Which universities are best in sciences?

<p>shizz is right. In LAC, you won't have chance to know those top guys in the fields. But people would argue one will learn more, chances are you will also forget more. For science and engineering, you get to do it to learn and know. This is just my understanding. And I think most academically successful people from LAC are due to their own potential or brilliance. I may be wrong.</p>

<p>Reed College ranks first among all colleges and universities in the United States in the percentage of its graduates who go on to earn a Ph.D. in the life sciences and third overall in all disciplines. Sounds like one can do okay at an LAC if one wants a career in science.</p>

<p>
[quote]
And I think most academically successful people from LAC are due to their own potential or brilliance. I may be wrong.

[/quote]
</p>

<p>On a per capita basis, the liberal arts college grads have a disproportionately high presence is the sciences, whether measured by future PhD production or membership in the National Academy of Sciences (19% of those elected to membership in a recent two-year period were educated at liberal arts colleges, according to Thomas Cech).</p>

<p>Here's a link to Cech's essay explaning the pluses and minuses of LAC science education. Cech is a Nobel Prize winning biochemist, who did his undergrad study at Grinnell, got his PhD at Berkeley, and was on the faculty at U Colorado when he wrote this essay. He's since left Colorado and is head of the Howard Hughes Medical Institute.</p>

<p><a href="http://www.collegenews.org/prebuilt/daedalus/cech_article.pdf%5B/url%5D"&gt;http://www.collegenews.org/prebuilt/daedalus/cech_article.pdf&lt;/a&gt;&lt;/p>

<p>Here's the per capita PhD production in math, sciences, and Engineering for the most recent 10-year period (1994-2003.</p>

<p><a href="http://talk.collegeconfidential.com/showthread.php?t=61288%5B/url%5D"&gt;http://talk.collegeconfidential.com/showthread.php?t=61288&lt;/a&gt;&lt;/p>

<p>I was wondering how long it would take for the literati to jump on a post that sheds negative light on LACs . I never said students from LACs don't do well. Big unis have an advantage over small schools for the aforementioned reasons, something that the smallest class sizes in the world can't make up for.</p>

<p>If you read Dr. Cech's essay, you'll see that large research universities and small undergrad colleges both have distinct advantages in the teaching of sciences. Of course, if your mind is made up, then don't read the essay because you probably won't like it. </p>

<p>The disproportionate representation of LAC grads at the highest levels of science in this country speaks to the efficacy of their science instruction. It would be hard for your post to "shed negative light" on the science instruction at LACs, since the data doesn't support your hypothesis.</p>

<p>Another good source for ranking the top science research universities--nothing on LAC's though. Really not much of a different result then just the list of top $$'s for research. Generally if you rank high in research $$$ you rank highly in NAS fellows and many other factors too.</p>

<p><a href="http://thecenter.ufl.edu/research2004.html%5B/url%5D"&gt;http://thecenter.ufl.edu/research2004.html&lt;/a&gt;&lt;/p>

<p>The results posted imply that 79% of Caltech graduates end up with doctoral degrees. The number, is high, but not that high. </p>

<p>According to the IPEDS completion survey, Caltech conferred 2,023 bachelors degrees in the 10 year period 1988-1997. Allowing 6 years to get a degree, and including all science fields listed by the NSF survey, 730 people with bachelors from Caltech earned doctorates in science in the period 1994-2003. This works out to about 36%. Still probably the highest in the country, but nowhere near 79%.</p>

<p>The problem with using current enrollment is that it is not an accurate reflection of the numbers of people who graduated from a school long enough ago now to have earned their doctorates. Some places have made policy decisions to change the sizes of their classes. Other places experience substantial variations in the numbers of students enrolled as their yields vary and they adjust class sizes in response. This happens at varying times, and ignoring it gives an inaccurate inference of the rate of doctoral attainment. </p>

<p>A further refinement would be to reduce the doctoral rate by accounting for attrition. Since Caltech, famously, fails to hold on to a significant proportion of the brilliant undergrads who matriculate, one would have to increase the number of Caltech students to account for the number of freshlings who entered, never graduated, and are not counted in the IPEDs data. I don't know of a central source of this data. If I use the estimate of Caltech's recent 6 year graduation rate of 88%, then there would have been 2298 graduates if everyone had made it through. This would reduce the rate at which people who enter caltech end up getting doctoral degrees to 32%. I think this will remain the highest in the country. However, some other places with less than 95% graduation rates might fall a few places in rank after this adjustment.</p>

<p>If one is considering the proportion of graduates who end up with doctoral degrees in science, the data, consistent through decades, shows that top LAC graduates are as likely to follow this course as are graduates of top universities. It is interesting that some of the top LAC's on this list have not captured the imagination of the public, or of USNews, as leading schools. Another indication, as if any more were needed, of why prestige and rankings are meaningless.</p>

<p>I wonder how the list would look if one could identify those who ended up with careers in engineering or science. Here the presence of engineering schools at most of the top universities, but few of the LAC's, might give a different picture for those who ended their education at the bachelors or masters level, but are spending their careers in science.</p>

<p>
[quote]
The results posted imply that 79% of Caltech graduates end up with doctoral degrees. The number, is high, but not that high.

[/quote]
</p>

<p>You may have infered 79%, but the post didn't imply that! I was pretty careful to give not only the formula, but also the sources for the data.</p>

<p>Actually, a decent rough guide for percentage of PhDs per graduating undergrad would be to take the ten-year number and divide by 5, figuring roughly a five-year graduation rate.</p>

<p>The best way would be to take the actual graduation numbers offset by five or six years, as you suggest. This would eliminate anomies for colleges that have grown or contracted by unusual amounts over the last ten or fifteen years.</p>

<p>One caution. I think if you do a NSF table on their listed "sciences", you also return Psychology PhDs. I realise that is an on-going debate, but I grouped Psych PhDs with the Social Sciences in tabulating the Science totals. It seems to me that the nature of Psych research more properly belongs with Social Sciences.</p>

<p>
[quote]
I wonder how the list would look if one could identify those who ended up with careers in engineering or science. Here the presence of engineering schools at most of the top universities, but few of the LAC's, might give a different picture for those who ended their education at the bachelors or masters level, but are spending their careers in science.

[/quote]
</p>

<p>The schools at the top of the PhD production list don't change much, with or without engineering included. Actually, a large engineering program should help, since engineering was the third largest field of PhD production over the 10-year period.</p>

<p>Under the assumption that major is determined by the student, and likelihood of getting a doctorate at least in part by the college (how much remains unknown), then correcting for student body composition is done by tallying the number of bachelors degrees conferred in science. Assume that few English majors get PhD's in science (while recognizing that some do exactly this), determine PhD's awarded to people who received their undergraduate degree from each institution divided by the number of science bachelors degrees conferred by that institution 6 years earlier.</p>

<p>This makes it possible to compare a place like caltech, where almost everyone majors in science, to regular colleges, with plenty of social science and humanities majors. This has little effect on places like caltech, MIT, and Harvey Mudd, but it makes the more broadly based colleges not trail these tech schools by as much. It also produces some re-rankings amoung the broad colleges. Swarthmore remains way up there.</p>

<p>Re engineering, I meant people who did not get PhD's at all, but who ended up with science careers. This is common for engineers, but the results will not show up in a list of PhD's.</p>

<p>I did misunderstand your statistic. I gather it is PhD's over the last 10 years, divided by total current undergrad enrollment?</p>

<p>I too exclude Psychology.</p>

<p>
[quote]
I did misunderstand your statistic. I gather it is PhD's over the last 10 years, divided by total current undergrad enrollment?

[/quote]
</p>

<p>Yes. Only because the grunt work was finding and entering a consistent enrollment number. Current enrollment is a conveniently available, albeit imperfect, number. I prefer your offset baccalaureate method and am playing around with trying to get the WebCaspar database to give me that. Alas, it doesn't seem to allow using two offset time periods, so I'm trying to see if I can export the data to two different spreadsheets with the same alphabetically sorted list of schools.</p>

<p>I'm not sure that equalizing for the number of science majors makes sense. To me, one of the key bits of information from these PhD lists is the relative importance of various disciplines within a school. For example, Econ is a huge department at Swarthmore and you see that in the per undergrad production of Econ PhDs, compared to, for example, the undergrad production of Linguistics PhDs. I'm not sure that you want to filter that out of the data by adding a "per Econ major" component. </p>

<p>Do so might be of interest in some sort of arbitrary "ranking" of econ department "value added". But to me, the usefulness of these lists is not so much in establishing any sort of ranking, but to identify the character of a school. A school that produces a lot of PhDs per capita probably has a different focus that a comparably selective school that doesn't. Or, a school that produces a lot of PhDs across all fields is different from one that has strong production in only one or two fields, etc. Or, schools that aren't ultra selective in admissions, but still produce high rates of PhDs might be really interesting to consider for a broad range of potential applicants.</p>

<p>"export the data to two different spreadsheets with the same alphabetically sorted list of schools." </p>

<p>That's what you have to do. But you probably will not be able to keep the sort synchronized across the two lists because there will be some schools that did not produce any PhD's. It is much easier to just identify the 30-50 schools that are interesting in this context, and pull their undergrad numbers from the full list. The vast majority of schools have very low PhD rates, and you can ignore them without changing any inferences.</p>

<p>Normalizing for the number of science majors make sense if you are interested in the number of science PhD's. If you are interested in total PhD's, then no normalization would make sense. The huge gap between Caltech and everywhere else is in part because essentially all their graduates are candidates for science PhD's, while at most colleges only a subset would ever consider such a thing. Even MIT has enough econ graduates to cloud a comparison with caltech if it is restricted to science PhD's.</p>

<p>Angels............head.....................pin. I doubt anyone thinks either Caltech or MIT or Swat is a bad place to go to college--especially if an academic life is your goal.</p>

<p>The original OP asked specifically about Universities.</p>

<p>Afan--you're trying to use statistics out of context. Firstly, Cal-tech is a primarily science/math-based school. Berkeley is multi-faceted. Thus the percent of faculty in NAS for Cal-tech should be higher, because it has less subjects and is more science-oriented. The percent for faculty in NAS for Berkeley does not take into account of the numerous other majors at the university.</p>

<p>
[quote]
The huge gap between Caltech and everywhere else is in part because essentially all their graduates are candidates for science PhD's, while at most colleges only a subset would ever consider such a thing.

[/quote]
</p>

<p>But, that's OK. That tells us something worthwhile about Caltech (not that it should be surprising). It would be just as worthwhile, and perhaps more surprising to the average college applicant, to see the same effect at Harvey Mudd.</p>

<p>One of the things I tell kids considering a place like Caltech or Mudd or MIT is that they better be darn sure they want to stick with science because the focus is so overwhelming. Normalizing on a per science major basis would mask that.</p>

<p>I don't think that any "all-purpose" college or university would feel too bad about slotting in behind Caltech, H. Mudd, or MIT on science PhD production. I mean, it would only be headline news if these three WEREN'T at the top of the science lists.</p>

<p>BTW, the two spreadsheet approach isn't that bad. The lists syncronize pretty well if you pull out the schools that graduated less than 500 kids and produced no PhDs in a ten year period. I don't have time to syncronize the lists right now, but, I will work on it. I don't like looking at just 50 schools. You miss a lot of hidden academic gems, like an Earlham or a Beloit. Again, the headline news is when schools that aren't on the ubiquitous USNEWS hit parade produce PhDs at a high rate. Those schools could well be terrific admissions and/or merit aid values for an academically-minded kid.</p>

<p>
[quote]
The original OP asked specifically about Universities.

[/quote]
</p>

<p>I wasn't aware of a forum rule restricting conversation only to the specific question asked by the OP. I'm rather enjoying afan's "take" on the statistics. His pointing out that the NSF database will provide actual baccalaureate degrees is really useful.</p>

<p>True, but for Caltech, it is astronomical. Correcting for the number of math/science faculty would be the next step if you wanted to make the NAS statistic more meaningful, but I don't have the energy. It will not get Berkeley up to the Caltech level. One would also have to correct for the size of engineering faculty, since engineers tend to be elected to NAE, rather than NAS. So a place where engineers are a large fraction of the science faculty may show up lower on the NAS/faculty ratio, but be just as prominent.</p>

<p>This reinforces my point that counting the number of NAS members, without accounting for the denominator, can be misleading.</p>

<p>Many people erroneously assume that all serious science students go to universities, rather than LAC's. If the OP was asking "what are the best places for an undergrad to study science?" then interesteddad's comments are relevant. If there was some other goal, then this may be a digression.</p>

<p>For the top schools in science, I think:
The Ivy League minus Dartmouth/Brown
Stanford, MIT, UC's, Caltech
JHU, Duke, Rice</p>

<p>in no order</p>

<p>The National medal of science is the highest honor that USA can give to an individual in science fields (even social science, economic science included).</p>

<p>Below are the universities that have won most numbers of national medal of science:
1) Harvard (33 winners)
2) Stanford (30 winners)
3) Berkeley (25 winners)
4) Caltech (23 winners)
5) MIT (21 winners)
6) Princeton (17 winners)</p>

<p>I think these 6 universites, and only these 6, are super strong in sciences.</p>