Better rates of grad school attendance from small LACs than an Ivy?

[Pushin30]

Wow! Thanks for the stats. I just bounced a couple of colleges off my list!

[bclintonk]

A couple of things:

1. interestedad mentioned this in passing, but it bears emphasis: these stats can be very misleading, especially for large, multidimensional universities. About 60% of Michigan's undergrads are in the College of Literature, Science and the Arts---essentially studying the same liberal arts curriculum as students at a LAC. The rest are in pre-professional schools (I won't call them "vocational schools" as intereteddad did); things like engineering, pharmacy, nursing, business, where Ph.D.s are far less likely. You can't assume that all the Michigan undergrads who go on to earn Ph.D.s are in arts & sciences fields, but if you looked just at LS&A grads, the percentage getting Ph.D.s almost certainly would be several points higher, perhaps more on the order of 8 or 9% rather than just under 6%. Also note than in sheer numbers of their graduates earning Ph.D.s, schools like Berkeley and Michigan dwarf the smaller LACs and even the Ivies.

2) The percentages of graduates from schools like Caltech, Harvey Mudd, Swarthmore, and Reed going on to earn Ph.D.s are impressive indeed. But I'm a little jarred when I compare Swarthmore's 21.1% Ph.D. rate with another figure interesteddad cited (post #24): 9% of Swarthmore grads are college professors. That's a high percentage, too, but it means fewer than half of the Swarthmore-launched Ph.D.s end up in academia. Where do the rest go? No doubt some end up with Ph.D.s in engineering, education, and other fields where that degree may be an additional job qualification in non-college/university settings. But we all know how tight the academic job market is in many arts and sciences fields---the fields Swat grads are presumably most likely to pursue. I fear it suggests a significant fraction of Swat grads may be slogging their way through Ph.D. programs only to find no academic job--or at any rate, not an adequate, full-time, tenure-track job---waiting at the end of the rainbow. And if that's the case at Swarthmore, one of our most distinguished academic institutions, I shudder to think how many graduates of less distinguished schools--an Earlham (10%), Beloit (9.6%), or Lawrence (9.5%), say, very good schools all, but surely no Swarthmores---are being led down a primrose path to Ph.D.s that end in professional frustration. I guess at a minimum I'd need to see a whole lot more information about where these people end up before I was persuaded that an extremely high rate of Ph.D. production is an unalloyed blessing.

[mini]

" interestedad mentioned this in passing, but it bears emphasis: these stats can be very misleading, especially for large, multidimensional universities."

But it's more than that. Ph.D. rates are only indirectly related to grad school attendance. True, you have to attend grad school to get a Ph.D. (at least in theory ; but a very large portion of individuals who attend grad school, even in Ph.D. subjects don't go on for Ph.D.s, and there is a whole range of areas from engineering to social work where a Ph.D. isn't the working terminal degree except for academics.

"I shudder to think how many graduates of less distinguished schools--an Earlham (10%), Beloit (9.6%), or Lawrence (9.5%), say, very good schools all, but surely no Swarthmores---are being led down a primrose path to Ph.D.s that end in professional frustration."

There is no evidence that those who enter Ph.D. programs from Swarthmore end up in better post-Ph.D. careers than those from Earlham, Beloit, or Lawrence.
.

[slipper1234]

Wow mini just because your world in Spokane doesn't recognize the value of a top degree does not mean the universe thinks that way. I encourage you to attend your Williams class reunion. My guess is 95% of your classmates would have a different view.

[Already?]

"There is no evidence that those who enter Ph.D. programs from Swarthmore end up in better post-Ph.D. careers than those from Earlham, Beloit, or Lawrence."

I think the real question is, do Swarthmore grads end up in better Ph.D. programs than do grads from Earlham, Beloit, or Lawrence? I see no reason why the Swarthmore B.A., Harvard Ph.D. would have any advantage over the Beloit B.A., Harvard Ph.D.

On the other hand, there is lots of room for individual differences. The already-driven high school student who goes to Swarthmore might have ended up just fine no matter what school she attended, whereas for the still-finding-herself high school student, Swarthmore might be a disaster while a lower-key but still intellectually stimulating Earlham or Beloit might give her the space to blossom.

[cellardweller]

Actually such a study was conducted at least for PhDs in the sciences for the 1990-1995 period.

The study found that while 37% of students with a baccalaureate degree from a highly selective college (by Barron’s selectivity index) earned a PhD form a top 25 PhD institution, 56% of students with a baccalaureate degree from a most selective college earned a PhD from a top 25 PhD institution, a full 50% increase. The proportion of students earning PhDs from top institutions dropped precipitously for the next category of very selective colleges.

So, baccalaureate origin seem to matter a great deal. Two students with equivalent GPAs simply do not have the same probability of enrolling in a top PhD program.

Even among elite colleges there were significant differences as to where the students earned their PhDs.

The proportion of students earning a PhD from a top PhD institution ranged from 62% at Harvard to around 25% at Amherst.

Harvard 62%
MIT 57%
Princeton 52 %
Yale 52 %
Berkeley 48 %
Swarthmore 44 %
Michigan 39 %
Williams 39%
Carleton 28 %
Amherst 25 %

This would seem to indicate that SAT scores of students also do not fully explain the differences in the proportion of students enrolling in top PhD programs. Top PhD institutions including Harvard, MIT, Berkeley and Michigan admit a disproportionate amount of students from their own undergraduates. They also appear to favor undergraduate applicants from other schools with strong PhD programs.

Despite the seeming disadvantage, Swarthmore at 44% and Williams at 39% perform surprisingly well at getting their students into top PhD programs. For whatever reason, Carleton and Amherst not so well. (less of a science focus?)

So, it can safely be concluded that Swarthmore grads are more likely to ean a PhD from Harvard or any top phD program than grads from other LACs, especially second tier LACs. But if you really want to earn a PhD from Harvard, nothing beats going to Harvard as an undergrad!

For the complete study
http://www.ilr.cornell.edu/cheri/con...onf2003_03.pdf

[bclintonk]

^ thanks, this is really interesting

[bclintonk]

^^ . . . except that as I read this study, it was for science and engineering Ph.D.s only. I wonder whether Carleton and Amherst's percentages would be higher if you included humanities and social sciences.

Also, while it's true that Harvard, MIT, Berkeley, and Michigan "disproportionately favored" Ph.D. program applicants who were their own undergraduates, that still meant something under 10% of their Ph.D.s went to their own undergraduate alums. Since their own institution's undergrads are likely overrepresented in their own grad school applicant pool, we actually don't know whether it's statistically advantageous to have gone there as an undergrad. Also note that for each of these Ph.D.-granting institutions, the percentage of Ph.D.s awarded to their own undergrad alums had declined sharply since the 1970s when closer to 20% of all their Ph.Ds went to their own undergraduate alums.

In any event, this study does suggest that nearly half of all Swarthmore grads going on to earn Ph.D.s in the sciences and engineering do so at "top" Ph.D.-granting institutions (defined in the study as those "with 10 or more top programs"). I wonder of those are the same half of Swarthmore alum/Ph.D.s who end up with the academic jobs?

[interesteddad]

I would be very careful about mixing and matching the 9% number (from the alumni office's database of 18,000 alums) and the 21% PhD rate from the National Science Foundation PhD completion records. For example, most of Swarthmore's living female PhDs were not allowed to teach college on the basis of their gender. However, one was head of Astronomy at NASA and oversaw the development of the Hubble Telescope.

Here are the fields for Swarthmore PhDs as a percentage of all Swarthmore PhDs over the most recent ten year period. I've only included those at 1% or more:

19.6% Biological Sciences
9.2% Psychology
8.3% English and Literature
7.4% Economics
6.4% Engineering
5.2% Arts and Music
4.5% Political Science and Public Administration
4.4% History
3.5% Physics
3.5% Sociology
2.9% Chemistry
2.9% Other Humanities
2.9% Non-Science Education
2.2% Foreign Languages
2.1% Mathematics and Statistics
1.7% Other Social Sciences
1.7% Religion and Theology
1.6% Business and Management
1.4% Computer Science
1.4% Anthropology

[cellardweller]

Since their own institution's undergrads are likely overrepresented in their own grad school applicant pool, we actually don't know whether it's statistically advantageous to have gone there as an undergrad.

Harvard alums do well across the board at other top PhD institutions, same with MIT grads, not just at their own institutions.

There is certainly some referral bias, i.e. applicants from the most selective colleges apply in greater numbers to the most selective PhD programs, which is to be expected. Logically, students will apply to PhD institutions where they have a reasonable likelihood of acceptance.

As far as the declining share represented by the top institutions, according to the study it is mostly attributable to the vastly increased number of foreign students as well as the overall reduction of students at elite colleges seeking PhDs.

It is not as if students from the most selective colleges are being displaced by students from less elective colleges. Actually, the spread is increasing not increasing. The study showed a drop from 62% to 56% of students at the most selective colleges earning a PhD from the top institutions a 9.6% decline. At the same time the proportion of students in the next category of highly selective colleges dropped from 44% to 37%, a 15% decline.

So while it is getting harder for a student at an elite college to get into a top PhD institution (because of the foreign influx), it is getting even harder for students from less selective colleges. The authors explanation is that the quality dispersion at elite colleges had narrowed dramatically since 1970, much more so than at second tier colleges. SAT scores across the board have been rising much faster at elite colleges. Many students who would have been admitted a decade or two ago are simply not making it today.

As selectivity at elite colleges has accelerated in the past decade, I am willing to bet that the share of students earning PhD at top institutions from second tier colleges is continuing to decline faster than at first tier colleges. The average Harvard, MIT or Swarthmore student is a higher performance student today than a decade ago.

As far as post-docs getting tenure track positions in academia, the overwhelming majority earned their PhDs at top programs.

[interesteddad]

The average Harvard, MIT or Swarthmore student is a higher performance student today than a decade ago.

I doubt that very much. The rationale for that argument is mostly increasing SAT scores, but that fails to account for the recentering of scores and a bit of SAT inflation with more intensive test prep and multiple test dates becoming more common.

Swarthmore publishes a chart of their historic SAT scores, including both pre-recenter and post-recenter of the 1995 scores for comparison purposes. Overall, the SAT scores are largely unchanged from 1970.

The falling acceptance rates don't necessarily imply better students enrolling. It could just be that the same students are beating out larger numbers of students who wouldn't have been accepted last year or last decade either. Same wheat, more chaff thanks on on-line one-click applications.

[cellardweller]

The spread of SAT scores has narrowed, not necessarily the mean.
You just need to look at the 25th percentile at most elite colleges.

For Swarthmore, as late as 1999-2000 according to their CDS 40% of entering freshmen had a CR SAT score of less than 700 and 44% a math SAT score of less than 700.

In 2007-2008, less than 29% scored less than 700 on the CR SAT and less than 36% scored less than 700 on the math SAT.

In 8 short years, well after the 1995 SAT recentering, there were 27% fewer below 700 CR SAT scorers and 18% fewer below 700 math SAT scores. Hard to argue the spread has not narrowed!

[bclintonk]

Originally Posted by interesteddad

For example, most of Swarthmore's living female PhDs were not allowed to teach college on the basis of their gender.

Really? Most? I find this a little improbable, first because there certainly were a significant number of women's colleges with women on their faculties going all the way back to the 1800s, as well as a number of enlightened coeducational institutions (like Swarthmore) going back just about as far. And certainly since the 1970s, a lot of women have successfully landed themselves in academic positions, although it's certainly been an uphill struggle. Granted, pre-1970s there would have been fewer opportunities. But hey, 1970 is a long time ago now. Are you telling me most female Swarthmore alums with Ph.D.s got them pre-1970? That would suggest extraordinary longevity among female Swarthmore alums; perhaps an unexpected reason to try to steer my D in Swarthmore's direction?

[mini]

Again, there is still no evidence that undergraduate degree, by itself, had any relationship at all to "post-Ph.D" career. All that can be shown is that more students at some schools, self-selected, are more likely to apply to and hence be accepted in Ph.D. programs.

"The study showed a drop from 62% to 56% of students at the most selective colleges earning a PhD from the top institutions a 9.6% decline. At the same time the proportion of students in the next category of highly selective colleges dropped from 44% to 37%, a 15% decline."

The increase? It's kind of obvious. Top graduates of Podunk.

[interesteddad]

The average years to completion of a PhD is now 10 years. So, we are only looking at graguates through the class of 1998 on average. Figure that Class of 1940 is still alive. So, you are probably right...the majority of living Swat female PhDs are now of an age where women were starting to teach college in significant numbers -- say the class of 1975 on. Of course, the numbers are still dwarfed by male professors.

The first female PhD in the United States, Helen McGill White was a member of Swarthmore's first graduating class in 1873.