotely repetitive in the lab? Nope. I can troubleshoot their results better than they can. I can tear down any machine in the lab and put it back together (within the limits of end user servicibility). I only do it rarely, but I have the ability better than anyone else….we all have our talents….
I think depending on what your lab does, you need to understand the techniques of your lab. You need to if you get a tech question when you do your Ringling Bros act in front of your colleagues at meetings in faraway fabulous destinations.
I think of it like a movie in which the evil guy sits around orders his thugs to wreak havoc on the innocents while he sits back and directs the show. In the end though when needed, he can come out and kickass better than all of them….I watch too many bad movies.
LikeLike
May 10, 2010 at 1:49 am
There is no One True Way to run a research group, but two good guidelines are:
1) Because the devil is in the details, you should always understand the details well enough to be able to scrutinize bad results…and, more importantly, good results. Nobody wants to publish a result that turns out to be too good to be true.
2) Because the devil is in the details, and because reproducibility and consistency are important, there should be some repository for detailed knowledge that lasts longer than the duration of a typical trainee’s sojourn in the group. Whether that is a staff scientist, a hands-on PI, a perma-doc, or a very, very well-maintained wiki, or whatever else, key technical details of hands-on work should not be transmitted solely through interactions between third-year grad students training first-year grad students.
Beyond that, there is no One True Way to run a group. There are many different ways, and all of them can work as long as the people involved are conscientious and diligent and devoted to maintaining consistency and integrity in the procedures. If those conditions are met, the rest will work itself out. If those conditions are not met, it really doesn’t matter whether the PI is in the office or in the lab or at conferences or whatever.
LikeLike
May 10, 2010 at 12:55 pm
Back to the original point (I’ve skimmed the comments, so hopefully I’m reading right and not repeating anyone here), I’d have to say that I agree with qaz, Drugmonkey, and Physioprof.
It’s definitely a balance, especially for a PI who’s just starting out. I’ve seen PIs who micromanage while quite successfully running a lab (at the beginning of their careers, and yes, they were work-a-holics… but who isn’t in science?). I’ve seen PIs who barely manage at all and barely know what’s going on in the lab (and the lab suffers). I’ve seen PIs who micromanage out of paranoia and end up doing all the work themselves, firing all other personnel (yes, it’s true – I couldn’t believe it myself), and failing.
Judging by what I’ve seen and experienced, it’s not bad, if one has the time, to take a few moments to get caught up on a new piece of equipment or method if it is going to become essential to a project. I think that this may have been qaz’s point? You never know, too – sometimes a person with more experience, such as a PI, will be able to see new/more/other applications for a piece of equipment or method that a person with less scientific experience simply can not envision.
But, training people underneath one’s supervision to take the reins themselves is extremely important.
LikeLike
May 10, 2010 at 1:38 pm
It is true that most established principal investigators do not have the time to carry out the experimental work themselves at the lab bench, despite the thrill that can come from seeing the research results firsthand. However, a PI should definitely be very familiar with the underlying theory and limitations of technologies that are being used by the trainees under their supervision. Moreover, the PI should always be on top of the experimental results and lead in the decision making process on how to proceed further, especially with graduate students and technicians. Many M.Sc. and Ph.D. thesis projects are unnecessarily prolonged by years due to insufficient supervision. In keeping with the old adage “the devil is in the details,” the PI with their experience and knowledge should be in the best position to judge how to lead a research enquiry.
LikeLike
May 10, 2010 at 4:24 pm
whimple:
I don’t think you want to go there, or you’d be inspecting the designs of your instruments and the source code of the software you use. Abstraction is how humans handle the problem of scaling complexity and is essential to doing almost anything nontrivial.
My favorite example is Kelly Johnson. Over the course of three decades he was in many ways the Sierra Hotel leader in high-performance aircraft design. He wasn’t, however, hands-on with fuel pumps, engine design, etc.
The same goes for the folk working on the Manhattan Project (I hope we’re not arguing that Richard Feynman wasn’t a scientist.)
Since nobody responded, I’ll repeat the question: does your definition of “scientist” exclude Stephen Hawking?
LikeLike
May 10, 2010 at 7:30 pm
D.C.Sessions: I’ll repeat the question: does your definition of “scientist” exclude Stephen Hawking?
The answer depends on what exactly Stephen Hawking is doing these days. And that part I simply don’t know about him (and I doubt that you do either). Does he still come up with his own ideas in science and then contribute significantly in testing them? Last news about him was the advice to stay away from aliens. Would voicing opinions like this be included into your definition of practicing science?
LikeLike
May 10, 2010 at 9:15 pm
He’s the Director of Research at the Department of Applied Mathematics and Theoretical Physics at the University of Cambridge. Here’s what his group is doing: http://www.damtp.cam.ac.uk/research/gr/subjects/
I have no idea whether that constitutes hands-on work or not. My experience with theoretical physicists is they often rely on a combination of math and observational data rather than experimentation. Does that make them not scientists?
LikeLike
May 10, 2010 at 11:41 pm
People who paint themselves into rhetorical corners are hilarious.
LikeLike
May 11, 2010 at 4:55 pm
I’ve known PIs who were the only ones in the world who could get a piece of equipment to sing arias and who had folk sending them samples to analyze and were on hundreds of papers and had mucho grants. Their students learned benchwork from them, but never surpassed them. I’ve known others that you let into a lab only if all the water and electricity had been turned off, the chemicals were stored away and they were handcuffed so they could not touch anything, BUT they had a unique gift to figure out what work had to be done, to write it up and get funding and to analyze the results. Their students passed the lab knowledge on to the newbies. And I’ve known PIs who were totally useless. YMMV
LikeLike