Tuesday, November 13, 2007

On failure.

This post was inspired by FSP's post of the same title, about the difficulties of spotting future researchers at the early stages.

Basically, FSP says it's hard. I say it's not. But I'm also the kind of person who would rather have no one in my lab than have the wrong people.

Here's how I describe the qualities of good researchers. I think these qualities are evident good researchers of all ages, even pre-PhD (note the sarcasm).

1. Good bullshit detector. Critical thinking. Objective thinking. Unwillingness to take anything on faith.

How to spot it: Journal clubs. If a student at the undergraduate level can't stand up and critically assess a paper, it's unlikely they will suddenly acquire that ability in grad school. If you don't think undergraduates can handle journal club, think again. They can and should be reading and discussing primary literature.

2. Creativity. Someone who is looking for what hasn't been done, and how it might be done in the future.

How to spot it: Do they ask good questions in class? In lab meeting?

Yes, I think undergraduates should volunteer in labs or work there in the summer at least once before applying to grad school. I would NEVER take a grad student who had not worked in a lab at least for 1 summer before applying. EVER.

3. Spine. You want someone who has their own opinions, will stick up for themselves, and will persist through the usual level of obstacles.

The last thing you want is someone who is book-smart but caves at the slightest setback or panics when things don't match exactly with the textbooks/journal articles.

How to spot it: This is the kind of thing you can tell from two tests if you know what you're looking for.

First, the interview. Ask this person if they've had any obstacles in their life. It could be anything. It will be illuminating.

The best student I ever had came from a family that ran their own business. She had a fantastic work ethic, and a fabulous no-nonsense attitude about everything.

More recently I had a prep-school student who knew the value of hard work through her sports training. She wouldn't take any crap from anybody, and she was always willing to speak her mind.

Second is in the lab. I've written before about how I think it's absolutely necessary to give undergraduates real projects, not pre-cooked kits like in lab classes (I think lab classes are worthless, and have blogged about it before).

Real projects are the best. This will tell you how they handle problem-solving and give them a real taste of what grad school will be like. You might not expect them to actually be able to solve any problems on their own at this stage, but watch for their reaction to the inevitable setbacks.

If they want to do something only once and then move on, beware! If they want to quit after 1 try that doesn't work, there's your answer! And theirs, too. The ones who can't handle failure cannot handle research, and they usually figure this out once they've had a taste of it.

4. Street smarts. By this I mean, someone who has common sense, who already has basic study skills, is at least somewhat organized, pragmatic, and not likely to get hung up on doing something one way when there are a variety of tools within easy reach.

How to spot it: Do they take notes? I require my students to take notes. That is non-optional. If they are not okay with that, they don't work with me. And I check their notebooks. The good ones grasp the concept that they will not remember everything. Even better are the ones who care that their notebook is the only record I have of what they did. The notebook tells you a helluva lot.

Are they willing to go look things up if they don't know them? One student really impressed me with this. We had a long discussion about how to do a calculation, and my way just was not working for her. The next day she came back and told me she discussed it with one of her instructors and came up with another way that would also get the job done. Bravo! I say. This is exactly what you're looking for in a researcher.

I also file a certain amount of maturity in this category. Do they keep a calendar? Are they always late and always making excuses?

Experiments are not forgiving of sloppy time management, and you shouldn't be, either.

5. Attention span. Most of the people I know who left research quit because they got bored easily. They enjoy the constant flow of new ideas that they get in patent law and science journalism. They did not want to work on the same project, or aspects thereof, for the rest of their adult lives.

How to spot it: Extreme cases are obvious when you've had a student in the lab for the summer. Here again, as with the question of 'spine', they will usually self-select when given the experience of having a project and being told to work on it exclusively for 3 whole months (!).

Less extreme cases are harder to spot, but easier to treat. Sometimes they don't know it themselves until they're mostly done with graduate school and thoroughly sick of their thesis project. Vacations help. Meetings and positive feedback help. If they get through all of that and decide to stay in research, usually finding the right postdoc project will cure it. Again, most will self-select after grad school since they've had the experience and a glimpse of the road ahead: more long-term, delayed gratification projects.

6. Big picture, little picture. The best researchers understand that sometimes you can fudge it, and sometimes you can't. It's like the difference between stir fry and baking (respectively).

Ideally you want people who can grasp the big picture, but still pay attention to the little picture when it matters. You want someone who is meticulous, but you don't want the fantastic technician who doesn't see that she needs to be reading several journal articles a week now that she's in grad school.

How to spot it: This is the tricky one. I myself am not meticulous by nature but rather by training. I've worked with some who are meticulous by nature but who learned to appreciate the big picture during grad school.

The most successful researchers I know are big picture people who take care of the forest and hire little picture people to take care of the trees. Unfortunately most graduate schools are selecting on criteria (like grades and GRE scores) that have more to do with meticulously trimming the trees and nothing to do with locating the forest.

7. Politicians. I wasn't going to include this, but on second thought I think I should.

If you want your students to go on and be successful in academia, better to pick the ones who are charismatic, good at persuading others to do their work for them, good at self-promotion, and unlikely to burn bridges. You know the type I mean. And yes, you can spot them as undergraduates.

Labels: , ,


At 7:44 PM, Anonymous Anonymous said...

It pains me to say this, but I think the real-world value of scientific creativity today is negative, particularly in the current NIH funding climate in the United States. What sells is more of the same, crank out the data and get (yet another) paper that only marginally advances progress. Good luck to you getting funded for something original, or translational, or heaven forbid, both. This really holds up in tenure/promotion decisions. The only things that actually count are #1: dollars brought in, and #2: number (yes, just the total number) of papers published. Creativity hurts both of these measures.

At 7:50 PM, Blogger Female Science Professor said...

Your science planet is different from my science planet. I have to make decisions about students based only on their written applications (statement, letters of reference, transcript, GRE scores) and perhaps a brief visit. We don't take students for a test drive here.

At 7:54 PM, Blogger Ψ*Ψ said...

I agree with almost everything you've said! A lot of undergrads DON'T bother to read the literature, which is kinda sad. Great way to figure out what you're interested in and who's doing it. For those who have blogs, writing lit posts is THE best way to get used to reading through journal articles critically but quickly, and also opens discussions with other scientists around the world.
Not giving undergrads their own projects makes them technicians at best. Otherwise, there is someone else to deal with the inevitable setbacks. It's best if they stick around long enough to see their projects to completion. Catching the tail end of something makes it less personal. Being around for the rough beginning means a lot of frustration for little reward.
Granted, since I am just an undergrad, I'm totally unqualified to comment.

At 12:47 AM, Blogger Dr. J said...

Okay. My question would be: how much of these do you think you can detect in a couple of hours interview with a potential student or postdoc? That's a different kettle of fish to having a chance to work with them a while to evaluate their abilities/ behaviour/ personality.

At 7:13 AM, Anonymous Anonymous said...

Yeah, the FSP, she don't know nothin'.
It's a wonder how she ever got to be a full professor at and R1.

Yet more evidence of how phucked-up The System is.

At 11:15 AM, Blogger ScienceGirl said...

I am amazed how one-dimensional some people seem to view a student's future potential. While my professors simply told me "you are a great student, you should go to grad school," I tried to self-evaluate myself along the lines of your post because I wasn't as easily convinced (I tend to be pretty hard on myself).

I would like to leave the same question here as I did on the FSP's site: I would like to hear about the professor's role in grad student's success. What would you consider "adequate advising"?

At 12:45 PM, Blogger yajeev said...

Ms PhD,

This was a good read. It was fun to mentally go through your checklist and see how I measure up to your standards for good researchers.

At 2:36 PM, Anonymous Anonymous said...

Lab classes are useful for screening inexperienced technicians. But for evaluating independent researchers, I agree -- they don't test what's important (creativity, determination, getting the right answer) and do emphasize things that are counterproductive (cheating, complete inability to distinguish between important and unimportant).

I've got to think I'm not the only successful scientist who simply refused to keep writing down the molecular weight of water in every organic chem lab report...

At 3:38 PM, Anonymous Anonymous said...

but no one's perfect. you can have a shitty undergrad experience or perhaps none to speak of, and become a stellar grad student. if everyone was great from the get-go, then there would be nature papers for everyone. if everyone wanted a go-getter student who had some research under his/her belt prior to joining your lab, then who would be the guinea pig PI who would take on the student with no experience?

i used to be the undergrad who didn't read papers and had no prior experience. :)

At 6:39 PM, Blogger Kristin said...

If one were to follow this set of standards, they would rarely accept students with degrees from liberal arts colleges. This is troubling on many levels.

Faculty at liberal arts colleges don't have summer labs in the same way faculty at R1s have. By that token, it would be difficult or impossible for students to get "summer lab" experience. Students at SLACs have opportunities to do their own research (e.g., formulate hypotheses, gather & analyze data, etc.) during the normal school year.

On the other hand, all the paper-pushing involved in my current lab's summer work is accomplished by undergrad RAs. I doubt that their abilities to run the photocopier are vital for the future careers as researchers.

Truly... which set of students are better prepared?!?

At 1:50 AM, Anonymous JaneB said...

Interesting post - but maybe only applies to specific kinds of science and undergrad institutions. I'm in the UK, so we have three year undergrad degrees (and 11 month masters, for those who take them), which cuts down on opportunities for summer lab experience - and anyway, in my subject area/institution/national context, there is NO FUNDING for summer/term time undergrad placements, and most of my students have to have paying jobs in order to avoid massive debt on graduation.

Lab classes are very useful for: teaching basic technical skills to largish classes, giving students a taste of what they might actually be good or bad at in terms of practical skills (many students never get this sort of opportunity at school, so it's important that they do have the chance to 'catch up' at university) and, if even half-way well designed, do NOT need to have fixed outcomes but can offer real opportunities for students to think for themselves - within a very limited time period, such as most modular courses allow for lab time. I think that what is bad is if students do short, few hour, constrained lab classes for every course throughout their time as undergrads - it's important to have some pedagogic progression.

Final anonymous - the most successful academic researcher in my generation that I know (not in my subject - but has won European prizes, has a lucrative private consultancy, has had a choice of positions even in the current market and even been headhunted for some all before the age of 40) was a total screw-up as an undergraduate, the archetypical too-busy-drinking student for her first couple of years. People mature at different rates, people find their passion at different times, and any selection process has to allow for that.

Re: advising - one element that is important is the personal dynamic between supervisor and student. My advising style will not suit all students, and I can only modify it so much. Similarly my expectations about work habits, record keeping, work rate etc. are lax compared to some labs and viciously strict compared to others. I'm the kind of supervisor who, when a student comes in and says should I do y? or what does x mean? provides a reading list and says come back next week and tell me what you decided and why. Some people thrive under this, others panic. And as the student's maturity and confidence develops, so the advising needs to develop, and sometimes that means that one person is great for you for a year or two and then you need something different that they can't provide... in other words, it's impossible to really define 'good advising' as it depends so much on the individual situation. The one critical thing I would say you need to find out about is the availability of your potential advisor. Advisors have to meet with students regularly, be available regularly, respond promptly to information, be clear about when they AREN'T around - or you never get to build a relationship of any kind.

At 12:43 PM, Blogger Ms.PhD said...

Kristin and JaneB-

I get what you're saying about schools that don't provide funding or summer lab opportunities. But I don't buy that as an excuse. There are lots of programs where students can go somewhere else for the summer, and they provide funding and housing for the summer.

But I'd also argue that it's not good preparation for grad school to go straight from a liberal arts school with no research opportunities. Not at all. Most of the people I know who hated grad school, and left academia, were ones who came from colleges where they had never met a grad student before they applied to grad school. I think I've blogged about this before.

I think students need to know what they're signing up for, and the tendency is to try to hide it from them. I think this is irresponsible and ultimately hurts everybody.

But, resourceful students find out about these things. If there aren't enough slots in summer programs to satisfy student interest, we've just identified an area where it would make sense to invest the money, since a little bit of money at the undergrad research level goes a long way.

I agree that in some fields, being on a publication as an undergrad is difficult if not unlikely. Actually I agree with most of what was said in these last few comments. All I can do is write about my field. As FSP said, some of us live on different planets.

At 4:55 PM, Blogger Ψ*Ψ said...

The problem with a lot of summer research programs is that there definitely aren't enough slots to satisfy the level of interest. Also, many med-school-hopefuls apply for them in hopes of strengthening their applications--these are the students with 4.0s, so they crowd out a lot of undergrads who are genuinely interested in doing science.
Sometimes I wonder if I'd have received a more complete education at a liberal arts school instead of a research university that treats its undergrads as a cash source and not as students. I don't think I'd be looking at grad school if I hadn't been taking classes and working in the lab with PhD students, though.

At 4:34 AM, Blogger Kristin said...

"But I'd also argue that it's not good preparation for grad school to go straight from a liberal arts school with no research opportunities."

That wasn't my point, because that's really not the case. The liberal arts students I've known have ultimately come to grad school with better research training than R1 undergrads. SLAC students who do research as part of their studies have to do more to address their questions (e.g., the whole study from the ground up). The opportunity is there, but they have to take the initiative. Those who are grad-school bound will take this initiative and run with it.

In contrast, undergrad students in R1 labs might have some data collection experience, and might end up doing a thesis project with their advisor's data. They don't have the same experience in generating research questions, selecting protocols, or troubleshooting problems in the field. They show up, run the copier, and get paid. There's little incentive for ownership of the project, because they don't have the responsibility to make thoughtful decisions.

See the difference?

At 9:47 AM, Blogger Ms.PhD said...

Sorry Kristin, I thought you meant SLACs that don't have labs at all. I have one friend like that, and she has had to work ten times harder to make up for her lack of prior experience. I think she'll make a great FSP, but I also think she's very unusual. And like all of us, she talks about quitting when she's having funding problems or worries that she can't get the kind of job she wants.

But back to your point- I think you're wrong. Undergrads in R1 labs can have projects, too. I'm actually kind of annoyed that you assume they don't.

I do agree that my friends who went to SLACs (where they took the place of grad student labor) got great training and are really independent researchers.

But here's the twist: my friends like that were horribly disillusioned when they got to grad school. While they all finished their PhDs and made great contributions to science, they also all left academia immediately after grad school.

Here's why: at a SLAC, people pay attention to you. There is mentoring and feedback, and you're usually not working alone.

In grad school and at all the R1s I know of, you're much more likely to be handed a bench, a notebook (and sometimes, but not always, a project), and left to your own devices.

Since FSP's original post was partly about the question of making more FSPs, I'm saying that partly because SLACs tend to be somewhat idyllic, it's false advertising for what grad school will be like.

This is very cynical, I know. In some ways I'd like to hope that students from SLACs are more likely to come in and try to change the system for the better. But so far no one is changing it.

At 8:57 AM, Anonymous Anonymous said...

Thomas Cech-you might have heard of him, he won one of those Sweedish prize thingies-wrote a lovely article that argues that graduates of SLACs have always been a disproportionately strong presence in science at all levels. We don't all get disillusioned and drop out because we see things are different after leaving our lovely idyllic environments. So your friend did. So my freind who went to Berkeley for undergrad also dropped out. Does that say anything in general about the prospects of Berkeley undergrads in science?

The more I think of it, the more colleagues I can name off the top of my head that were undergrads at SLACs. Most are in faculty positions at R1's or med schools, and the ones that I can think of who are not are doing other things that they are happy with, like being profs at SLACs. Oh, yes, sure it was sad when I went from being the center of the universe where all was warm and love was everywhere with field-stone dorms and fire places and pipe-smoking tweedie rumpled philosophers discusses Kant and Hume with us over cognac, to the cold hard world of grad school in a lab with a working centrifuge and more than one agarose gel apparatus and postdocs who taught me much of what I know about genetics and all, but the $14K a year stipend helped.

Sorry you'r having such a bad time, YFS, but don't project your misery on the rest of us.

Here's Cech's article, for anyone interested in upbeat fact-based arguments.

At 9:23 AM, Anonymous Anonymous said...

who wants to change the system after they've spent their whole careers trying to play within the system?

At 12:57 PM, Anonymous Anonymous said...

Just curious, but why are all your good students referred to in the third person female pronoun? Have you not had a promising male student?

At 1:31 PM, Blogger Ms.PhD said...

@Anon 8:57 AM,

You're right, I do know one person (1 person) who went to a SLAC and now has a tenured position at a top-tier university running her own lab.

I don't care what Tom Cech said.

@Anon 9:23 AM,

Indeed, why change the system if it works for you. Because you don't want other people to have to go through what you went through?
If you can remember who you were before all that contorting to fit through tiny, flaming hoops.

@Anon 12:57 PM,

When I wrote this post back in 2007, indeed the best students I'd had up until then were female. Since then, I had several trainees who were male and also very good.


Post a Comment

Links to this post:

Create a Link

<< Home