How to Do Research With Undergrads

Collaborating with undergrads on research can be both helpful and rewarding! Remember that the primary goal is mentoring: think of it as a form of service where you help introduce a talented undergraduate to research and, as a side benefit, you get some extra help with your research.

Adrian has some draft notes on advice for working with undergrads. This advice is still in progress but may someday become “official” CS department advice.

Logistics for Mentors

Here’s a step-by-step guide to working with undergrad researchers:

1. Decide what you want to do with an undergraduate researcher. It’s probably a good idea to talk to your advisor.
2. Find a student to work with.
   • You can direct interested students to our instructions and “application” form. New issues filed in the undergrad-research GitHub repository get posted to the #general channel on Slack. You can chat with the student in their issue comment thread.
   • Going through the issue system is not required! If you meet someone some other way, you might find the questions on the form useful to get a conversation started. Or if you think a student might be good but don’t have the bandwidth to work with them immediately, you can redirect them there.
3. When you have someone you want to work with, it’s important for your advisor to meet them to help set expectations.
4. To get things bureaucratically squared away, the student should either register for CS 4999 research credits or get hired for hourly pay (but not both). Please ask them which they prefer and help them take care of the logistics. See the next section for details about getting signed up either way. In either case, please pay attention to the timing for the logistics—there is either a deadline or a “lead time,” depending on the route.
5. Finally, when it’s time to start doing research, please orient your new collaborator as a member of the group! Be sure to invite them to our Slack workspace. A good place to start with the orientation is to send them a link to the root page on the Capra internal site and the “New to the Group?” bullets therein. Invite them to lunch meetings and the Slack channel for the project they’ll be working on. Set some expectations about how you will be coordinating: probably by scheduling a weekly meeting. As much as possible, encourage them to post to Slack, especially in the #overflow channel, with questions and comments—try to counteract their natural tendency to be intimidated by the group by encouraging communication.

Logistics for Undergrad Researchers

So you’d like to start doing research with us—great! First things first: please make plans with the people you’ll be working with—probably a PhD student. Then, here’s what you need to do to get officially signed up.

Your first logistical decision is whether you want to do research for credits or for money. For the former, you sign up for CS 4999 for the semester. For the latter, you get hired as an hourly employee. You can only choose one of these for a given semester, but it’s pretty much up to you. But in the summer, you almost certainly don’t want to do credits because it will involve paying tuition—so please do the money option.

Credits

Please just register yourself for CS 4999 credits in Student Center. You do not need a PIN from Adrian. You will need to decide whether you want a grade or a pass/fail outcome.

The big decision is to think about how much time you will devote to research this semester—and to pick a number of credits accordingly. Roughly speaking, one credit equals about 3 hours of work per week, but you should think of other classes you’ve taken and how much time and effort you spent on them. Set your credit count according to the average amount of time you’ll devote every week.

Recently, the College of Engineering has gotten more serious about enforcing a deadline for adding research credits. This new deadline is later than the “add date” for normal courses but it’s not that late. Currently, this calendar on the Engineering site lists the date, vaguely but screamingly, as the “ABSOLUTE LAST DAY to ADD Project Team courses.” (This date apparently also applies to research credits, not just project teams, despite the wording.)

Money

This route works by hiring you as a research assistant—typically (but not always) as an hourly position where you fill out a timecard every week, in which case there is less up-front commitment about the time you will spend.

Shailja can help with hiring. The PhD student mentor for an undergrad researcher should email or Slack Shailja, probably cc’ing the person to be hired, to start the process. [Everyone: Please include notes here about what information should be included in this initial email/Slack message, as you discover what’s most important.]

There is no specific semesterly deadline, but getting hired takes time and needs to go through several levels of bureaucracy. So as soon as you know you’re going to do this, get the process started—at least a week or two before you plan to start working.

Guidelines

Here are some general guidelines for how to approach undergraduate research, which partially duplicate the aforementioned [notes][ugr-notes]:

• For research during the normal academic year, try to pick a target number of hours per week that you’ll be able to focus on research. Take in to account your other obligations, such as coursework and TA duties. Research during the semester should be flexible: some weeks, you’ll need to focus on exams or homework and won’t get much done.
• For research in the summer, treat it like a full-time job. Don’t attempt to balance internships or summer classes with research. To get the full value out of a research experience, aim to spend 40 hours a week on it. Summer is a unique opportunity to make lots of progress on your project, so don’t take on other big obligations that will take away from your focus.
• Researchers should take on only one project at a time. Both during the summer and concurrent with a semester, it is probably a bad idea to balance multiple projects in multiple labs. Taking on one research project at a time will make sure you can focus and do great work. (It is a common misconception that working in multiple labs is good for getting into grad school—this is false, but that’s a topic for another page.)
• Be sure to ask questions frequently on Slack or in person. This doesn’t come naturally to everyone—it can be tempting to work in isolation to get past a problem. But your team and mentors are here to help! Try to force yourself to ask questions earlier and more often than you might do naturally, even if they seem simple.