Archive for the 'Postdoctoral Training' category
This repost is via special request from some n00b Assistant Professor who has apparently lost access to Google.
It was originally posted 25 Aug, 2008.
The comments following a recent post touched on the newly independent investigator dilemma of who to hire first: A postdoctoral fellow or a technician? We'll leave aside the best answer ("both") as impractical because, as Professor in Training noted,
I only have enough money to pay ONE postdoc's salary for 18 months ... or ONE tech ... that's it. While that would be great for me, that's certainly not enough time for a postdoc to get more than one study done (in my field probably only 75% of a study). Is it even advisable to employ a postdoc for such a short time with no great certainty of being able to pay their salary beyond that point?
To define terms just a little bit for those less steeped in the biz, I covered the job category of "technician" here:
A "technician" in the biomedical sciences is an employee of the laboratory (well, actually of the University) who is not "in training" (such as graduate students and postdoctoral fellows) and does not (usually) have a terminal doctoral degree. (For example long-term PhD scientist employees who are too far along to really be "postdocs" and are not PIs are not really "techs".) Most typically the tech has a bachelor's degree in a scientific major and a few will have advanced credentials such as Veterinarian Tech specialties or subject-based Master's degrees. It is not unusual for the tech to have continued her education while working in the laboratory by taking advantage of University educational repayment policies.
One of the most important parts for today's discussion is that the technician can be viewed, non-pejoratively, as 100% an employee working "for" the lab head or Principal Investigator (PI). Someone who is expected to do what is asked at all times with the goals of furthering the lab agenda. In this employment relationship the PI is unequivocally understood to be "the boss".
A postdoctoral fellow/trainee is a person who has acquired a terminal doctoral degree (Ph.D., D.V.M, D.D.S., M.D.) and is working under the supervision of an independent investigator. Here one key difference is that the postdoc is in a dual role, the balance of which is debated. The postdoc is considered a "trainee" in the sense that s/he is working in part for her (his) own benefit, to acquire skills and tools that will be required to obtain and launch an independent research career. This notion implies a degree of independence from the PI, an ability to work on stuff other than what the PI has "told" her (him) to do, possibly to work on stuff that is only of benefit to the postdoc (and not the lab or PI directly), etc. I happen to believe that the postdoc also has a responsibility to help the lab and, in essence, to do the job for which s/he was hired and to make all data generated accessible to the PI...but not every postdoc agrees with that. In truth it is also likely that some PIs either explicitly or implicitly think that postdocs are basically indistinguishable from technicians when it comes to the employment relationship. So there's a spectrum.
Okay, so why do I think a tech is more important to secure for a brand new PI?
That isn't your job anymore. Scutwork. Tedium. The stuff that is absolutely necessary to the running of your lab which is not particularly demanding in intellect and may be incredibly time-consuming. Yes, you specialized in this as a lowly graduate student and took pride in the fact that you were able to do this work as a postdoc while still doing more high-falutin' intellectual labor. Postdocs around you who couldn't find their behind with both hands when it came to the basic work were to be derided. Fine. But this is not your job anymore!!! My position is that the more time-consuming scutwork you can get off your plate the better. Since nobody likes scutwork, it is far better to rely on an employee who is paid to do a job, can be readily fired and replaced, for this sort of thing.
Your first deceptively hard question as a new PI is that of determining which tasks in the lab really do not require your input, after basic training and given that you will continue to supervise and troubleshoot. I say deceptively hard because my experience in talking with some fairly advanced postdocs and even junior faculty is that they have not really thought about this question. They get stuck in the usual traps. "It is more work than it is worth to train somebody to do this." "I only trust my own work/data/analyses." "My hands are the best." "This is too important to screwup." Etc.
All true. Being a PI takes a big leap of faith in the work of others. This is, in my view, part of the deal. For the huge increase in scientific terrain you are able to cover as a PI directing the efforts of other scientists, you are accepting the risk that someone else is not as good as you*. So get over this. Your job is to learn how to set up your management style such that you can tolerate human frailty and still make excellent progress on what interests you.
Progress and Work Ethic. Management of personnel is one of the hardest things for new PIs to learn. After all, we were motivated self-starters so we can't really understand why everyone else would bother to be doing this stuff if they weren't self-wackaloon-motivated. Sadly, not every one is just like you (ibid), new PI! Which means that you may have to evaluate an employees work effort and apply some judicious boots to the posterior. Perhaps even with threats of dismissal and actual dismissal for poor performance. It is very much easier to do this with a technician who is supposed to be working 100% at your behest.
Stability. In most cases postdocs will be transient visitors to your lab, lasting 3-5 years at best. So sure, a good one may get you through tenure. But postdocs can and do leave for all kinds of reasons. Their interests and relative focus on your stuff necessarily changes as time elapses. The tech on the other hand, can be a more or less career employee in whom your investment over time continues to pay you back over intervals of a decade or more.
Availability. Unless you are very lucky or very HawtStuff, recruiting a good postdoc to your laboratory is far from given. I've seen this from all ends, as a postdoc, peer of postdocs, mentor to postdocs , as a PI seeking fellows and as a peer to other junior faculty looking to hire. I've seen situations in which the fellow (or grad student) was very focused on joining the lab of the local BigCheez and quickly evolved to be working most closely with local junior investigators because the fit was so obviously better. The bottom line is that for a postdoc the prospect of joining a starting Assistant Professor's lab is very much of career concern.
An additional concern with availability is that it is not unusual for postdocs and PIs to come to arrangements far in advance of the actual start date. Very anecdotally, I'd say the better the candidate, the more likely this is the case. So a junior PI who manages to recruit and lock-in a highly promising candidate may have to accept that she will not be arriving in the lab for a year. That's a big hit, especially if that is the first year in which teaching loads have been reduced. Technicians are typically hired with a fairly short lead time on the order of weeks at worst.
Data Stream Strategy for the Long Haul. This one verges dangerously close to the argument over being an investigator who operates on the cutting edge at all times versus the small-town grocer. So YMMV. If, however, you have an aspect of your research program which can putter along with relatively little input from you, is of lasting importance to your scientific goals and interests and, most importantly, can support a steady stream of bread-and-butter publications I recommend getting this going. It is not necessary that the tech does all of the work up to publication-quality figures, of course, mainly that s/he is able to generate good quality and interesting data at your direction. Or the tech is capable of doing most of the work with you sailing in for the essential parts.
It is all very well and good to shoot for GlamourPubs. If you manage to get them, you are set. I get this. Not getting them is, however, excusable. In most environments, meaning that even if not in your specific department you can get a job elsewhere. Perhaps one theoretical tier down, but still a research-focused job. What is not optional is publishing somewhere, anywhere**. When it comes to most decisions that matter, tenure and review of prior progress when it comes to grant review, 0-fer is not excusable. Published articles can be debated on their merits with respect to actual impact, importance, brilliance, what have you. A lack of publication can not be debated or defended***. Admittedly the postdoc who is half-decent has a greater possibility of getting all the way to a submittable manuscript. But the bread-n-butter tech is near guaranteed to make sure the data are available to writeup when you are feeling the publication pinch.
Final Thought. It basically comes down to risk management from my perspective. If you can get a very good, hardworking postdoc right away the choice is pretty clear in opting for the postdoc. I am quite pessimistic, however, that new PIs can pull that off. When it comes down to a postdoc who will not show up for 12 months, a postdoc who is lazy, distracted or really focused on interests that are not sufficiently in line with yours...well, the technician wins every time.
Update: One thing I forgot to mention originally but was reminded of by the first comment. The chances of getting a technician working for you for free are next to zero. It is possible to get a postdoc for free, however. A postdoc may come with their own fellowship (there are many international-study type fellowships where a country funds individual fellows to go abroad, for example) or you may be able to secure a slot on a local institutional training grant for your postdoc. This is not a guarantee but these situations are considerably more likely than getting a technician working for you but paid for by some other source of support.
*almost by definition, the fact that you are a PI now means the smart money bets you were a better-than-average postdoc. Very likely you had more motivation, intellectual curiosity and, yes, better experimental hands than the average bear. Which means that on average, postdocs that come into your lab are going to fall short of the standard of you! (Yes, even accounting for an inflated view of self.) Deal.
**"anywhere" means "peer reviewed" and is environment specific. Whatever your field considers the supposedly lowest denominator or a minimally respectable "dump journal" or whatever.
***usually. I could tell you some stories. But really, make it easy on yourself and publish something already!
We all know about the oversupply problem in academic science wherein we are minting new PhDs faster than we create / open faculty jobs to house them.
Opinions vary on what is the proper ratio. All the way from "every PHD that wants a hard-money traditional Professorial job should get one" to "who cares if it is 1% or even 0.001%, we're getting the best through extreme competition!"
(I fall in between there. Somewhere.)
How would we detect it, however, if we've made things so dismal that too many trainees are exiting before even competing for a job?
One way might be if a job opportunity received very few qualified applicants.
Another way might be if a flurry of postdoctoral solicitations in your subfield appeared. I think that harder flogging of the advert space suggests a decline in filling slots via the usual non-public recruiting mechanisms.
I am hearing / seeing both of these things.
This has to be the longest winded and most indirect way of putting it.
My interpretation (I could be wrong) is that yes, you can use R-mech research grant funds to send your postdocs to meetings, give seminars and do other postdoctoral training activities that are not directly related to the goals of the research grant.
Continue Reading »
Is a lab that has 12 or more concurrent postdocs really "training" them?
Is this in and of itself evidence that this is the scientific workforce?
If, say, three-quarters of them ended up in faculty appointments would this change the equation?
There was a discussion on Twitter regarding postdoctoral compensation that veered off into the assertion that it is "easy" for PIs to give a 20% bump to their postdocs if they wanted to. It eventually incorporated some allegation that particular subspecialties (like computer/informatics jockeys) required a bonus bit of pay above normal for postdocs. Eventually it became clear to me that there are both some postdocs and indeed junior faculty that basically think they are able to just pay their postdocs whatever they want if they happen to have sufficient grant or other money to do so.
In one case I tried gently to raise the spectre of unequal treatment, bias and discrimination in wage compensation. I do this not just because of the general observation that in far too many workplaces women just happen to get paid less. I do this also because I have been around training environments where the direct-experience pain of unequal postdoc pay ("because DewdDoc has a family to support and you, WomanDawk do not") was made clear to me. While I have not heard of anybody pulling that crap in my current department...let's just say I'm not the trusting type.
I have tried to adhere to postdoctoral compensation policies that were external to my own (inevitably biased) self in my actions as a lab head. In brief, I have stuck to the NRSA guidelines. At one point that meant that my postdocs were being paid better than the mean in my department. I think we are now all doing better but my sense is that my institution only requires the starting pay to be at NRSA minimum and doesn't enforce the yearly steps.
I tried the subtle way but it is not, apparently, working.
PhysioProf, growing tired of the shenanigans got shouty.
Hint to all of you: THERE ARE FUCKEN FEDERAL COST PRINCIPLES THAT YOU ARE OBLIGATED TO FOLLOW BY FEDERAL LAWS AND REGULATIONS. THESE PRINCIPLES DETERMINE THE ANSWER TO DOUCHEMONKEY'S QUESTION.
Head on over to the NIH Grants Policy Statement where it talks about Salaries. Emphasis added for the slower readers.
Allowable. Compensation for personal services covers all amounts, including fringe benefits, paid currently or accrued by the organization for employee services rendered to the grant-supported project. Compensation costs are allowable to the extent that they are reasonable, conform to the established policy of the organization consistently applied regardless of the source of funds, and reasonably reflect the percentage of time actually devoted to the NIH-funded project.
and on "bonuses", a mechanism used to supplement postdocs that I have heard of.
Allowable as part of a total compensation package, provided such payments are reasonable and are made according to a formally established policy of the grantee that is consistently applied regardless of the source of funds.
The grants policy has this to say about Fraud and Waste:
Examples of fraud, waste, and abuse that should be reported include, but are not limited to, embezzlement, misuse, or misappropriation of grant funds or property, and false statements, whether by organizations or individuals. Other examples include theft of grant funds for personal use; using funds for non-grant-related purposes; theft of federally owned property or property acquired or leased under a grant; charging the Federal government for the services of “ghost” individuals; charging inflated building rental fees for a building owned by the grantee; submitting false financial reports; and submitting false financial data in bids submitted to the grantee (for eventual payment under the grant).
The Federal government may pursue administrative, civil, or criminal action under a variety of statutes relating to fraud and making false statement or claims.
It doesn't precisely say that overpaying staff outside of what is "reasonable" and that "conform to the established policy of the organization" is the same as "theft for personal use" and they talk about "inflated...fees" only in the context of building rental. But c'mon.
It doesn't take a neurobrainrocketscientistsurgeon to deduce that an individual NIH-grant-holding PI has some obligation to pay their staff only what is consistent with institutional policies external to their own personal preferences.
I take no position whatever on what institutional postdoc compensation policies should be in various locations in this fair land. Ok, that's a lie, I think NRSA scale should be a minimum. But if there need to be local adjustments, no problem. If there needs to be institutional recognition that some job types (like a informatics / computational biology postdoc) require a consistent bonus to compete with the local killer-app startup, fine.
But it is VERY clear to me that individual PIs do not just get to willy-nilly decide what to pay their trainees without any reference to what is "reasonable" and what has been made "established policy of the organization".
I got tough-love comments only about twice in my postgraduate academic training prior to becoming a PI.
Once was from a friend in a far different graduate program when I was in the depths of writing my dissertation. This one is in my top 5 life touch-stone conversations I would guess. Basically it boiled down to "If you find all of this so unpleasant, get the hell off your ass and do something else. I mean seriously, wtf is wrong with you?"
The second tough-love convo came from a mentor who had just seriously pulled the rug out from under me and I had disappeared (essentially) from work for several weeks. That one was more----nuanced in effect on me. But in the end s/he was right. Shitty things happen but dude, it's your fricken career and your fricken life. So sack up and get back to work.
On reflection, I didn't have anywhere near enough people doing this to me throughout my career. Two was not by any means as many as I could have benefitted from.
As a mentor, I rarely start these conversations. I doubt I've had a trainee that doesn't get at least one such conversation from me that could qualify, but I find them unpleasant and avoid them.
No doubt just as my mentors did.
They are hard conversations to initiate and hard ones to receive.
Even on the Internet from some anonymous blowhard.
I got it mostly out of my system on Twitter but there IS a lesson in this idiocy from today. Do not EVER assume that just because [insert some assumption you made up inside your own head] that your postdoctoral mentor is going to continue to support you beyond some timepoint agreed upon explicitly. That time point may be established when you first join the lab, that time point may be adjusted. That timepoint may depend on forces outside of either party's control and change despite your best mutual intentions.
Do not assume!
@andpru Yes, but usually you'd expect the PI to pick up the salary after the fellowship money runs out.
— D List Scientist (@DListScientist) September 17, 2014
No. There is no "usually". There is no "expect".
What there is, postdocs, is discussion.
Money to support you does not grow on trees, as much as one might like to fantasize about it:
@GidRothschild not sure, 3-4. But productive and fundable. If it were me, I'd keep them for same number of yrs as they had the free salary.
— D List Scientist (@DListScientist) September 17, 2014
Ok, sure. There are people who have unlimited funds to tap. Sounds fine to stipulate. Most people in the biomedical sciences do not. Most of them support trainees on research grant funds, if no fellowships are available. And those grant funds cannot be just stockpiled for three years waiting for the trainee to rotate off the training grant. The grant award, for the most part, is only somewhat guaranteed for 5 years. So assuming the postdoc on a fellowship comes in to work on a given project, its clock will be ticking during the fellowship. Three years of fellowship and three years of the "expected" payback and....uh-oh. It doesn't add up even under the perfect scenario.
so even though it is essentially impossible to make this guarantee to someone on a training fellowship..
— Gideon Rothschild (@GidRothschild) September 17, 2014
The PI must take "financial responsibility". For how long?
— Gideon Rothschild (@GidRothschild) September 17, 2014
Oh, a "year or two"? Is that all? Well I suppose that is better than the equal time DLister is calling for.
and why should the PI take this responsibility for postdoctoral stints past the fellowship duration?
— Gideon Rothschild (@GidRothschild) September 17, 2014
This may be correct for certain subfields. It may take you 5 years of postdoctoral work to get a paper. But 1) that's your choice and 2) you and the mentor should have had a decent discussion about this stuff before you joined the lab. You should have a mutual plan in place. Don't depend on some expectation that the PI will share your view or even have the capability of fulfilling his or her best intentions three years down the line.
Talk early, talk often.
And work as if you expect not to have a job in that lab past the most obvious end point. I.e., your fellowship end, the end of the grant that is funding you or the end of the "well I think it should be about a three year postdoc" conversation you had with the PI at the start.
Naturally at this point the conversation turns ad hominem and I am accused of being a horrible bastard of a PI.
@drugmonkeyblog You're a heartless science republican.
— D List Scientist (@DListScientist) September 17, 2014
That may be but I have explicit conversations with my postdocs all the time. About my view of their training arc when they start in my lab (3 years is the target, with some description of average duration being longer), about sources of funding (on starting with me, and as we go along), about upcoming funding cliffs (I've weathered more than one and less than 5 imminent lab-is-closing funding scenarios in my time) and about what results need to happen with which grant reviews along anticipated timelines for me to continue to keep them in the lab. I work to structure my lab so that no postdoc goes three years with no first-author pub in sight- that would be an abject failure within the sphere I operate. Total dereliction of my duty and/or complete failure to produce on their part. We have so far avoided this fate and I plan to continue this path. But things happen.
To recap: Don't. Bloody. Expect. Or. Assume. That. Your. PI. Can. Float. You. After. Your. Fellowship. Ends!
More from Richard Harris at NPR:
That's because if you want a career in academia, it's almost essential as a postdoc to make a splashy discovery and get the findings published in a top scientific journal. Hubbard-Lucey is working on an experiment that she hopes will be her ticket to a professorship — or at least to an interview for an academic job.
Whether she succeeds or not, she's part of a shadow workforce made up of highly qualified scientists who work long hours for comparatively little pay, considering their level of education: about $40,000 a year.
Potnia Theron wishes to discuss this last assertion: NPR story on Postdocs: what is your salary? edition.
And, I am glad she found a position in NYC. I am sure she loves The City. But $47K is a lot more than median salary in the United Sates right now. Maybe its not enough to live in NYC, but it is elsewhere.
I have an older post for your consideration of what trainee salaries look like compared to when I was a wee trainee. My conclusion that scientific trainees enjoy a 30% or more bonus in inflation-constant dollars over my day was not of any comfort to the Millennial types, apparently. So good luck with your point, Potnia.
Anyway, you'll be happy to know the subject of the NPR piece came out okay in the end:
That first conversation took place in May. Later in the summer, while Hubbard-Lucey was still working on her scientific paper, she heard about a job where she could make good use of her Ph.D. She wouldn't be running a lab or working in academia. But she would be advancing cancer research at a nonprofit institute. She got the job. And now, she says, she's happy with the new path she's chosen.
Way to blow the lede, Richard Harris. Way to blow the lede.
from a Twitt:
@drugmonkeyblog Actually PI #1 is passing around drafts of upcoming R01 renewal. So there's exposure, but not formal "training".
— Kate Stafford (@kastacholamine) August 29, 2014
Let me explain something to you trainees. You are not undergraduate students anymore. You are not given a syllabus, quizzes and office-hour responses to "Is this going to be on the test" or "What do I need to know".
When your PI gives you a draft of a grant or a manuscript for you to read and provide feedback, this is not ONLY about asking for your help. This is about training you in how this person accomplishes these tasks, what manuscripts look like in nascent form, how a grant should be structured and how you incrementally improve an academic work.
The PI can lead you to water but it is not her job to force you to drink. It is YOUR JOB to drink the water.
Exposure is training.
Another one of the twitts identified a problem I had in writing papers as a postdoc. It boils down to the fear of showing your PI something that is less than perfect lest she think that you are a fool, incompetent and nowhere near the scientist-prospect that you hoped was her impression of you. I used to delay and delay showing anything to my PI until it was looking really good.
Let me tell you a little something. PIs do not lose respect for trainees for sending them crappy drafts. At the worst, they shake their heads ruefully over the shitty training you received in your last stop. Mostly, they just saddle up to train you how to write a paper their way.
They lose respect over other things. A lack of any sign of a manuscript. You can say you are "working on it" but the PI has no concrete way to distinguish the fact you are in Draft XXVII of a master work from the scofflaw who hasn't done much more than write a title page into a Word doc. So show them something.
Another thing that PIs lose respect for trainees over is a failure to make changes in response to what the PI has said or shown them. This is key. You are being trained. If a PI tells you to do something, bloody well DO IT. Don't spend weeks bitching to your spouse, fellow trainees and the Internet about what a taskmaster the your PI is. Just write, edit, change, fix. DO IT.
There are ways to really get on the PI's good side. For example when the draft is on the PI's desk for editing/review? There is no reason you can't also be working on it. And updating your PI on your new drafts.
Write more. I think Comradde PhysioProffe had a blog post or extensive comment on this some time ago in a prior discussion of the topic. Trainees are often blocked from writing because they are thinking to themselves how to be as
lazy efficient as possible. "I'm not sure what she wants here so I need clarification before I write a whole bunch of stuff". Or "Last time I wrote four pages and she didn't use any of it in the manuscript!". PP's point was that sometimes you have to write something out to see for yourselves that it is the wrong direction to go in. It is not wasted effort, it is part of the process. The science communicator types preach on about being willing to "kill your babies". I believe this is similar. Do note, however, that often enough some major passage that you decide to leave out of the present manuscript comes back as useful material for the next manuscript (or grant or review article). So writing is rarely a total waste in this business.