<p>US News has several components of its ranking system. The Selectivity sub-rank combines 3 measures of selectivity to produce an overall Selectivity ranking: enrolled student SAT scores (50% weighted), % of enrolled students who were in their HS top 10% (40% weight), and acceptance rate (10% weight). </p>
<p>Other than the weights, US News does not directly say how they calculate an overall Selectivity score.... I had to take some liberties in order to do my own calculation so I could combine National U's with LACs....in essense, I added the SAT 25%ile + 75%ile, ranked each category, then weighted the ranks according to USNews weights & added them up, then re-ranked. The intent of the Selectivity ranking, I presume, is to basically measure the challenge of getting admitted...or how picky schools are. </p>
<p>I combined the National University list with the LAC list to form a single ranking. Of course hooks like athletic admits, big-money donors, or demographic preferences, etc, which lower SAT averages and the like should be washed out a bit. Note that the bigger schools can 'afford' (within their ranking) to accept more of these hooks. Also, there is probably some fudgery going on in what's reported to USNews (and the CDS)....I've certainly heard about SAT scores being bent (using accepted vs enrolled for instance) and there is plenty of opportunity to do the same for accepted top 10% stats. </p>
<p>Any case, did this for grins, reporting on the top 75 ranked below.....</p>
<p>[also note that I used data from the 2009 USNews, so the data are a year or so old.]</p>
<p>Very nice. Very useful guide for prospective students.</p>
<p>My only comment is that UC Berkeley is ranked with U Chicago and Middlebury but the Berkeley SATs are about 100 points lower. But, Berkeley has 99% in top 10% versus 83% at Chicago. This is all hard to reconcile.</p>
<p>Berkeley does not superscore, so that will add 20-30 points to the SAT score (/1600) or 35-45 points (/2400). That narrows the gap some…</p>
<p>Additionally, the state of CA is missioned to reach out to students from every type of public school in CA, including those severely underperforming the CA statewide standards. This implies that though good raw material enters (top 4% from these underperforming schools), the test scores reflect generally a lack of careful preparation for the tests (and in most cases, the tests are taken COLD, with zero preparation), and certainly zero professional prepping.</p>
<p>In the end you end up with students performing in the top 4% of their peer group, but the sub-par classroom instruction coupled with lack of test preparation makes the test scores for all UCs seem out of whack.</p>
<p>PC, having gone through the same exercise in the past, I have to admit that there are severe limitations. In particular, the low weight applied to the acceptance rate somewhat skews the conclusions about how hard it is to be admitted. </p>
<p>For instance, take a look at this small group:</p>
<p>7 Washington U
8 Harvey Mudd
9 Penn
10 Swarthmore
11 Stanford </p>
<p>Inasmuch as Harvey Mudd might be the most selective LAC, it is not the hardest to “get in” --and this despite having stratospherically high SAT averages. Now compare Mudd with Stanford! And then compare, Stanford and Penn. Penn ranked in the 9th spot admits students at more than 30 percent in its early round. </p>
<p>Fwiw, the inclusion of the top 10% in the USNews rankings is a clear testament of the organization’s … intention of “leveling the playing field” and boosting a number of schools that would otherwise be ranked lower. </p>
<p>In the end, I think there is greater value in presenting the data without attempting to weigh the three criteria and presenting the thee set of data separately. Of course, such a position also conveys my own bias, as I would discard the 10% factor. :D</p>
<p>On the other hand, the SAT scores offer a glimpse at the relative value of the 10% criteria. It shows that the 99% percentage reported by Berkeley is not directly comparable to the scores of universities and colleges that not only recruit on a national basis but also at schools where rankings are not relevant.</p>
<p>HMC is the hardest LAC to “get in” for the average college applicant. However, you are right in that HMC might not be the hardest LAC to “get in” for HMC applicants because of the huge self-selectivity of HMC. The acceptance rate is high compared to the other selective LACs because of this self-selectivity of HMC applicants.</p>
<p>I’m pretty sure Reed College should be on this list. It’s ranked artificially low because it doesn’t participate in peer assessments, but all that information you listed should be in its common data set. I’d be interested to see where it stacks up against the others.</p>
<p>At multi-college universities which consist of distinct colleges having separate admissions pools:</p>
<p>A conglomerated university wide “selectivity” ranking may not accurately describe the selectivity of the particular college that you are actually applying to at that university.
The differences in selectivity among the various colleges at a particular university can be significant.</p>
<p>An applicant can only apply to a particular college of a university, not some theoretical amalgam of its different colleges. So this aggregate number may have little meaning, or even be significantly misleading, for purposes of estimating one’s chances of admission to that college at the university that they are actually applying to.</p>
<p>At least for the top ~ 20 schools, the % ranked in the top 10% of hs class is a meaningless stat as someone indirectly above brought up in the SAT vs class rank comparisons of UChicago and Berkeley. Some of these schools are rejecting 70%+ of applicants ranked in the top 1% of their hs class; they could easily have 100% in the top 10% if they want. So, to me whether these top schools have 89% in top 10% or 83% isn’t telling of anything other than maybe to what extent admissions is based on non-academic factors; yet USNWR is making it appear as a huge important distinction.</p>
<p>I’d like to see a ranking of standardized values for “(GRE/LSAT/MCAT/etc.) minus (SAT/ACT)” to see what schools do to or for students. The results could be presented in bands according to SAT/ACT. This would be significant, IMHO.</p>
<p>kwu, from just a few data points, almost all colleges are 10-15% more selective for entering 2008 than for entering 2006. I think that rising difficulty tide moves all the boats pretty equally.</p>
<p>I don’t see how this is significant at all in the case of generalist tests like GRE/GMAT/LSAT. These schools aren’t preparing their students for these exams. What med school/law school/other PhD programs/jobs students attain admittance to with the same standardized test score or GPA would go much farther towards determining how schools are helping their students. Unfortunately, this type of info is probably very difficult to separate out and discern anything meaningful.</p>