I have a lot of complicated feelings about the process of hiring
student researchers, and the world that makes advice like this (and
similar advice across the .edu Internet) applicable. I
think it is effective for navigating the world as it is, but I am less
convinced that the world should be that way. The main body of this post
has been aimed primarily at students; this section is for my fellow
PIs.
Is familiarity with academic social protocols really a prerequisite for being a graduate assistant? To what extent are our heuristics filtering for irrelevant things like that rather than relevant skills, experience, and thought processes? How can we improve this process?
I’ve been impressed by Thomas Ptacek’s thinking on tech hiring, and the work that he, Erin, and Patrick are doing at Starfighter to try to make it easier for more employers to test the things that really matter. I’m trying to incorporate some of these ideas into how I look for graders; I just asked a candidate to do a little bit of demo grading on a broken solution to an old homework, and if that seems useful I may develop it into a ‘standardized work-sample test’ for grading. It’s less obvious how to apply this to research positions, and how to balance the need to get research output with the goals of training and education. The best thing I have so far is taking my class, especially Recommender Systems, as the assignments do give me a taste for a student’s ability to do the kinds of work they need to do in a research context; I’ve also tried to use simple LensKit tickets, but the onboarding process there isn’t good enough yet for that to be particularly effective. I’ve also had success hiring a student first as a grader, and then as a GRA after I had already worked with them for a semester.
The balance between skill coming in and the educational goals of the university system is also complicated. The most technically skilled candidate is not necessarily what we should be looking for. Also, people in general are not static, and the point of being a student is to not be static (or maybe just pick up some credentials, but let’s assume for a moment that higher education isn’t only signalling). There’s some debate over the legitimacy and appropriateness of the ‘growth mindset’, particularly when applying it by berating yourself for perceived lack of growth, but I find it to be a useful prior (in the Bayesian sense) for interacting with people. I try to assume, absent evidence to the contrary, that each student is able to grow and improve as a programmer, a scientist, a scholar, and a person, and that I may be able to play some beneficial role in that development. If I did not make such an assumption, being a teacher would be a very questionable occupation.