In an effort to measure the strength of undergraduate (rather than graduate) level CS research I have taken the number of undergrad research awards over the last ten years for both an a major industry research award program (sponsored by Microsoft) and a major academic research award program (sponsored by NSF) and normalized the results by dividing the number of awards by the number of CS and Computational Math majors at each school. For transparency, all the data are provided so the reader can get an idea of the relative size of the CS departments at each school as well as the absolute number of awards and the normalized rank for each type of award. Here are the results:
2016 Undergraduate Computer Science Research Ranking
School…# uSoft…# NSF…Total…# Comp Sci …# NSF …# uSoft…Total
Name…Awards…Awards…Awards…Graduates…/# Grads…/# Grads…/# Grads
Princeton…13…32…45…117…0.039…0.022…0.032
Harvey Mudd…10…13…23…78…0.024…0.026…0.025
Cal Tech…0…15…15…60…0.036…0.000…0.021
Olin College…0…8…8…38…0.030…0.000…0.018
Columbia…10…15…25…126…0.017…0.016…0.017
Harvard…5…14…19…96…0.021…0.010…0.016
Yale…0…10…10…51…0.028…0.000…0.016
Tufts…8…6…14…83…0.010…0.019…0.014
U Rochester…6…1…7…43…0.003…0.028…0.014
MIT…2…40…42…283…0.020…0.001…0.012
UCB…10…32…42…286…0.016…0.007…0.012
UVA…5…10…15…115…0.012…0.009…0.011
Cornell…10…15…25…213…0.010…0.009…0.010
Northwestern…0…6…6…52…0.016…0.000…0.010
Stanford…1…23…24…217…0.015…0.001…0.009
CMU…3…19…22…211…0.013…0.003…0.009
Rice…0…7…7…70…0.014…0.000…0.008
Wash U…0…5…5…53…0.013…0.000…0.008
U Wash…13…20…33…357…0.008…0.007…0.008
Dartmouth…1…3…4…47…0.009…0.004…0.007
U Mass…5…5…10…124…0.006…0.008…0.007
Georgia Tech…3…19…22…288…0.009…0.002…0.006
Duke…0…5…5…68…0.011…0.000…0.006
Brown …5…1…6…103…0.001…0.010…0.005
U Penn…0…7…7…127…0.008…0.000…0.005
UIUC…7…8…15…275…0.004…0.005…0.005
UCLA…0…6…6…163…0.005…0.000…0.003
Cent Florida…0…6…6…165…0.005…0.000…0.003
Virginia Poly…0…6…6…178…0.005…0.000…0.003
UCSD…0…10…10…304…0.005…0.000…0.003
Purdue…0…6…6…225…0.004…0.000…0.002
U Michigan…0…7…7…408…0.002…0.000…0.001
Good feedback. I went back to double check the numbers in College Navigator and they now have the 2015- 2016 graduation numbers listed. I had to go into the raw IPEDS database (kind of a pain) to check the UCB and UCLA numbers that I used - and they check out.
What is interesting is that when one compares UCB’s 2014-2015 CS graduation number (286) to their 2015-2016 CS graduation number (380) there is a huge increase. Does that make sense to you? If the numbers are representative, then they would suggest that the average class size must have increased significantly…
UCLA (and many other schools) numbers have gone up as well, but not by as much. Since the number of awards is looking backwards over the past 10 years, updating the calculation with the latest numbers would actually increase the existing “unfair” bias that this calculation has against schools that have experienced recent growth.
I was surprised at the size of UCLA’s CS department - given the size of the school - but that is why it is important to normalize by the size of the department and not the overall size of the school.
CS enrollment has increased dramatically almost everywhere.
Thought I read somewhere that Cal has 1000 start out in their intro CS sequence a year ago, though obviously not all of them will graduate as CS majors.
Yale had to outsource/merge their Intro CS course to/with Harvard (presumably because both schools are short of money) so the combined course is bigger than Cal now.
CS departments across the country are feeling pressure from increased demand for their major, but the highest pressure point at many schools comes from that fact that an Intro CS course is almost perceived as a core graduation requirement these days.
This pressure is especially acute at schools where the CS department is small relative to the School of Liberal Arts.
At some schools 50% or more of the students sign up for the Intro CS Course.
Tech oriented schools have not felt this pressure as much because a much higher percentage of students were taking the intro CS course to begin with.
Those numbers are probably for only one of the CS majors, since UCB has EECS in its College of Engineering and CS in its College of Letters and Science (L&S CS). See https://career.berkeley.edu/Survey/Survey (“What Can I Do With a Major In…?” links) which show graduation numbers for each major. Of course, some EECS majors emphasize EE and some do a mix of EE and CS, so it is not completely obvious how many EECS majors should be counted as “CS majors”. Upper division EE courses are relatively small compared to upper division CS courses, though.
As others have noted, CS has generally been increasing in popularity rapidly in recent years.
UIUC prioritizes the seats in the CS intro sequence (not including CS101/CS105 which are for non-majors) for CS students, but there’s definitely seats in the intro sequence for non-majors, especially in the spring semester when there aren’t many CS students taking the intro class.
UCB reports both BA graduation numbers and the EECS (BS) graduation numbers in the Association of Engineering Education (ASEE) database. There is also a degree that includes materials engineering, but it is very small.
The BA degree numbers track the IPEDS numbers I have and the EECS numbers are slightly less than the source you provided (because they do not include double majors).
To be consistent with the data used for other schools, the sum of the two numbers should probably be used.
The fact that UCB offers both types of degrees provides an opportunity to look at the relative numbers over time
The BA degree is growing much faster than the EECS degree so just looking at the BA numbers (as I was above) gives a misleading picture of the overall rate of growth of the department.
Note that the EECS major is gated by admission to the school (as a frosh or transfer). The L&S CS major used to admit all L&S students who passed the prerequisites (L&S is the largest division, and admits new students undeclared). Then a 3.0 GPA minimum was used, later raised to 3.3 GPA, to try to prevent enrollment from exceeding department capacity.
The introductory CS course (CS 61A) went from a few hundred to over a thousand students enrolled.
In general, the “NSF award” is the National Science Foundation (NSF) Graduate Fellowship Program (GRFP) which awards fellowships for students in research based Masters or Doctoral programs. There were roughly 13,000 applicants and 2,000 awards in the last round. Applications are reviewed via the NSF Peer Review Process. Three to five reference letters are required so there is a lot of self-selection in the applicant pool. A goal of the program is promoting STEM diversity - so there is some bias toward encouraging underrepresented minorities to pursue graduate research.
In this specific context, the “NSF award” is based on the undergraduate institution that the winner attended - not the institution where the winner attends grad school.
To give credit where credit is due, the NSF award data was sourced from another thread on CC. That data was combined with the “Microsoft award” data that I had gathered and posted on yet another CC thread. Award data was then normalized using the graduation data by major from CDS (as reported on College Navigator). Based on feedback from this thread, I cross-checked some of the cases with the ASEE database. The ASEE database has more detail for engineering based CS programs, but it doesn’t always list non-engineering based programs.
I have normalized NSF awards data for other subjects as well, but extracting it out of spreadsheets into CC legible form is very time consuming (at least based on my knowledge of the CC post editor) due to formatting issues…
This is clearly not a perfect process, but it seems to represent a much better proxy for undergraduate access to leading edge research than using graduate school research rankings (which are skewed heavily toward criteria more relevant to graduate students, not undergrads).
Discussions/critiques of the process/data aimed at promoting better understanding and/or improvement are welcome.
Much of this was done during commercial breaks during football games, so it is a good thing that the Patriots made it the Superbowl!
The “Microsoft Award” is an undergraduate research award program set up by the Computing Research Association (CRA) and sponsored by Microsoft and also Mitsubishi. Applicants need to be nominated by two faculty members and approved by the department chair.