About this time last year, I asked advice about where a high school senior should consider for college engineering. At the back of my mind, I was figuring that this is a pretty smart kid and both of his parents have PhDs, so finding a place that might get him fired up enough to consider education beyond the bachelor’s degree would be a good idea. But what places really have an environment that inspires students to go on to still-higher education? While I had thought a smaller school would be a good choice, I also bought into the argument that students are inspired by top-notch research around them, arguing for a larger research-oriented institution.
The NSF has done a study on this, and is listing the undergraduate programs whose students are most likely to go on to get a PhD in a science or engineering field. Thanks to Laura McLay for pointing to this article on the subject. Here is the table that lists the top 50 schools in terms of fraction of students who go on to get a science and engineering doctorate:
The results are pretty stunning for a lot of reasons:
- The list is dominated by private schools, with only three public schools in the top 50.
- About half the schools are “Research – Very High”, denoting the most active research institutions (there are three levels of research activity in the Carnegie Classification), with the other half being small undergraduate-oriented colleges.
- The effect is significant, with a factor of seven difference between number 1 and number 50 on the list. The average “Research – High” university has a value of 1.5 (1.5% go on to doctorates), so the lift for number 50 is more than 3, while that for Cal Tech is 23 (a graduate of Cal Tech is 23 times more likely to go on for a science and engineering doctorate than a graduate of an average research university).
- Berkeley (who also has the largest number of graduates with PhDs) is the only large public university on the list. No Michigan (who has the second third largest number of graduates with PhDs) , Georgia Tech, or other large public university is to be seen.
Now this sort of study has limitations. In retrospect, it is not surprising that a school that graduates lots of, say, accountants will naturally have a lower fraction who go on to get doctorates. Most big public universities have honors programs or other structures to nurture those with further educational aspirations, and I am sure that the results for those in these specialized programs look like the results for the schools in the table above.
But if you want to be surrounded by those likely to go on to get doctorates in science and engineering, you should either go to one of the very top private research schools or go to a small private liberal arts college. Carnegie Mellon or Oberlin (or, particularly, Cal Tech), that is the question!
I think this point:
“In retrospect, it is not surprising that a school that graduates lots of, say, accountants will naturally have a lower fraction who go on to get doctorates.”
— has more weight than this post seems to suggest. Table 3 in the NSF report, which displays information on where PhD graduates obtained their undergraduate degree has over 50% of public schools in the top 50 (of course, state schools tend to be large to begin with, so that’s not the whole story either). Michigan is #3 on the list, since Mike pointed us out — but Michigan also has an ENORMOUS student population in humanities and social sciences, so if ratios are to be taken, the determinator will work against us to a great extent.
Notice, on the other hand, how far ahead CalTech and Harvey Mudd are from everybody else in terms of that ratio — not surprising given that almost all undergraduate degrees they grant are in S&E to begin with. (The same might be at play for New Mexico.) Even MIT is pretty far behind the first two — I knew having a degree granting Music department was a source of trouble. 🙂
Agreed on the denominator issue, but Swarthmore, with a total of 8 engineering faculty members (out of 172) and others also seem to have a pretty big humanities and social science program. I’m not surprised MIT and Cal Tech are there (though the ratio is pretty darn impressive) but the large number of graduates from small liberal arts colleges surprises me.
And I pointed out Michigan since Michigan is the most impressive public university I have seen.
I suspect that students with no ambition to go beyond a baccalaureate degree (in any field) are more likely to attend a public university than a private university (typically pricier and often perceived, rightly or wrongly, as more challenging academically). So I’m not surprised that private schools dominate the list.
“And I pointed out Michigan since Michigan is the most impressive public university I have seen.” Working off a pretty small sample? 🙂
I know that many students prefer to attend private schools due to the prestige (or that their parents want them to attend the school). Nonetheless, there are still many prominent public university that can offer great research opportunities, higher education, etc. I guess it would then depend on the preference of the students. (ie get any degree or a special kind of degree, job advancement or become the next Einstein, personal achievement or not…)
Is there a more recent ranking by any chance?
The statistics presented in this table are somewhat biased. As it happens, my institution (“New Mexico Tech) is #15 on the list (and please don’t confuse us with “New Mexico” as in the perennial football also ran up the road in Albuquerque…) The fact that nearly all of our undergraduate students are majoring in subjects that the NSF considers science and engineering definitely biases the results in our favor.
Mike- many smaller engineering/science oriented institutions like NMT, the Colorado School of Mines, Missouri-Rolla, Michigan Tech, etc. do a very good job of undergraduate education in science and engineering.
We’re definitely not “liberal arts colleges”, but we often have small or nonexistant doctoral programs and thus don’t always get classified as research universities. For example, NMT only graduates a dozen or so PhD’s a year, but does over $100 million per year in federally funded research. We’re a Carnegie 2005 “Master’s” institution, but there’s no way to compare us reasonably with most of the regional comprehensive master’s degree granting institutions.
The numbers do surprise overall, but Engineering (see Figure 3 discussion) is the one sub-area that doesn’t fit the main table.
Instead the results pattern is dominated by natural sciences and social/behavioral sciences (which apparently count as “S&E” doctorates, who knew?).
For engineering, the interest of the original student discussed, Figure 3 shows us the research-intensive private school group has 4-5 times the doctorate production rate of the Oberlin 50. In the other two subject areas, the two groups are more evenly matched.
Paul: Not small, but not all-inclusive either! Biased by IE schools (with a smaller bias towards business schools: I think very highly of Michigan State for reasons you can guess).
But it seems that I am only ticking people off (yes, Brian: I know NMT, Colorado School of Mines, etc. etc. do a very good job: we need far more smaller engineering and science oriented places!).
Suffice it to say that I love schools that employ, hire and/or graduate people who read this blog. Perhaps that is a ranking I should be concentrating on.
This is no surprise. If i have a list like this in my country, no public school would be on it. Most public school in my country are left to decay due to the lack of funding and whatever crumbs allocated for it are left for the vultures who calls themselves administrators.
I envy you. Your list still include 3 public schools. Better than none at all.