How do I grad school: management skills edition
August 20, 2013 3:46 PM   Subscribe

I'm a grad student in the sciences, going into my (mumble) year. I've been assigned an incoming grad student to show the ropes. Great! What the hell do I do?

Pretty much what it says above the fold. I don't really have any point of reference for how to do this, because I joined a lab that was just getting started during my first year; I didn't get trained by senior students because there were no senior students. There were a billion things that needed setting up, and I figured out how to do them either by reading and trying things on my own and asking my adviser when I was at an absolute dead end. Now that the lab's been around for a while, everything's kind of settled into a few established projects that are chugging along in their own way, and there's a bit more structure to it all, which includes having senior students mentor newer ones.

We're both chemists, but it's an instrument/methods development lab, which is super-fun and all, but it means that I can't really direct him to make up a batch of precursor reagents or run samples on an off-the-shelf instrument or any of the other tasks that are standard to undergrad chemistry training. There's a fair amount of mechanics and electronics and such that I'll need to teach him.

What we've done so far: the first day, he just sort of observed and I narrated what I was doing and why as I tried to fix my instrument for the nth time, and that seemed like a good introduction to basic principles. Also, he got to witness an argument a spirited discussion between my adviser and I about what was wrong this time (my adviser's a great guy and we actually get along pretty well, but this tends to be how we troubleshoot.) Today I set him up on Labview putting together a simple local-maxima-finding routine to add to the data collection program that I put together (something that I've been meaning to do for a while, but it's been low-priority compared to all of the other stuff.) I worry that that's not an appropriate level of difficulty, and that I need more ways to get him involved with the running of the instrument (which at this point is really more the fixing of the instrument.)

Also, at the risk of sounding like an insecure jerk, I'm not really someone who went into research science despite an outgoing personality and winning social skills, you know? I'm pretty oriented towards working by myself, and I don't have much middle ground between being silent and rambling a blue streak. I worry about both going so slowly that I'll insult his intelligence or running through everything so quickly that it winds up being overwhelming; I worry both about being too hands-off and being overbearing; I just kind of am not sure what balance to strike in general?

So: what kind of things make for a good introduction to a laboratory? How do you design tasks that are appropriate for someone who's new to a project? If you manage people in the sciences, or not in the sciences, what are the important things you've learned about that? If you had a more traditional introduction to grad school than I did, working mainly with a more senior student, what did they do that you liked/didn't like?
posted by kagredon to Work & Money (9 answers total) 7 users marked this as a favorite
Sit him down and say, "I will not be impressed by how few questions you ask. So go ahead and ask them, even if you think you know the answer." And mean it.
posted by Etrigan at 3:53 PM on August 20, 2013 [2 favorites]

I'd find him something to do. Give him some slice of the project to do and allow/expect him to ask you lots and lots of questions. It will take him 10x longer than you to get it done (5x if he's very very good), but he'll learn important skills and be better next time.
posted by Betelgeuse at 3:54 PM on August 20, 2013 [2 favorites]

Best answer: And make sure the initial slice is small. The work I have my students do for semester projects are usually something I could do in a few days. They're undergrads, so I'd expect more from grad students, but still.
posted by Betelgeuse at 3:56 PM on August 20, 2013

Best answer: I recommend Teach What You Know: A practical guide to knowledge transfer using peer mentoring. It is aimed at helping engineers transfer knowledge to other engineers -- most engineers I know don't have winning social skills either :).
posted by elmay at 4:07 PM on August 20, 2013 [1 favorite]

Best answer: Oh god, how painful.

'Show the ropes' is incredibly vague. Can you get some clarification from your supervisor as to exactly what he should be learning? At least then you've got some guidelines for what he needs to be included in. He probably doesn't need to watch you order parts, for example. Also, I go completely nuts if I've got someone hovering near me all day. So he's going to need something to do during down times.

I assume the priority is to get him working on the instruments independently? Do you have procedures or manuals? If yes, he can read them. If no, he can start to write them. Give him a pad and pencil, and get him to take notes as you work. He can then collate them into a troubleshooting manual, or even just his own personal how-to. Also, what's he studying? Does he have a lit review that needs doing? Obviously, if you've got stuff he can do that's relevant to what he's going to need to know, then have him do it. But I don't think there's anything wrong with telling him to entertain himself for an hour or two.

And make sure you get away from him for lunch. Just go hide somewhere out of the way.
posted by kjs4 at 5:02 PM on August 20, 2013

But do be nice. And do encourage questions. Lots of questions. Mostly because he'll learn faster, but also because he'll be less likely to break an instrument if he asks before he does something stupid.
posted by kjs4 at 5:07 PM on August 20, 2013 [1 favorite]

Lots of realistic advice about which kinds of work they should be focusing on in order to graduate would probably be more useful than anything technical, honestly.
posted by no regrets, coyote at 5:17 PM on August 20, 2013

Best answer: Today I set him up on Labview putting together a simple local-maxima-finding routine to add to the data collection program that I put together

This is a good start.

First, I try to identify tasks which a student can do more or less work on on their own. Simple ones are fine to start with, but don't be afraid of a little chucking into the deep end if the student is capable of tasks you give him. These tasks can start grunt work you haven't had time to do, but should evolve to bigger projects. My first grad project was porting the lab analysis software from fortran in unix to c in dos/windows. Later, I was given the task of developing FT and peak-finder/spectral analyzer code. Give him a walk through, lay out what you expect, give him an approach to the problem and let him go. Be available and able to answer his questions. Programming or data analysis tasks are great as it's difficult to cause lasting harm (take backups) and they tend to be purpose-directed or finite, and therefore solvable.

On the other hand, if he's struggling as you give him more complex thing to do, back-off and recalibrate. Maybe you need to show him something again, or maybe he's missing something fundamental.

Secondly, check that he has his readings. He and your supervisor should be doing this already, but make certain that he's got all of the background reading that he needs to understand to do the labwork. Profs sometimes forget the more practical works.

Thirdly, pull him into your work, slowly, at a pace that works for you and him. You've got your own stuff to do, but he should be sitting beside you when you shim the magnets/adjust the beam optics/whatever. He needs to be shown how the maintenance cycles work on the instruments, so when something breaks, you need walk him through the basics. I use a show, supervise, do cycle: I show how to do it, I get the student to do it with me present, then the student does it on their own and I check to make certain everything is ok. The last step is quite important, and you need to be able to judge when he's ready for it.

If desirable/possible, I use a validation exercises for student introductions to methods: do a repeat set of a known standard, paired-replicates, spike addition recovery, whatever. Repeat measurements until the student can reliably meet an RSD that's acceptable/within historical or QC limits for some statistically significant set of data. The exact target you use is going to be specif to your instrument and methods, but showing the new guy how to numerically understand performance/success with the instrument gives him a tangible goal to work towards.
posted by bonehead at 8:50 PM on August 20, 2013 [1 favorite]

Best answer: Secondly, check that he has his readings. He and your supervisor should be doing this already, but make certain that he's got all of the background reading that he needs to understand to do the labwork. Profs sometimes forget the more practical works.

When I was in the neophyte's situation 20 years ago, a sheaf of white papers on the lab research would have been nice. Someone to discuss them with as I was draining the compressor and degassing the vacuum chamber and scrubbing the e-beam gun would have been heaven.

It's very easy to forget the learning curve that you've gone through... please be patient with your newbie.

(Oh, and no sequence structures or locals in the LabVIEW. :)
posted by underflow at 10:40 PM on August 20, 2013

« Older Maybe we need to upgrade the installer too?   |   How do I load a file of stamps into Adobe Acrobat... Newer »
This thread is closed to new comments.