It doesn’t make economic sense for Universities to do Research

September 6, 2010 at 9:25 pm 11 comments

I just found Rich DeMillo’s fascinating response to my query of several weeks ago, “Why do Universities do research?”  Rich’s bottomline is “It doesn’t make economic sense.”  A massive intellectual property win is a pipedream.  Research isn’t necessary to be really great — Rich just did another fascinating blog post on that topic, including a list of top PhD producing undergraduate institutions (most of whom are liberal arts institutions without a research program).  A research program costs far more than it brings in. So what’s the point?  One suggestion: Institutional envy.

Viewed through this lens, Guzdial’s questions are even more interesting.  It frequently makes little economic sense for a university to conduct research. It may be part of the mission of a multiversity, but it is not the only mission — and there are plenty of examples to guide other choices.  If the dream of IP commercialization success drives  institutions to build their research programs, what about the data that predicts little chance of success? And if a university is concerned about reputational hierarchies, does building a research portfolio actually help?  Among the many components of a modern multiversity, few could survive without the instructional programs.  Academic programs, on the other hand, might do quite well without hospitals, theaters, or fancy football arenas. So, why should a university do research?

via Why Universities Do Research? « WWC.

Entry filed under: Uncategorized. Tags: , .

New APCS trial at UCSD Unemployment high in Tech Sector – NYTimes

11 Comments Add your own

  • 1. Alfred Thompson  |  September 6, 2010 at 10:06 pm

    I heard the head of a private research lab (for a major non US company) once tell a group of academics that their research was basically uninteresting to his company. What they were interested in was “researchers” that were produced by the universities.
    At the same seminar were shown some data that indicated that most of the IP income for all universities in the US was from GatorAide. So apparently some research does bring in some income. Not much of it in computer science though hopes seem to run high.
    Not all researchers become faculty – many go to work for companies. It’s a different sort of research sometimes but the skills developed in university research can be valuable in commercial companies.

  • 2. Alan Kay  |  September 7, 2010 at 6:07 am

    Hi Mark,

    This is a complex issue.

    1. It is really a bad idea to try to pay for graduate research from dinging undergraduate education — this is a sickening practice

    2. Universities are the best place with the best kinds of people to do many kinds of important research that companies wouldn’t touch with a 10 foot pole (but have been benefiting enormously from for many decades).

    3. The funding for good university research has not been there for a long time now (large scale ARPA-IPTO funding a la the 60s).

    4. The paper writing tenure and promotion system is generally specious and (in CS at least) has led to mediocrity and tiny almost empty gestures with little advancement of the state of the art.

    Looked at since 1940 or so, we can see how it was done well and led to enormous advancements in knowledge, technology, and GNP, and we can see how it got corrupted in various ways to the sorry state it is in today.



    • 3. Mark Guzdial  |  September 7, 2010 at 2:30 pm

      Hi Alan,

      The point you make “we can see how it got corrupted in various ways to the sorry state it is in today” is the part that I’m most curious about. Given all of Rich’s well-researched points, why is it that so many universities are growing or starting a research program? If it’s a money-losing proposition, why do more of it? I’m curious as to how we got here. The loss of the ARPA-IPTO funding, as you say, means that we’re scrambling more for smaller pots of money — but that would seem to deter more from going into research. What caused the corruption from the 40’s through the 60’s and 70’s to today?


      • 4. Alan Kay  |  September 7, 2010 at 5:23 pm

        Hi Mark,

        I don’t think I have enough of an historical perspective wrt universities in our field to give a definitive answer.

        I was a grad student during the golden ARPA/ONR etc., years, and was concentrating on our projects at PARC with other former ARPA folks during the 70s after the ARPA funding was done in by the Mansfield Amendment changes.

        One part of this seems simple: the cessation of “Grand Challenge” research in universities correlates most strongly with the change in DOD related funding. The 80s and 90s were also about the commercialization of the inventions of the 60s and 70s, and this led to both large money flows and new jobs.

        And, there’s the Bayh-Dole Act (in the 80s?), which perniciously gave universities title to IP even when it was created via public funding.

        The universities for their own reasons decided they were actually businesses, and started to try to conduct themselves along those lines.

        This could have been catalyzed by a number of factors, including the baby boom. An interesting book “Imposters In The Temple” by a Stanford prof, discusses in some depth what happened to universities as they tried to cope with the baby boom as it happened rather than planning ahead for it.

        When did “publish or perish” start in computing? I don’t know. This was not any kind of an issue in the ARPA community when I was a grad student, or when I was a post-doc at Utah and Stanford. I’m guessing it was already there in other fields throughout this time.

        There were only a few worthwhile places to publish a paper: most people who were doing good work and who wanted to publish, would do it in one of the few national conferences, or in the few ACM publications of the time.

        NSF was not a big factor in computing research from the 60s through today (I don’t know exactly why). They had some good people (like Howard Moraff, who did give Nicholas Negroponte some useful funds).

        I sit on a number of NSF committees and get to watch NSF do its thing. The NSF staffers could hardly be better or more dedicated. The politics at every level is complex, and intertwined with the two main influences on NSF: Congress, and the various scientific communities which have great say about what is important and what not.

        My view is that wrt computing, the field has a far too mundane (even boring and irrelevant) notion of computing, and this affects NSF severely. I have advocated to NSF that could make a huge positive difference to our field if they did a little visionary funding without asking the field whether they should our not. 5-8 years of this would revolutionize how R1 universities (and many others) conduct themselves wrt computing.

        As I said this is complex and needs to be discussed more.

        Best wishes,


  • 5. Alejandro Echeverría  |  September 7, 2010 at 8:22 am

    I have two answers to the question:

    1. I agree with Alfred, one main goal of university research is to generate researchers. As a Ph.D. student myself, I have realized that my research is no what is relevant, is my ability to do research, that can be transfered later into the private sector.

    2. The advancement of knowledge is not instantaneous: there’s a multi-year gap between the initial idea and its possible application in the real world. Specially in technology: I am thinking about the mouse, multitouch, etc. Whatsmore, you have to generate many ideas to have one successful, so I think that if only the private sector was involved in research you would have probably less research projects in general, and surely less research projects that don’t have a direct application in the short run.

  • 6. Steve Tate  |  September 7, 2010 at 8:26 am

    You seem to be implying that the main “benefit” a university gets from a strong research program is the intellectual contribution (or intellectual property rights?) of the research itself. I don’t think that’s so – research is the whole point of graduate education, and that trying to separate “research” from “education” in a doctoral granting program is impossible. Think of it this way: surveys show repeatedly that the key qualities for a person to be particularly successful are to be good at critical thinking, creativity, and communication. So we have a general education curriculum for undergrads that claims to support this, and maybe it does to some extent, but can you honestly say that your average undergrad (not at a top school) is particularly strong at these skills? Now think about what research provides for a student, without thinking about the value of the research results – if a student succeeds at research, they have these skills in spades. So doing the research, even if the results themselves aren’t “valuable”, is an educational experience that is many times the value of an undergraduate education as measured by those desirable skills.

    So where does this leave undergraduate education? Well, I can argue that faculty who are active in research/scholarship are more “cutting edge” and hence more inspirational to students, or that getting undergraduates exposed to or involved in research inspires them to reach higher. Both OK arguments, but maybe the best measure of how convincing I find these arguments is how I take them as a parent of kids getting closer to college age. While I might have different answers for my two kids, I think the ideal combination would be a strong undergraduate-only liberal arts school as an undergrad, followed by graduate school at an R1 university. The half-hearted arguments about the value of research to undergrads? Well, let’s just say I find them less than compelling for my kids, which says something interesting…

  • 7. gasstationwithoutpumps  |  September 7, 2010 at 10:59 am

    “critical thinking, creativity, and communication … – if a student succeeds at research, they have these skills in spades.”

    I have seen many “successful” PhD students with no communication skills and limited creativity. It is particularly a problem in the lab sciences, where a grad student may just be “hands” for a postdoc or faculty member who provides most of the thinking and creativity. In CS, communication seems to be the problem, as many CS grad programs have no requirements for public presentation of research and only limited writing requirements.

    I’m at a public R1 university that has only 10% grad students, so we do involve a lot of undergrads in our research, though not in all fields (majors that attract students from the bottom of the pack tend not to involve their students, as it would be a waste of everyone’s time).

    I would want my son involved in research as an undergrad, and I note that several of the top schools on the list (Caltech, Harvey Mudd, MIT, …) have a reputation for including their undergrads in research.

    Of course, I think that a big chunk of the “produces students that go on to get science PhDs” is a selection effect at admissions, not the result of the training at the college. Almost all the schools on the list have very selective admissions, and it would be interesting to compare the students who attend them to students with similar qualifications coming out of high school who went elsewhere. That won’t entirely negate the selection effect (unless you believe that college admissions officers are making essentially random selection from the pool of qualified applicants), but may be enough to see whether there is primarily a training effect or a selection effect.

  • 8. Rob St. Amant  |  September 8, 2010 at 11:49 am

    I also agree with Alfred about the idea that producing researchers is of economic value. This is pretty hard to fold into some measure of the economic value a university produces, though. There’s a good example close to home at North Carolina State University. James Goodnight graduated with a Ph.D. in statistics in 1972; after a few years as faculty he founded SAS, Inc., which is now the largest privately-held software company in the world, with its headquarters just 15 minutes away. If we were only to count income from IP to the university (leaving aside grants and gifts from the company), I don’t know of any related to this. On the other hand, it’s reasonable to think that if NCSU hadn’t been a research university, the company wouldn’t have happened, at least not here. How should that be counted?

  • 9. Barry Brown  |  September 12, 2010 at 11:46 pm

    The common fallacy among prospective students and their parents is that good research goes hand-in-hand with good teaching. Incoming freshmen flock to the top research institutions, yet I see little evidence that the teaching there is any better than the lower-tier schools. Certainly every college has its shining stars, but the correlation is shaky at best.

    Perhaps universities see this perception among students and rush to start research programs so their teaching reputation can ride coattail.

    • 10. Mark Guzdial  |  September 13, 2010 at 8:25 am

      Actually, Barry, I think research credentials and teaching performance (depending on how you measure it) are negatively correlated. At ITICSE 2003, Lecia Barker had a really interesting poster showing how the higher up the R1 ladder a school was, the lower was the percentage of undergraduate women who graduated. The more research-y you are, the worse you were for women, and I think that’s a pretty important measure of teaching performance.

  • […] heard the argument that the Bayh-Doyle act was the downfall of undergraduate education in America.  By allowing universities to keep the intellectual property rights to sponsored research, an […]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Trackback this post  |  Subscribe to the comments via RSS Feed

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 10,185 other followers


Recent Posts

Blog Stats

  • 2,039,511 hits
September 2010

CS Teaching Tips

%d bloggers like this: