<p>I’ve now gone through and compared Machen’s PA ratings for all 263 schools he was asked to rate, and compared them to the 2009 US News PA scores for these same schools. I have to say, I cam away impressed. Not only are the vast majority of Machen’s scores well within the mainstream of the industry as a whole, but I think they evidence some thoughtful judgments. I don’t agree with all his judgments, and a small percentage of the scores seem to be off base—obviously not enough to significantly damage the schools he low-balls, or to significantly help those he gives high marks. But the pattern is pretty revealing.</p>
<p>Consider this. Some CCers criticize Machen for giving Michigan where he previously served as provost, a 5 while giving the University of Chicago only a 4. But in both cases he’s well within the mainstream of PA raters. Michigan’s overall PA score is 4.4. That means at least 40% of PA raters gave it a 5—as did Machen. </p>
<p>Think about it this way: to get an average score over 4, some significant fraction of the individual scores had to be 5s. If all the scores were 4s and 5s, then 40% of the scores were 5s and 60% were 4s. But we can probably safely assume there were a few 3s, possibly even a few lower (especially if any presidents, provosts, or admissions deans hold the kind of virulent anti-public bias we hear from some on CC). But if some of the scores were lower than 4, then an even larger percentage must have been 5s to keep the average at 4.4. Bottom line: Machen’s 5 for Michigan was well within the mainstream, as at least 40% of PA survey respondents agreed with him. Also note that even if he had gone with the majority and given Michigan a 4 instead of a 5, it’s extremely unlikely that would have made any difference in Michigan’s average PA score. Three people at each of the 263 schools, or 789 persons, were invited to respond. We’re told the response rate is just under half, so let’s assume about 350 actually responded. That means Machen’s vote comprises 1/365th of the total—just a little over 2/10 of 1%. On overage it’s going to require a change of about 35 votes even to move just one tenth of a point, e.g., from 4.4 up to 4.5, or 4.4 down to 4.3.</p>
<p>What about Chicago? Machen gave it a 4; its overall PA score was 4.6. Again, Machen’s in the minority, but well within the mainstream, as we can assume roughly 40% of PA voters gave Chicago a 4 or lower. And again, the likelihood that his individual vote is the tipping point in any movement in Chicago’s PA average is practically nil. </p>
<p>What about Notre Dame? Didn’t he hose them? No, not really. Machen gave them a 3. Their average PA was 3.9. That means at least 10% of PA votes agreed with Machen and gave Notre Dame a PA score of 3 or lower; probably more, because views on Notre Dame vary widely, so there were probably some 5s as well, meaning there were probably more than 10% at 3 or lower to result in an average of 3.9. You might disagree with the score; I do, I’d give them a 4. But that’s beside the point. This score isn’t an outlier, it’s at a still fairly fat part of the tail of the curve in a normal distribution.</p>
<p>What about everyone’s favorite atrocity story, Brown? Machen gave them a 3; their PA average is 4.3, so Machen’s score falls a full 1.3 points below the average. i’ll admit to being quite surprised by this. I’d have given Brown a 4 or a 5, and most PA voters apparently agree with me. But once again, in fairness this has to be seen as another “tail of the curve” story, except Machen’s farther out on a narrower part of the tail on this one. does it reflect ignorance, carelessness, or inattention? I doubt it. I suspect he has his reasons. I especially think that because he scored Dartmouth–in many ways a similar school, smaller and more LAC-like than other research universities, not at all fitting the model of a large, multi-dimensional, strengths-in-all-disciplines university that many people, including Machen, have in mind when they think of a research university. Machen’s view of these schools, although not a true “outlier,” is pretty far out on the tail of the curve; but there’s also a sense in which these schools are themselves pretty far out on the tail of the curve for what even counts as a research university (or “national university” to use USNWR’s parlance). So perhaps it shouldn’t be too surprising that there’s the occasional mismatch in perceptions. But clearly Machen’s view doesn’t hurt them much, as the PA scores of these schools hold up pretty well without Machen’s help.</p>
<p>What about Georgetown? Hosed? Not really. Georgetown’s average PA score was 4.0, but with that average you wouldn’t expect 100% 4s, you’d expect perhaps at least 10 to 20% 5s and a roughyl equal number of 3s. Another tail-of-the-curve story.</p>
<p>Sounds like too many “tail of the curve” stories? I don’t think so. Out of the 263 or so schools Machen rated, there are only a tiny handful like this—in short, the number of stories like this itself reflects the “tail of the curve,” not the predominant story. In only 16 cases was Machen’s score more than 1 PA point away from the school’s PA average. Two of those he scored on the high side—Florida and Utah, the two schools where he’s been President, reflecting an obvious but harmless hometown bias.Three others were in his current home state, where he scored Florida Atlantic, Florida International, and Florida Institute of Technology at 1, whereas each had an overall PA average of 2.1. Hard to say he was egregiously wrong in any of these cases, and he may have had better inside information than the average PA rater, who did not hold these school in high regard in any case. The rest? Brown and Dartmouth we’ve already mentioned. Machen gave Pepperdine a 2 v. its 3.1 average. George Washington, 2 v. 3.4; Clark Atlanta, 1 v. 2.2. Boston College, 2 v. 3.5. Carnegie Mellon, 3 v. 4.1. Lehigh, 2 v. 3.2. Clemson, 2 v. 3.1. Memphis, 1 v. 2.3. Baylor, 2 v. 3.1. Tail of the curve to be sure, but not wildly aberrant, and mostly consistent with the “not-quite-full-strength-university” story I told about Brown and Dartmouth. </p>
<p>Again, you may think he’s wrong, as do, apparently, most people doing the PA scoring. But I’m impressed with the degree to which his views are broadly in accord with the mainstream views of academic administrators—except on this question which affects only a few schools in the survey and certainly doesn’t come close to biasing the results against them, since Machen’s got only one vote and on this question may be a minority of one. Broad consensus doesn’t mean the PA ratings are “right,” of course. But it does suggest they may do a pretty good job of doing what they purport to do, namely to reflect the relative standing of academic institutions in the eyes of their peers.</p>