Advice for students on interviewing

This week’s CS table topic was technical interviews. We considered a variety of essays that gave advice on interviewing people who have applied for programming positions, or related positions. We read some interesting claims about the small percentage of people who apply for programming positions who can actually program. We concluded by talking about implicit and explicit biases in the traditional CS technical interview processes [1].

But I’m not going to address any of those issues. Rather, I’m going to try to provide some advice to students. Now, I must admit that I do not interview people for industry positions. Nonetheless, I do interview students for my summer research positions. And I do care that my students succeed in their interviews. So I may be able to provide reasonable advice about how to do better in the process [2]. I realize that recent studies have shown that interviews have much less predictive power than most interviewers think [3,4]. But I can confidently say that my interviews have contributed to me hiring large numbers of spectacular students and very few duds [5,6].

Students tell me that my interviews are hard and painful. That doesn’t bother me. Part of my job as a faculty member is to help prepare my students for other interviews, whether or not I hire them. If my interviewers are hard but I reveal the process, my students will be better prepared for both the easy interviews and the hard interviews they have in the future. I’m not going to tell you everything about my interviews [7]. But I will reveal some of the things I’m looking for.

We’ll start with an important fact, one that I tell to my interviewees early on in the interview: I tend to have about three times as many applicants for a position as I can hire [8]. As a candidate, their job is to help me figure out why I should hire them rather than some of the other applicants. I try to make it easy on them; I ask it explicitly:

Historical evidence suggests that I will have about twenty-four applicants for this position. I can hire eight. What sets you apart from your peers? What characteristics do you have that should lead me to hire you other than sixteen other people? You don’t have to be my top candidate, but you have to convince me that you are better for my project than sixteen people.

While the numbers are different for every position, the candidate’s job is the same for each interview: They must help the interviewer figure out why they are better than other applicants, even if they don’t know who the other applicants are. Most interviewers won’t ask this question explicitly, but the question is at the center of most interviews [9].

I’ve set the stage. Let the advice and commentary begin.

Unfortunately, most students begin to answer my question by saying I’m a hard worker. I work well in teams. I’m really enthusiastic about the project. Now, those are important characteristics. And it’s okay to mention them. But they don’t set one candidate apart from another. I can usually get candidates to realize this by asking What fraction of my applicants give that answer? If they aren’t sure, I follow up with Most of them [10]. So, what should you do? It’s fine to describe strengths that you share with other candidates, but make it clear that you understand that they are not unique. One might add or imply something of the form that I expect that most people say something similar and I have other strengths. Nonetheless, I mention them because I know these are core skills required for the position.

We’ve gotten the common skills out of the way. What next? Well, you really should figure out what your unique or particular strengths are and be prepared to describe them. How do you figure those out, particularly given that very few of us like to promote ourselves? Introspection is good. Asking your friends is often helpful. Asking your faculty can be helpful [11]. Here’s one other tip Learn about the position and the employer and try to identify the skills you have relevant to the position and employer.

In my most recent set of interviews, I was looking for students who will help me develop curricula for two code camps and one course, with one camp and the course emphasizing data science. What might help with that? For the code camps, experience with kids would be a huge plus and for the data science aspects, experience with stats would also be helpful. But I had to drag those skills out of some people. I see here that you took a course on Political Economy. Tell me more about the course. Well, we had a large project in which we …. So you did a lot of data analysis and manipulation in that course? Yes. It was stressful, but fun. Given that we are trying to find a good set of examples for doing data science, might that be relevant? Oh, yeah, I guess so.

That dialog also suggests one other issue that is important for interview: Be prepared to provide evidence for your claims. Grinnell students should have learned that in Tutorial [12]. I find a surprising number of students have trouble when I ask them to further explain their claim. How do I know that you work well in groups? How do I know that you’re organized? [14] Why should I believe you’ll be able to provide particular insight about data science? [15] Some interviewers will ask; some won’t. You should volunteer the information. I know everyone says this, but I also know it’s essential to the success of your project: I work really well in teams and have a variety of experience in teams. Let me give you some examples … [16]

That’s enough advice for now. I didn’t italicize all the advice before, so I’m going to summarize [17]. Be prepared to promote yourself. In describing your skills, it’s okay to mention common characteristics, but acknowledge that they are common. Describe what makes you special. Do your homework about the position and the employer. Explain why your skills are relevant to the position. Provide evidence. Yeah, that should set you apart.

Hmmm … doesn’t that advice make me seem warm and generous? When students don’t promote themselves, I work with them to figure out what is special. When they make unsubstantiated claims, I encourage them to provide evidence and even help them figure out what evidence might help. That seems supportive, rather than scary. Nonetheless, my students tell me that the interviews are scary [17]. Perhaps it’s the ways in which I push them to promote themselves. Perhaps it’s that I’m explicit. Wow, that was a bad answer. Did you hear yourself? Really, that’s all the evidence you can give? Let’s try harder. Or perhaps it’s that I ask a few more questions that are intended to push students to think carefully about what makes them special.

Hmmm … Maybe I should remember the adage that You’re not just interviewing the applicant; the applicant is also interviewing you. Nah. They’re going to have to deal with me for a full summer; they should know what they’re getting.

[1] Given the results these processes give, it seems fairly likely that the processes are biased.

[2] And yes, most of our students need to do better.

[3] That is, it seems that interviews have almost no predictive power.

[4] Have recent studies shown that? I’m pretty sure I saw an article in the New York Times this week. Ah, yes, it’s called The Utter Uselessness of Job Interviews. I should read that article to learn more.

[5] Things that should undermine my belief that my process is successful, even if there is broader evidence that interviews are generally unsuccessful: (1) something like 80% of drivers think they are much better than average; (2) if I choose randomly among my applicants, I have fairly good odds of getting excellent students.

[6] The joy of typos: I had not previously thought about the fact that duds and dudes are only one letter apart. I have hired very few duds. I don’t consider whether or not my candidates are dudes.

[7] If you know more about my interviews, please don’t tell others.

[8] When I’m lucky, I can round up funding for eight students. I usually have about twenty-four applicants.

[9] Okay, in some tech interviews, the question may be the other way around. Given that there are a dozen places that top candidates can choose between, why should they choose this company? But let’s pretend that the interviewer has the upper hand; they do in most cases.

[10] Sometimes the applicant knows to answer All of them. In that case, I respond with Yes, all of them, or at least most.

[11] Ask your faculty assumes that you are at an institution in which your teachers are likely to know something about you.

[12] Oooh! That brings me to three things every Grinnell student learns in Tutorial: know your audience, cite your sources, and provide evidence for your claims.

[14] My favorite answer to that question: Let me show you my color-coded notes for this interview.

[15] In my political economy course, I did a large project on X. We gathered data from Y and Z. Obviously, we had to massage the data.
We found these visualizations to be helpful. That could be a good process to use in the class. I can also rely on my project and my experience thinking about my peers’ projects to help come up with good examples for your class.

[16] Okay, perhaps that’s getting a little long.

[17] And italicize.

[18] One of my favorite students said, I was ready to declare a major with you. Then we did the interview. I realized that I should choose a different advisor. And this is a student who I hired!

Version 1.0 of 2017-04-14.