<p>Here are the top-50 undergrad schools in per capita PhD and Doctoral production from 1994-2003. Rank, followed by name, followed by number of PhDs per 1000 undergrads. This covers all PhDs and doctoral degrees included in the NSF data base. </p>
<p>Per Capita Undergrad Production of PhDs and Doctoral Degrees </p>
<p>Academic field: Biology and Health Sciences </p>
<p>PhDs and Doctoral Degrees: 1994 to 2003 from NSF database </p>
<p>Enrollment from 2004 USNews </p>
<p>Formula: PhDs divided by undergrad enrollment times 1000 </p>
<p>1 California Institute of Technology 125 2 Swarthmore College 103
3 Reed College 92
4 Massachusetts Institute of Technology 84
5 Harvard University 79
6 University of Chicago 62
7 Haverford College 61
8 Yale University 59
9 Kalamazoo College 59
10 Princeton University 59
11 Mount Holyoke College 58
12 Carleton College 58
13 Oberlin College 57
14 Cornell University, All Campuses 57
15 Earlham College 56
16 Stanford University 56
17 Harvey Mudd College 55
18 Rice University 53
19 Wellesley College 51
20 Amherst College 50
21 University of Rochester 48
22 Brown University 48
23 Grinnell College 47
24 Bowdoin College 46
25 Lawrence University 45
26 Pomona College 44
27 Duke University 44
28 Hendrix College 44
29 Williams College 44
30 Johns Hopkins University 42
31 Bryn Mawr College 40
32 University of California-Berkeley 40
33 Bates College 38
34 Wesleyan University 37
35 SUNY College of Environmental Sci & Forestry 37
36 Allegheny College 36
37 Smith College 36
38 Occidental College 35
39 St Olaf College 35
40 College of William and Mary 35
41 Hiram College 33
42 Beloit College 33
43 Macalester College 33
44 Case Western Reserve University 33
45 Dartmouth College 32
46 Knox College 32
47 Hampshire College 31
48 Bucknell University 30
49 Colorado College 30
50 Juniata College 30
51 University of Pennsylvania 30</p>
<p>Note 1: Some have complained that these lists don't provide useful data. Proposed Solution: ignore the lists. </p>
<p>Note 2: Some have complained that these lists don't include Law, MBA, MD, or Masters degrees. Proposed Solution: find the data and make your own list. </p>
<p>Note 3: Some have complained that I should go school by school and selectively remove engineering from one school or music from another, but leave them for still others. Proposed Solution: Be my guest.</p>
<p>Yes. The NSF WebCASPAR site allows you to generate reports from their PhD database. You can select any time from from one year or ten years or all the way back to 1966. You can include a single school, designated schools, groups of schools (private or public) or all schools. You can select whatever departments you like and have the results totalled or displayed in separate columns. I just generated a ten-year report with schools listed down the left side and columns across for each field of study. There's an export feature to save the table as an Excel file. The only thing missing is the undergrad enrollment data necessary to get a per capita number. From a master spreadsheet, it's a piece of cake to generate the individual lists.</p>
<p>That data did not reflect the actual number of graduates. Here’s an updated list. The first column is the total number of undergraduate graduates over a ten year period. The second column is the percentage of the total graduates who completed a PhD in biology or life sciences during the following ten year period (offset by by five years).</p>
<p>Note that using the actual # of grads as a denominator tends to favor schools with poor graduation rates. For example, if you ran the numbers using enrolled freshmen over 10 years, Reed would fall dramatically as they have a relatively low graduation rate. So think of this list as answering this question, “if you have graduated from this school, what are the odds that you have also gone on to get a PhD in bio or life sciences?” The schools at the top of the list have pretty big numbers. As many as one out of every twenty graduates has gotten a PhD in bio. That is going to make for a pretty noticeable hard-core bio/science community at the undergrad level just as it would if 1 out of every 20 grads played in the NFL or became a member of a professional orchestra or became an investment banker.</p>
<p>This is a fairly good proxy for the impact of hard-core biology science geeks in the student body and the relative focus on bio on an undergrad campus. Schools with a more science PhD orientation will tend to do better. Schools with a lot of bio majors will tend to do better. Medical school is missing. It would be nice to be able to do a combined bio PhD and med school since those are the two parallel tracks, but the med school data is simply not publically availalble in the way that NSF has tracked all American PhD recepients every year since the 1920s.</p>
<p>**Percent of PhDs per gradutate
Academic field: Bio and Health Sciences</p>
<p>PhDs and Doctoral Degrees: ten years (1994 to 2003) from NSF database
Number of Undergraduates: ten years (1989 to 1998) from IPEDS database</p>
<p>Note: Does not include colleges with less than 1000 graduates over the ten year period** </p>
<p>
2059 5.39% California Institute of Technology<br>
2599 4.77% Reed College<br>
3657 4.40% Swarthmore College<br>
8270 3.29% University of Chicago<br>
11348 3.08% Massachusetts Institute of Technology<br>
1015 3.05% University of California-San Francisco<br>
17855 3.04% Harvard University<br>
2565 3.00% Kalamazoo College<br>
1335 2.92% Harvey Mudd College
2410 2.82% Earlham College
9260 2.68% Johns Hopkins University<br>
11101 2.60% Princeton University<br>
2773 2.60% Haverford College<br>
4936 2.57% Mount Holyoke College<br>
12941 2.50% Yale University
6432 2.47% Rice University
2598 2.46% Lawrence University
4561 2.46% Carleton College<br>
16662 2.45% Stanford University
7067 2.43% Oberlin College
33736 2.37% Cornell University, All Campuses<br>
3229 2.26% Grinnell College<br>
2041 2.25% Hendrix College
2879 2.12% Bryn Mawr College<br>
3740 2.11% Bowdoin College
5840 2.11% Wellesley College<br>
4179 2.06% Amherst College
3578 2.04% Pomona College<br>
2308 2.04% Beloit College<br>
14669 2.02% Brown University<br>
11830 2.00% University of Rochester
2047 2.00% Long Island University Southampton Campus<br>
6751 1.97% Case Western Reserve University
15531 1.94% Duke University
2361 1.91% Hampshire College<br>
1535 1.89% Ripon College<br>
2966 1.85% SUNY College of Environmental Sci & Forestry<br>
2199 1.82% Knox College<br>
5082 1.81% Williams College<br>
3821 1.78% Occidental College<br>
3989 1.75% Allegheny College<br>
2911 1.75% Philadelphia College of Pharmacy and Science<br>
38488 1.75% University of California-Davis<br>
2462 1.75% Juniata College
6901 1.74% St Olaf College
30559 1.74% University of California-San Diego<br>
4113 1.70% Bates College<br>
3945 1.70% Macalester College<br>
56363 1.69% University of California-Berkeley<br>
12784 1.65% College of William and Mary
1363 1.61% New Mexico Institute of Mining and Technology<br>
1971 1.57% Centre College<br>
1092 1.56% Rush University
7081 1.53% Wesleyan University
1277 1.49% Fisk University
1753 1.48% Wabash College<br>
2640 1.48% Hiram College<br>
13887 1.48% Washington University<br>
2081 1.44% University of Dallas<br>
21761 1.40% University of California-Santa Cruz
7162 1.40% Smith College
<p>“Note that using the actual # of grads as a denominator tends to favor schools with poor graduation rates. For example, if you ran the numbers using enrolled freshmen over 10 years, Reed would fall dramatically as they have a relatively low graduation rate.”</p>
<p>Only if we think that eventual PhD earners drop out at Reed’s average rate. Like Swarthmore, Reed is one of the hardest schools; my guess is that the least capable drop out more often than the most capable, but I have no backing data.</p>
<p>It doesn’t matter. I originally did this chart using current enrollment for each school. Then, I found that the IPEDS database would give me the actual number of graduates over a 10 year period. Schools with a low graduation rate fair much better when you use a smaller denominator (actual graduates). The schools that improved with the smaller denominator were schools like UChicago that have grown significantly in recent years. Using their actual grads is a much smaller denominator. In fact, I went back and redid the numbers because I was scratching my head to understand why Chicago wasn’t looking so hot on PhD production. Didn’t make sense given the nature of the school. It was all wrapped up in their recent growth.</p>
<p>Most schools didn’t change much because graduation rates at the top are universally high and because most of them have only seen incremental growth in enrollment over time.</p>