<p>There are also departments that will be relatively small at almost every college. A lot of PhD production from small departments is anindication of a pretty serious program.</p>
<p>Yes, but, let me remind you of what you just said:</p>
<p>Quote:
"I'm not interested in that kind of breakdown [Ph.D production as a percentage of majors.] To me, the more interesting use of the PhD stats is to compare a school's ranking across a wide range of fields. To use an extreme example, CalTech is #1 in Physics PhDs, but I would dare say off the bottom of the charts in English PhDs. That tells me something about the relative emphasis/popularity/resources of the two departments and something about the interests of CalTech's students."</p>
<p>Someone posted this report:
<a href="http://www.hhmi.org/bulletin/summer2004/wellspring/producers.html%5B/url%5D">http://www.hhmi.org/bulletin/summer2004/wellspring/producers.html</a></p>
<p>Has anyone noticed that the table that they list there is unadjusted for the size of the schools? In fact half of these schools are almost double the size of a typical LAC...</p>
<p>Oberlin's size is almost 3000. Same for Wesleyan in Conn..3000. Wellesley is 2300. Colgate is 2800. Bucknell is 3500. Furman is 2800. </p>
<p>Can anyone post the adjusted numbers? My guess is that after adjustment Reed will be at the top and other schools will appear on the list like Wooster, etc.</p>
<p>Well, that has sort of been the point of half Interesteddad's posts (both here and elsewhere in CC). My point is why stop there; why not adjust for the number of individual undergraduate degrees in that major? We seem to have arrived at an impasse.</p>
<p>johnwesley:</p>
<p>Sounds like a great idea. I imagine that building a spreadsheet with the number of majors in each department over a ten year period at each of 1300+ colleges and universities will be quite an undertaking. But, I'd be interested in taking a look once you're done!</p>
<p>I'd be happy to e-mail you my spreadsheet with 1300 colleges and the number of grads getting PhDs in each field as a starting point, if you'd like.</p>
<p>Adjusting for "the number of individual undergraduate degrees in that major" will not be the same thing. </p>
<p>It could be that some schools have better cultures for producing science PhDs than PhDs in other fields and this will be and should be preserved in the numbers. Or the fact that some students have more information about about professors in various departments than the outside observers (personality, approachability, willingness to help of professors) may very well be some of the reasons why students major in a certain department and not another. </p>
<p>Sorry, I did not read all messages by other users in the thread...I just noticed that other people had the same comments. </p>
<p>Adjusting for size shouldn't be that difficult...and unless you do that you can't draw conclusions about who the "leaders" are. That table is pretty much useless because it is not ceteris paribus.</p>
<p>This is a lot closer to what I meant:</p>
<p><a href="http://www.rescorp.org/ae/AE-rpt3.pdf%5B/url%5D">http://www.rescorp.org/ae/AE-rpt3.pdf</a></p>
<p>Colgate, Furman, Bucknell practically disappear from that list when you adjust the numbers.</p>
<p>You're correct and the surprising thing is that Reed, Wesleyan, Haverford and Swarthmore are all within a hundred graduates of each other in terms of total baccalaureates awarded in the sciences between 1985 and 1997:</p>
<p>This would tend to validate the Howard Hughes Medical Institute's figures except that Oberlin is perhaps exaggerated a bit because it, indeed, has more total science graduates, nearly 50% more.</p>
<p>You just invalidated the whole field of econometrics with such a comment. </p>
<p>Not to mention that when one looks at numbers in the range of 600s, even a difference of 100 could make a significant impact on a ratio like college grads/phd degrees. </p>
<p>Schools like IC, Macalester, Wabash could have a much bigger impact than that poorly justified report that you keep posting suggests ust so that you can mention the name of your school in it.</p>
<p>johnwesley:</p>
<p>You can't combine the numbers like that. The report you cite for science majors means just that -- majors in science departments. Note that the Howard Hughes report also includes PhDs in Psych and all Social Sciences. Two different things. Comparing apples and apples plus oranges.</p>
<p>Plus, the Howard Hughes numbers have some errors, such as no Engineering PhDs for Swarthmore which does not jive with the NSF PhD tracking numbers that the HHMI chart is based on.</p>
<p>To paraphrase Lee Iacocca (who knew a thing or two about econometrics): if you can find a better chart--quote it.:)</p>
<p>you can paraphrase whoever you want but the fact that correlation doesn't mean causation in the world of science is not going to change. (in this case bigger school size obviously well-correlated with number of graduates and hence people earning phds).</p>
<p>Which is why I'm suggesting (and have been for about four posts, now) that the Hughes data be read in conjunction with the Web Caspar data which tends to, at least, mitigate size as a factor regarding those schools with similar numbers of science baccalaureates. Excuse me, but, this isn't <em>really</em> rocket science. And, yes, Interesteddad, I understand that there are going to be anomalies between two different sets of data, but, I'm less interested in whether every last one of Swarthmore's engineers gets captured or whether too many of Wesleyan's Anthropology majors get included than in painting the broad picture which is really all we have to work with until a better study comes along.</p>
<p>You can get both undergrad degrees granted, by major, and PhD's received, by field by undergraduate school from Webcaspar. If you have lots of energy, you can match up the two reports. Problem is that the undergrad list and the PhD list don't match up right out of the box, and you have to adjust by hand. Interesteddad had the patience to do it, so no one else has to.</p>
<p>If you don't need up to the minute data, then refer to the Weighted Bac Origins study. It shows results for all colleges, weighted by size of the undergraduate class, but does not adjust for number of science majors.</p>
<p>A much more manageable task is to inspect a limited number of colleges, the ones that interest you. Most people care about the elites, so just run those, and perhaps use the WBO to add in a few you would have missed otherwise. If you do it your self, you can adjust for undergrad major easily. But remember that people move between undergrad major and PhD. You don't need a BSE to enter an engineering grad program. </p>
<p>In general, you find places right at the top tend to be very scholarly environments (anywhere from 18-20% and up of students getting PhD's). The next group includes colleges with lots of very smart students, many of whom go to grad school, but as many not to doctoral programs. Think Ivy, Williams, Amherst. Lots of JD's, MD's, masters of engineering in the alum ranks. Lower still, good colleges with, ON AVERAGE, less talented or intellectually engaged students OR state schools with lots of people who cannot afford to be out of the labor market for the time it takes to get a PhD. I have not seen data for PhD's but for MD's the average income of medical students is about double the median family income in the US. </p>
<p>Interesting thing about the WBO is that it shows the results over time. Although the top handful of places have been quite stable, below that, say at the Ivies, the rankings bounce around quite a bit.</p>
<p>Afan -I think cherry picking the schools you are interested in is the way to go; a lot of people use the USNews charts to build lists of their own.</p>
<p>Reed students have one of the highest % of PhDs in Bio, but their med school acceptance isn't that great (57%). What does that mean?</p>
<p>It's probably much easier to get a PhD than to get in to medical school. </p>
<p>The PHD% seems to be a recurring theme on CC. However, it really doesn't tell us ANYTHING at all about the quality of a particular program or the quality of a particular school in relationship to other schools.</p>
<p>Only in the rarified atmosphere of LAC medical school applications would a 57% acceptance rate be looked down upon. One thing, however, you must consider is that your baccalaueate degree of origin is a fairly minor ingredient when it comes to med school; gpa and aptitude test scores (the MCATs) figure more importantly than perhaps in any other form of grad school. A 57% med school acceptance rate from Reed may just reflect fewer high scorers on the MCATs that particular year.</p>
<p>"Reed students have one of the highest % of PhDs in Bio, but their med school acceptance isn't that great (57%). What does that mean?"</p>
<p>It likely means that they haven't pre-screened applicants (so that anyone is allowed to apply), and the "weed-out" that occurs at other schools doesn't happen there.</p>
<p>Since the vast majority of students don't go on to Ph.Ds in the subject in which they major, quality of a program is best measured by the satisfaction of the AVERAGE student in the program. (the best ones are going to do fine wherever they go.)</p>
<p>I don't do stats- but I did notice that for students who applied more than once their acceptance rate was higher- ( 78%)
also it doesn't seperate out students who have undergrad degrees in other fields- just those who applied to med school</p>