Cultivating our seed corn (after Eric Roberts)

In which I report briefly on my first LACS meeting, discuss the hiring crisis in computer science, and plea for CS Ph.D.s to return to academia.

Last week I visited Williams College for my first meeting (as a member) of the Liberal Arts CS Consortium (LACS). Since I’m hosting the group at Whitman College next year, I got to take minutes. Topics ranged from enrollments to hiring to space to curriculum to programming languages to organizational strategy. I led a discussion of senior assessments, which will definitely figure into the design of Whitman’s CS major, and probably into this blog as well.

One question was surprisingly engaging, leading us far away from our planned agenda: How can we expand the pool of potential CS faculty who want careers at liberal arts colleges?

It’s a question of great relevance: Out of the 16 institutions participating in this year’s LACS “show and tell” survey, 11 positions are going unfilled. I consider myself very lucky that we were able to hire John and Andy.

It’s also a question I’ve been thinking about for quite some time. It’s motivated me to organize three panels on faculty careers at the Grace Hopper Celebration of Women in Computing, and, this fall, my first panel at the Richard Tapia Celebration of Diversity in Computing. I revised one of those panel proposals into an article for XRDS, the ACM’s student magazine. (Editorial requirements being what they are, the XRDS article is nowhere near as good as the panel proposal it is based on. It also pisses me off to no end that a stock photo of a man was chosen for an article written by six women.)

My focus has always been on persuading grad students to consider liberal arts colleges as an alternative to research universities. But the truth is that few CS Ph.D.s go into any kind of faculty career. Most go on to careers in the tech industry.

At my own commencement from the UW CSE Ph.D. program, I was the only one going to a liberal arts college. Only three of us had jobs at any kind of academic institution. The rest were off to industry. A majority of my classmates were hired by Google.

While my situation was a bit extreme, the overall trend is borne out nationally. The Computing Research Association conducts an annual survey measuring “the enrollment, production, and employment of Ph.D.s in computer science and computer engineering (CS & CE).” Over the past ten years, around half of all CS Ph.D.s have gone into industry:

This chart does not cover the first computing boom of the early 1980s, but it is easy to identify the dot-com bubble of the late 1990s, the subsequent bust, and the recovery leading to an even greater boom today. Right now, just about 10% of new Ph.D.s are going to non-Ph.D.-granting departments, such as those at liberal arts colleges.

At the height of the dot-com bubble, Eric Roberts (Stanford) argued that hiring too many CS Ph.D.s in industry is “eating the seed corn.” We need Ph.D.s in academia to produce more CS graduates and more CS Ph.D.s. He is making that again argument today, and recently launched a new Masters of Education in Computer Science for experienced teaching faculty who want to change fields.

What was new in the LACS discussion last week was a different way of framing our recruiting problem: How can we get CS Ph.D.s to consider liberal arts colleges as an alternative to industry?

The group brainstormed several exciting ideas: persuading Google to send Ph.D.s back to academia on a short-term basis; developing programs to recruit women who are contemplating leaving the tech industry; changing how we advise our undergraduates about grad school. My contribution: I think CS faculty should be trying to recruit undergraduates who are interested in faculty careers, as their prospects of landing a position are probably higher in CS than in any other field.

Our brief conversation was incredibly generative. This seems like a fruitful topic for a birds-of-a-feather session or workshop at SIGCSE 2017.

More specifically, Valerie Barr (Union College) floated the idea that someone should write a funny, tongue-in-cheek PR piece for the Communications of the ACM pitching liberal arts college careers to Ph.D.s in industry.


My first draft is below. Feedback from both industry and academic colleagues is very welcome.

Tech industry Ph.D.s: Academia can be nicer than you think

Dear industry colleagues:

I understand some of the reasons why you took your Ph.D. and ran: The glamour and mystery of Silicon Valley. The opportunity to work with smart colleagues on products that make a difference in people’s lives. The excitement of a rapidly changing field. The opportunities to change jobs and take sabbaticals. The hours. The perks. The paycheck.

And being in industry means never having to write a grant proposal.

I, too, recoiled at the thought of a research career in which I would spend my time writing grant proposals, managing grad students, and writing more proposals so that I can continue to support my grad students, never touching a line of code. I, too, considered a career in industry, albeit very briefly. Instead, I’ve pursued a career teaching in computer science at liberal arts colleges, along with doing a little research and more administration than I ever thought I would. Ten years and still going strong!

Liberal arts colleges differ from research universities in a few key ways:

  • Liberal arts colleges focus on undergraduate education. By definition, a liberal arts college does not grant Ph.D.s. Some liberal arts colleges offer masters’ or professional degrees, but these programs tend to be quite small.
  • Liberal arts students are “liberated” by learning many different ways of thinking. In a typical engineering program, major requirements might constitute 60% of a student’s coursework, or even more. At a liberal arts college, major requirements typically comprise 25% – 35% of a student’s coursework, with the remainder spent exploring other disciplines, and perhaps developing depth in a second discipline through a double major or minor.
  • To earn tenure, liberal arts faculty are typically expected to demonstrate excellence in teaching, an active program of scholarship, and contributions to their institution and profession. Faculty might spend 60% of their time on teaching, 30% on scholarship, and 10% on service.

Teaching at a liberal arts college has more in common with working in the tech industry than you might think. Let me make a feature-by-feature comparison:

  • The glamour. Silicon Valley is glamorous. But liberal arts profs are mini-celebrities among their students and often in the towns where they work.  “Scientist” and “teacher” are the fourth and fifth most admired professions in the United States. Chances are good you’ll be the only local expert on your research area, and many other things besides. It’s a small pond, but we are all big fish!
  • The mystery. Most people don’t understand our jobs any more than they understand yours.
  • Working with smart people. Believe it or not, most academics are pretty smart. Working at a liberal arts college means nearly all of your colleagues are not computer scientists, which means you can learn a lot just by talking with them.
  • Meetings. We have meetings just like you do. What a great opportunity to learn from our diverse and thoughtful colleagues! I’ve gotten to know some really cool people by serving on committees with them. The very best meetings are the ones where something gets done.
  • Making things. Teaching is all about creating learning experiences. I still write code: I write my students’ homework assignments.
  • Getting your hands dirty. Undergrad researchers need a lot of supervision. Chances are good you will find yourself working right beside them, whether it’s to teach them how to do it right or figure out what the <expletive> they did wrong. You’ll get to investigate bugs you never even dreamed were possible!
  • Making a difference. Okay, I admit it: Nobody uses the things we academics make. Except our students. Strangely enough, students mostly do what we tell them to do. They seem to trust that our instructions are intended for their learning. Sometimes, students and alumni even come back and tell us how much they learned from our classes, projects, and assignments. Not to get sentimental or anything, but in the last month I’ve gotten notes from two different alumni thanking me just for spending time with them.
  • The excitement of a rapidly changing fieldIt seems I never teach the same course twice. What we teach in academia is influenced by what you are doing in industry, so we’re changing right along with you. And the body of foundational knowledge in CS is still expanding. For example, between studying computer architecture as an undergrad in 1996 and teaching it for the first time 15 years later in 2011, pipelining and superscalar architectures made it into the textbooks. No big deal, right? CS is also expanding into interdisciplinary applications. Liberal arts colleges value learning across disciplines, and our relatively flat organizational structures facilitate collaboration. And even if by some miracle the technology and content of a course haven’t changed, the students have.
  • Changing jobs. It seems like folks in Silicon Valley change jobs every 2-3 years. I change jobs all the time. Every semester I am teaching different courses on a different schedule. Every year I have different administrative responsibilities. Bored? Frustrated? Just wait three months!
  • Sabbaticals. Sabbaticals are a unique opportunity in the tech industry, compared to most other industries. (As fast as everyone changes jobs, no one will notice a few months’ gap on your résumé.)  By contrast, sabbaticals are de rigueur for faculty. The traditional sabbatical is one year out of every seven (hence “sabbatical”), but some colleges offer more than that. The truth is, we need it—a time to renew, reflect, research, write, and rest. Undergraduates are young. We’re old. They only stay for four years. We’d die if we tried to keep up with them all the time.
  • Summers. You don’t get summers off? We do! Well, not exactly off. Most liberal arts college faculty have 9-month contracts, so we are on our own recognizance over the summer. Some use this time to take on short-term contract work, or they write summer salary into their (gasp!) grant proposals. Most write, research, travel, supervise students, and/or prepare for the coming academic year. It’s important to rest up during our summers, too, even if we sometimes feel guilty for not getting as much done as we planned to. Fortunately, our 9-month salaries are spread out over all 12 months of the year, so we don’t starve like grasshoppers.
  • The hours. Like you, we sometimes work long hours to meet release deadlines. That means not just meeting conference deadlines, but also getting big assignments out,  finalizing grades, or making recommendations for hiring. But much of our work falls into a routine, and we don’t get brownie points for arriving early or staying late.  No one is watching (except maybe our students, who have been known to remind the less self-disciplined of us to go home and eat dinner). Much of our work is our own choice—usually how we do it, and often what we do. When I’ve had the longest hours, I’ve been doing things I enjoy: meeting interesting people, teaching new classes, and working intensely with my students. Also, see “Summers” above.
  • The perks. You know all those perks that Google and Facebook offer? They are trying to create a college-like environment. Why not go for the original? Free lunch might not be offered every day, but it’s often an incentive for attending seminars and committee meetings. The dining hall is cheap. Candidates and speakers need to be taken out to dinner at fancy restaurants. And why is there so much alcohol at faculty events? To get us to stop working on our own things and talk to each other. Many colleges offer inexpensive housing near campus for new faculty, so you can roll out of bed and walk to class just like your students do (though maybe not in your pajamas). Trust me, after a couple years you won’t want to be quite so close to campus anyhow.
  • The paycheck. I won’t lie: Academic paychecks are nowhere close to Silicon Valley paychecks. Even though computer scientists tend to earn more than most faculty, some of your new grads will earn a higher salary than you do. There are no annual bonuses or stock options in academia. On the other hand, the same salary goes a lot farther in Grinnell, Iowa or Walla Walla, Washington than in Silicon Valley. And in the eternal words of Jessie J: “It’s not about the money.”
  • Funding research. While most liberal arts colleges will help you write external grant proposals—and I have many colleagues who have received external research grants—few liberal arts colleges actually require faculty to seek external grants. If your equipment needs are modest, there is often internal funding to support students. Internal proposals are reviewed not by other computer scientists (or by the C suite), but by your colleagues in the liberal arts. This means that as long as you can explain it convincingly and eventually get it published, you can do pretty much whatever kind of research you want. There’s no need to reframe your real interests to fit into a corporate or national research program.

I haven’t written this just to hear myself talking (though it sometimes seems faculty are prone to that). CS is facing a crisis in hiring. An increasing number of new Ph.D.s are bound for industry, which means faculty positions are going unfilled. Whatever good experiences you had as a CS undergrad, students today are not necessarily getting the same experience. Fewer faculty means bigger classes, less hands-on mentoring, and a chillier climate for women and minorities. We need you to help us make undergraduate education everything that it should be.

Thinking long term, if more Ph.D.s don’t return to the academy, then there won’t be enough graduates capable of taking all those high tech jobs. Thinking even longer term, if there are not enough CS graduates getting Ph.D.s, the whole pipeline could grind to a halt.  And liberal arts colleges make a real contribution to maintaining the Ph.D. pipeline: In measuring yield of science students who go on to earn Ph.D.s, baccalaureate colleges are second only to “research universities with very high research activity,” and constitute the majority of the list of top 50 institutions. It is critical that we get new faculty colleagues who can fill open positions at liberal arts colleges.

Please consider a return to academia, whether forever or just for a year. I promise, there will still be jobs in Silicon Valley when you go back.

Your undergrad professors
(We always believed in you!)

Updated October 13, 2016, based on feedback from Valerie Barr. Thanks, Valerie!

10 thoughts on “Cultivating our seed corn (after Eric Roberts)

  1. Tom

    I remember way back in ’97 when I showed up at Whitman planning to go into the CS program only to be told that the program was being cut becasue it didn’t fit in the mold of a Liberal Art School (I am glad to hear that has changed now!) I, along with dozens of others were forced to choose other academic and career paths becasue of that decision. How many other institutions made similar decisions back then? So my real question is, how much of the fault lays with academia itself? In particular, did the liberal arts schools, who so often cultivate the students who desire to stay in academia for their careers, end up pushing those students either in to other fields or, if they decided to stay in CS, to more technical schools where they were pushed into industry?

    1. Janet Davis Post author

      Tom, it seems that most CS faculty at liberal arts colleges have had to persuade their colleagues that CS belongs in the liberal arts. Math faculty who pursued graduate-level work in computer science, such as my Grinnell colleague Henry Walker, have played a major role in this by getting a foot in the door. But at most schools, that happened years ago. From what I know, Whitman is very unusual in that it had a CS program in the mid-1990s, just before the dot-com boom, and chose to cut it rather than expanding it. By now, most of Whitman’s peers already have well-established CS programs.

      I also find it interesting that a disproportionate number of STEM grad students (including CS grad students) got their undergrad degrees at liberal arts colleges.

  2. Darakhshan

    This is hilarious.
    On a more serious note, we should ask NSF’s help on this. We should create more Fellowships (similar to Wellesley’s Norma Wilentz Hess Fellowship) both for fresh PhDs and PhDs in industry (maybe biting up with Google, etc.) to lure them into our institutions, and see for themselves what we offer. This creates a two-day pipeline when/if we partner with industry: our students into Google, Google’s researchers into our institutions, perhaps temporarily, perhaps not. Our students are phenomenal, and we need to provide a way for engaged, bright people to discover all the joys of working with them in several capacities.

    We should deliberately create exemplary mentoring models in our departments (very much like Wellesley CS department did for me) to invest in and retain teachers (especially when we don’t have tenure-track positions for them) through these Fellowships.

    1. Janet Davis Post author

      Indeed, the fellowship you held at Wellesley came up in our conversation! Engaging the NSF is a good idea; I think other LACS members who have connections in the NSF may already be starting to do that.

  3. Amy Csizmar Dalal

    This is an excellent post, and I love your column draft!

    I wonder to what extent we need to start getting the word out to PhD advisors, too. I know that in my department in grad school we were *actively discouraged* from pursuing *any* type of academic post. There was tons of mentoring and networking opportunities if you wanted to go into industry, but if you wanted to go into academia, particularly to a teaching institution….well, you’re on your own then. And it’s not like I was at a Top 5 school either. I wonder how much this has changed — I fear not very much.

    1. Janet Davis Post author

      Whoa, seriously? Your profs didn’t want to create mini-mes? I can only think they were caught up in the dot-com bubble, or really unhappy. (Admittedly, I also got advice to go to industry, but that was because someone thought I didn’t have the work ethic for academia.)

      In any case, I agree that this is a message for PhD advisors, too. I’ve been learning from folks at UW (my alma mater) that having even one person who can advise students who want liberal arts jobs can make a big difference. Every panel I’ve organized, I’ve listed the audience as “Ph.D. students and their advisors” but it’s often seemed like we were mostly preaching to the choir.

      1. Amy Csizmar Dalal

        Maybe this is the next “in a box” resource that NCWIT should develop. “Advising PhD Students for Liberal Arts Careers In-A-Box”!

        1. Janet Davis Post author

          That is a wonderful idea. At SIGCSE this spring, Tammy VanDeGrift and I talked about being ready to write a followup to our 2006 ASEE paper now that we have significant search committee experience. Perhaps the right target audience for that is PhD advisors, since they rather than their students are now our peers. I may be in touch to ask you more about this!

  4. Valerie

    Hi Janet! I think it’s quite good — the one thing I’d add is this: those of us who teach in that high-touch setting are molding the people who will be the creative thinkers, developers, creators of the future. So we may not be the ones sitting in Silicon Valley making the next new shiny thing, but without us there’d be many fewer new shiny things. Particularly because of the high-touch setting of liberal arts colleges, and the focus on undergraduate research, our students are uniquely prepared to go on and make significant contributions in industry — so we are the grandparents of a lot of new technology!


Leave a Reply

Your email address will not be published. Required fields are marked *