<p>
</p>
<p>The only reason Harvey Mudd was not on my list, above, is because it falls into a different Carnegie class than either the LACs or research universities shown above. Harvey Mudd is grouped in the “Engineering and Technology” class, not the “Baccalaureate/Liberal Arts I” or "Research I " classes I pulled from the NSF data to generate that list. For the same 5 year period shown above (2006-10), Harvey Mudd alumni earned 30 Math/Stat doctorates. Adjusted for institution size, that is approximately the same production rate as Caltech’s (39 PhDs per 1000 currently enrolled undergrads). So it would be tied for #1 by this measure. </p>
<p>The only other school in the “Engineering and Technology” Carnegie class to produce 10 or more math/stat doctorates for that period was New Mexico Tech. Normalizing by institution size (1565 current graduates), NM Tech’s rate works out to 6.4 math/stat PhDs per 1000 ~current undergrads. That would place it slightly above Williams (6.3 per 1000) and slightly below Princeton (6.6 per 1000) by this measure.</p>
<p>It is quite possible that confounding factors (unrelated to department quality per se) affect these rates (or virtually any other post-graduate outcome metric). So I would not make too much of fine differences in the rank order. On the other hand, I don’t think schools on the above list are there by coincidence. For what it’s worth, all 6 of the research universities on that list also show up among the USNWR top 10 graduate math departments. </p>