Observations on Running a (CS) Research Lab with Undergrads
Hi. My name is Jordan, and I’m a teacher. Well, professor. I like that title too, but I think teacher is maybe a little more accurate. It highlights the part I love the most.
I’m also a visual analytics researcher. Visual analytics, in a nutshell, means using interactive data visualization and machine learning to help people find unexpected things in data. A lot of this research takes place at R1 institutions, or at similar research-focused institutions around the world. It also takes place at liberal arts colleges and other institutions focused on undergraduate education, like the one where I teach.
My students do all kinds of cool things, like survey how visualization has been used in cybersecurity and change how we think about ancient Syriac handwriting. They even help us learn how “laggy” systems influence how people search (with thanks to our collaborators at the University of Maryland, who invited us to join their project). Undergrads make real contributions to science, and I’m deeply proud to have the opportunity to work with them.
Recently, while attending the IEEE VIS Conference in Vancouver, BC, a colleague at a large public research university asked me to share my thoughts on how to manage a cohort of undergraduates in a research lab. Undergraduate students are a lot like graduate students in many regards, but there are also some key differences (and maybe a few misconceptions). Though I’m not a regular contributor to Medium, I thought I’d try writing my thoughts down here to make them easier to share. I hope you find them helpful, or at least a useful conversation starter.
Q: How do you manage expectations during the term when they [undergraduate students] have other priorities (like classes)?
A: It may be a bit different for those of us at PUI (primarily undergraduate institutions) than for colleagues at research universities, but I try to be really up front with students: I tell them that it’s my expectation that they are students first, and research assistants second. I choose projects that are engaging and have clear, “bite sized” pieces, and I have them work in teams. (VIS research, it turns out, is great for this.)
We have a shared lab calendar where students mark in advance when their exams are, weeks that they expect to have especially heavy workloads in their courses, when they plan to go home for break, etc. I find that this has the practical benefit of helping us communicate better (and making me worry less when they go silent), but also helps scaffold them as they learn to juggle multiple commitments.
This is especially true for the first-year / “freshman” researchers in my lab. Remember: traditional-aged undergraduate students (18–22) undergo significant brain development during their college years, and for many of them, this is the first time they’ve ever had to manage this much scheduling and planning without “adult” intervention. I don’t mean this to sound patronizing; I want to reiterate that my undergraduates do amazing work, and are able to bring a lot of truly insightful thoughts to the conversation. They’re just still growing.
Q: How do I include them or get them up to speed on academic writing?
A: First, I think everyone (undergrads included) gets better at writing when they read good writing. I have a selection of papers that I think are very well-written, and they’re part of the “onboarding cannon” — every student in my lab reads them, shares their thoughts on them, and starts to get a sense for how people in a particular research community talk about their work. We also sometimes read papers that aren’t so well-written, which gives us an opportunity to talk about what makes some writing styles more compelling than others.
Second, blank pages are my personal kryptonite. Luckily, my undergrads are universally awesome at turning blank pages into (admittedly, sometimes crappy) drafts. From there, we iterate and iterate and iterate… but importantly, I share with them the aforementioned fact about my difficulty getting started. They know that they’re an integral part of getting our lab’s work out there, and I think that helps the editing we wind up doing feel less painful and more collaborative.
Last, we do a lot of writing outside of deadline crunch time. They keep online lab notebooks, and write in them every day: they write down their plans, celebrate their successes, and document where they got stuck. They draft sections for the monthly reports I send to our funding agencies. They send papers to undergrad-specific venues, as well as to more prestigious conferences. Two-page poster abstracts are a great way to give them reasonably low-stakes practice: since the format is so short, I find it’s much easier for me to help edit them.
Q: How do I recruit students who may underestimate what they can offer?
A: I think the most successful approach (measured by the number of students who wind up in my lab) is that I talk a lot about research opportunities at venues where I think those students are likely to be: events for newly-admitted students, orientation, etc. My institution also has several programs that help with this: I write a brief abstract about my current work and the program matches me with one or two students a year. Their research with me counts as their work-study position.
It’s important to be honest that undergraduate students, especially very young ones, aren’t always immediately “productive.” For me, that’s okay. My funding situation is sufficient for me to have ~10 students work with me in some capacity each year, so it’s not a big deal if some of their projects don’t pan out. Whether we like to admit it or not, failure is an integral part of all research. We talk a lot about failure, what we learned from it, what they’d change if they had it to do over again, and we try to set up enough firebreaks during the planning process that the world doesn’t come crashing down when a project flops.
Q: Independent vs. group projects? If group, how do you make sure that one student getting overwhelmed doesn’t hold back the rest of the project?
A: Most definitely group, at least during the onboarding process. However, it’s also important for students (undergraduate or graduate) to feel some level of ownership and investment in what they’re doing. In my lab, we balance this by having students work on multiple projects and playing different roles, spreading out who gets to “lead” each project. This also helps with fault-tolerance when one student needs to step back for a bit, as does that fact that we try to communicate about when that’s going to happen.
Eventually, a subset of them are ready / want to lead an independent project, and this usually takes the form of an honors thesis (if it’s a research-y project) or an independent study (if it’s an application or practical tool). Even in this case, their connection to the group stays important. Their labmates give feedback, help debug, listen to practice talks, and serve as initial pilot users.
They also do the same for me: when I implement something, they review my code and offer suggestions. When I write something, they correct errors. We really don’t do anything alone in my lab… but not because my students can’t handle it. Instead, it’s because I believe that it’s important to cultivate a community culture where seeking help and regular input from others is viewed as normal, healthy, and welcome.
Closing thoughts
I’d like to thank the person who asked these questions, as well as the person who asked if I would be willing to share them more broadly: it has been very grounding. I hope that when these students head off to be (your) PhD students, these practices will have prepared them well, shielding them somewhat from imposter syndrome, and ultimately making them stronger researchers.
Thanks for reading.
