<p>
</p>
<p>He did say that selectivity correlated .837 with PA. He did not indicate all the other undergrad characteristics he investigated. So his results may not be that different from yours.</p>
<p>Note, Collegehelp, that he looked at USNews graduate program RANKINGS, not the NUMBER of graduate students, which was the hypothesis of this thread. This is essentially a PA for graduate programs. It hardly seems surprising that the graduate program rankings would correlate with the undergraduate program rankings. The grad rankings are based on what faculty think of the graduate programs at these universities. This is likely largely determined by what they think of the faculty. Since the same faculty teaches grad and undergrad, one would expect the by-program rankings for undergrad and grad to be correlated. </p>
<p>Since Hicks computed average grad program rankings, he in effect estimated the reputation of the university faculty across all fields. The result would be expected to follow the undergrad PA.</p>
<p>From the grad student numbers provided by Hawkette, it looks like a high correlation between grad ranking and PA, but a much lower correlation between grad student numbers and PA.</p>
<p>His methods certainly leave his results vulnerable to restriction of range, but for his purposes- comparison of Emoryto other elite colleges, this made sense. If he had included lots of much lower ranked places, he would have been open to criticism that Emory should not be compared to such colleges.</p>
<p>In short, he says that a university cannot be much better than its faculty. High faculty reputation attracts top students. He makes a pretty good argument that in order to raise its overall reputation Emory will have to improve its graduate programs. But the only way to do this is to raise the reputation of its faculty.</p>