How to test (admitted) grad students before hiring them in your lab

Question

Recruiting "bad" PhD students who only become emotional, temporal, and financial drag is not good for anybody. I was listening to the recent freakonomics podcast episode, and thought it would be great to use tricks for testing the candidates before hiring them. Interviewing only can tell you so much, but unlike industry, PI's have more time to test the students before committing to seriously bring them into their research program.

In a related question, some qualities of successful students were discussed. Specifically, I want to test for

• persistence/focus
• creativity
• logical/systematic thinking
• communication
• teamwork
• "smartness"
• some basic skills (programming, computer skills, writing, etc.)

At the same time I do not want to punish for lack of knowledge or experience. Also, I'm okay with eccentric personalities to a certain degree.

I've seen PIs testing their candidates by giving them mini research problems before hiring them (or while they are rotating in the lab). Some students will quickly finish the task and further explore the science on their own write fantastic reports/papers, while some students never finish the simplest first step. This seems to work fine (but some students might think it is not fair). I would like to learn if there are some quick tests that would reveal the quality of the candidate.

What tricks/procedures do you use?

P.S. I am in a computation/theory heavy science in US.

EDIT: I am especially looking forward to King Solomon's cutting the baby in half type of creative solutions. Perhaps PI can ask the student to do an impossible task and see how long it takes before the student says he/she thinks it is impossible.

Answer 1

What tricks/procedures do you use?

I am especially looking forward to King Solomon's cutting the baby in half type of creative solutions. Perhaps PI can ask the student to do an impossible task and see how long it takes before the student says he/she thinks it is impossible.

You're contemplating starting out a long-term, intense professional relationship with this person. "Tricks" like this are dishonest. Dishonesty is not a good way to begin such a relationship.

The situation is similar to an ordinary job interview. In a job interview, it's not just the employer who is judging the candidate. The candidate is also judging whether the employer is someone they'd want to work for. Manipulative interview techniques are a flashing red light telling the candidate that this is not an employer who respects his employees.

Another thing that a job candidate is looking for is an employer who will support him in his long-term professional growth, rather than treating him as a labor commodity who needs to be productive from day one. This holds even more for a student getting started in a PhD program; the reason universities claim that grad students are not employees is that they're supposed to be focused on learning, not on acting as cheap labor.

In the usual case where the student does a rotation in your group, that's the opportunity for each of you to see if you're a good fit for the other. It's meant to be an experience that helps the student to learn and grow, not one that produces a lot of research. Part of what the university is paying you to do is to provide these opportunities for these students to learn and grow. If, by the end of the rotation, the student has shown a lack of aptitude, then you've provided the service you're being paid to provide.

Answer 2

What tricks/procedures do you use?

I work with each student as a potential colleague. I meet with them regularly, monitor their progress, offer what feedback and advice I can on what classes to take, papers to read, problems to work on, techniques to apply, people to work with, conferences to submit to, writing, presentations, and so on.

If they show sufficient promise/progress after a semester or two, I offer to continue working with them in an official capacity as their advisor. If they don't show sufficient promise/progress, I offer to help them find another advisor who better matches their interests, background, and working style. If necessary, I help the student navigate a change in degree programs, departments, or universities.

Finally, if I don't have enough time to effectively evaluate and/or advise a student, I just say no at the beginning.

In other words, I don't have any tricks. I just do my job.

Answer 3

I don't think it is fair to judge a person by tests. People have ups and downs. Look on their overall profile. Have they done something creative even once, in their career? Some students are exam-phobic and do not get A grades all the time, but if you look at their research and thesis in the past, those are very creative and novel. Those kind of people are the ones, that succeed, not always the ones who run after a 4.0 gpa(on a 4.0 scale).

If the person doesn't has any novelty to show in their past projects, ask them the reason. What was the idea behind the work they have done, what did they learn from it, and what do they intend to learn from their PhD. Why are they pursuing a PhD? Why do they want to join your lab? Tell them what are you looking for, in the candidate and ask them to do a self-evaluation. Tell them to be frank, if they do not know something that you want them to know, and that you'll be willing to work with them if you really see them as the right candidate.

Keep your expectations real. You might never find a student with 100% qualities matching your requirement, but if you find that right candidate who is willing to give his 100%, you can make their 70% match better than anything else you'd ever get.

At the end of the day, a good student is the reflection of a good teacher.

Oh and by the way I'm also a Bioinformatics major and it is not always possible to know everything on the computer skills side. They might be excellent in one programming language but know nothing in others or they might be average good in a number of languages being the master of none. At the end, the thing that matters is, if they can get the job done, using whatever technique, whatever approach!

Correct me if I'm wrong. :)

Answer 4

Not to sound facetious, but have you ever considered your role as the mentor in understanding why these collaborations don't work out? Relationships are bidirectional. As such, I am struck that in your initial post the issue is the deficit of the student.

I'm a grad student. I've worked with many professors on different projects. In every instance you get what you give. Most normal people will do their job (i.e., the grad student meeting expectations) if you do your job, at least in the context of a collaboration across time.