USNWR 2009: Looking at the Top Strata II (Top 10% students)

College Search & Selection
1

<p>The publication of the 2009 USNWR College Rankings provides an opportunity to compare schools based on a wide variety of data points. In this and in other threads, I urge the reader to think less about the absolute rankings and more about the nature and value-added of the data point being discussed. </p>

<p>% of entering students ranked in Top 10% , National Universities</p>

<p>99% Caltech
99% UC Berkeley
97% Yale
97% MIT
97% UCLA
96% Princeton
96% U Penn
95% Harvard
95% Wash U
94% Columbia
92% Brown
91% Stanford
91% Dartmouth
90% Duke
90% Georgetown
88% Emory
87% Cornell
87% U Virginia
86% Notre Dame
85% Northwestern
83% U Chicago
83% Rice
82% Johns Hopkins
80% Vanderbilt
73% Carnegie Mellon</p>

<p>% of entering students ranked in Top 10%, LACs</p>

<p>93% Harvey Mudd
91% Swarthmore
89% Williams
89% Haverford
87% Pomona
85% Amherst
85% Bowdoin
84% Claremont McK
83% Davidson
82% Middlebury
80% W&L
78% Wellesley
74% Carleton
74% Hamilton
71% Wesleyan
69% Vassar
69% Oberlin
68% Macalester
66% Grinnell
64% Colgate
63% Smith
62% Bryn Mawr
60% Colby
56% US Naval Acad
55% Bates
47% US Military Acad</p>

2

<p>lol thanks for the list</p>

3

<p>Here's an expanded list that includes the top 50 national universities (putting hawkette's ranking of the top 25 in a slightly broader perspective):</p>

<ol>
<li>99% Caltech</li>
<li>99% UC Berkeley</li>
<li>99% UC San Diego</li>
<li>97% Yale</li>
<li>97% MIT</li>
<li>97% UCLA</li>
<li>96% Princeton</li>
<li>96% U Penn</li>
<li>96% UC Irvine</li>
<li>96% UC Santa Barbara</li>
<li>95% Harvard</li>
<li>95% Wash U</li>
<li>95% UC Davis</li>
<li>94% Columbia</li>
<li>93% Lehigh</li>
<li>92% Brown</li>
<li>92% Michigan</li>
<li>91% Stanford</li>
<li>91% Dartmouth</li>
<li>90% Duke</li>
<li>90% Georgetown</li>
<li>88% Emory</li>
<li>87% Cornell</li>
<li>87% U Virginia</li>
<li>86% Notre Dame</li>
<li>86% USC</li>
<li>86% U Washington</li>
<li>85% Northwestern</li>
<li>83% U Chicago</li>
<li>83% Rice</li>
<li>82% Johns Hopkins</li>
<li>80% Vanderbilt</li>
<li>80% Tufts</li>
<li>80% Boston College</li>
<li>79% Brandeis</li>
<li>79% William & Mary</li>
<li>76% UNC Chapel Hill</li>
<li>76% U Florida</li>
<li>73% Carnegie Mellon</li>
<li>72% U Rochester</li>
<li>69% U Texas</li>
<li>66% NYU</li>
<li>66% Georgia Tech</li>
<li>66% Case Western</li>
<li>64% Wake Forest</li>
<li>64% RPI</li>
<li>61% Yeshiva</li>
<li>60% Wisconsin-Madison</li>
<li>55% UIUC</li>
<li>45% Penn State</li>
</ol>

<p>What's striking to me is that 5 of the top 10 are UCs, and 9 of the top 25 are publics. Indeed, the top is pretty much dominated by publics, Ivies, and Caltech-MIT-Stanford. But there's a huge skew among publics: this is one area where the UCs, Michigan, and Virginia clearly stand apart from other top publics. And once again, some very highly regarded privates are surprisingly far down the list---Northwestern #28, Chicago and Rice tied at #29, Johns Hopkins #31, Vanderbilt #32, Carnegie Mellon #39, Wake Forest #45. Of course, the figures may not be comparable; the top privates may be selecting their student body from a more elite group of high schools where second-decile students may easily be as qualified as top decile students from run-of-the-mill public high schools who end up going to public universities. But it does appear there's almost as sharp a skew among privates as among publics: CHYMPS and the other Ivies skim off the cream, and Northwestern Chicago, Vandy etc scoop up the rest.</p>

4

<p>Many people both in California and OOS are not aware that to be eligible to apply to any campus in the UC system, a student must be in the top 12.5% of students statewide. The OOS requirements are even a bit more stringent. University</a> of California - Admissions California has a three-tiered system with the 9 UCs aimed at the top 12.5% of students, the 23 California State Universities aimed at about the top 30%, and California Community Colleges aimed at students who either want to reduce the cost of their education or who did not qualify for UC or CSU admission.</p>

5

<p>While there are some notable differences among the Top 10% students among the USNWR Top 25, I found the differences much more striking among the LAC Top 25. The range from top to bottom among the Top 25 was 26% among the national universities (Caltech & UC Berkeley at 99% down to Carnegie Mellon at 73%), but expanded to 38%/46% among the LACs (Harvey Mudd at 93% down to Bates at 55% Army at 47%). </p>

<p>Among the colleges ranked in the USNWR Top 25 National Universities, 15 had 90% or more of their enrolling students as Top 10% scorers. When expanded to include the USNWR Top 50 National Universities, this expands to 20 colleges (the additional colleges are all public universities). By comparison, the USNWR Top 25 LACs had only two colleges (Harvey Mudd and Swarthmore) at 90% or higher. </p>

<p>In the case of the publics among the national universities, it seems likely that, whether explicitly acknowledged or not, most make class rank a defining element of a student's application. This would appear to be less the case with the privates, either in the National University universe or the LAC universe.</p>

6

<p>DD would be top 10% in four out of five area HSs.</p>

<p>She attends the fifth; how do colleges control for that?</p>

7

<p>Most privates do not rank. Most publics do. </p>

<p>
[quote]
While 85 percent of public high schools rank students, only 19 percent of private schools rank their students

[/quote]

Fewer high schools calculate class rank for college admissions</p>

<p>Hence these statistics are very likely to be based heavily on public school admittees as students from non-ranking schools will simply be omitted from the data. The Ivies could be (as a speculative example without any evidence) admitting many second- and third-decile private school students with no impact on these figures.</p>

8

<p>descartesz,
I think you may have hit the nail on the head. I had some information that I got from collegeboard about a year ago that listed the % of students who came from public schools. Below is how those schools compared. </p>

<p>The top 7 colleges and 9 of the top 11 for % of students coming from public high schools were public colleges (and this does not even include several publics that didn't provide information, including W&M, U Wisconsin, UCSD, Georgia Tech, U Washington, U Florida, Penn State and U Texas). </p>

<p>% of freshmen coming from public high schools, School</p>

<p>85% UC Berkeley
85% UC Davis
85% UC S Barbara
84% U North Carolina
84% UC Irvine
80% U Michigan
80% UCLA
80% Rensselaer
77% U Virginia
75% MIT
75% U Illinois UC
74% Cal Tech
73% Northwestern
71% Rice
70% Brandeis
70% Case Western
69% J Hopkins
69% NYU
65% Harvard
65% Duke
65% Emory
65% Wake Forest
64% Lehigh
63% Wash U StL
62% Stanford
61% Dartmouth
60% Brown
60% Vanderbilt
60% USC
60% Tufts
59% U Chicago
55% Princeton
55% Yale
52% U Penn
51% Georgetown
50% Notre Dame
49% Columbia</p>

9

<p>Another pointless list of criteria that means nothing about making a University "great". The implied meaning here is that all the other colleges are full of stupid people, or the lower down the list you are the more likely you are to run into stupid people. </p>

<p>The measure of a college academically is more about its faculty, how good at TEACHING they are, how insightful they are, what kinds of course offerings they have. In short the quality of the classes. </p>

<p>Presumably kids in the top10% are very bright. But some of them are as stiff as cardboard. Kids in the top 25% may be brilliant, creative, vocal, helpful and innovative, as much or sometimes more than the others.</p>

<p>And we also know that at a lot of the feeder prep schools, they make their money off of their high tuition and the parents/students expect RESULTS in college admissions (otherwise, why bother?) so grade inflation is a huge problem there. </p>

<p>And finally, I am always suspicious of these kinds of stats.</p>

10

<p>If you take all the scores on these items and standardize them, you can plot them against the overall distribution and compare them to the other measures U.S. News includes. This shows that among the top schools, there is remarkably little variation in retention, top 10%, and grad rate. These three measures really don't differentiate much between the top schools. Of course, that's already apparent just from looking at the non-standardized range, but comparing them to all measures further drives the point home.</p>

11

<p>The fact that schools like UC-Davis/UC-Irvine have 97% (or something like that) of students in the "top-10%" shows you high schools have HUGE range.</p>

12

<p>Given the great disparities in public high school districts across the country and the meaning of Top 10% across these districts, I thought it might be more insightful to consider where the students are coming from and the degree to which colleges draw from the public high schools. </p>

<p>To do this, I created a measurement using the Top 10% data provided by bclinton and the data that I provided for % from public high schools (taken from collegeboard). Taken together, I think that this might be a more accurate indicator of enrolled student strength than a standalone Top 10% number. </p>

<p>What do you think of the results?</p>

<p>Rank , % coming from Public HS/Top 10% , % Public Schools , Top 10% , College</p>

<p>1 , 192% , 49% , 94% , Columbia
2 , 185% , 52% , 96% , U Penn
3 , 176% , 55% , 97% , Yale
3 , 176% , 51% , 90% , Georgetown
5 , 175% , 55% , 96% , Princeton
6 , 172% , 50% , 86% , Notre Dame
7 , 153% , 60% , 92% , Brown
8 , 151% , 63% , 95% , Wash U
9 , 149% , 61% , 91% , Dartmouth
10 , 147% , 62% , 91% , Stanford
11 , 146% , 65% , 95% , Harvard
12 , 145% , 64% , 93% , Lehigh
13 , 143% , 60% , 86% , USC
14 , 141% , 59% , 83% , U Chicago
15 , 138% , 65% , 90% , Duke
16 , 135% , 65% , 88% , Emory
17 , 134% , 74% , 99% , Caltech
18 , 133% , 60% , 80% , Vanderbilt
18 , 133% , 60% , 80% , Tufts
20 , 129% , 75% , 97% , MIT
21 , 121% , 80% , 97% , UCLA
22 , 119% , 69% , 82% , Johns Hopkins
23 , 117% , 71% , 83% , Rice
24 , 116% , 85% , 99% , UC Berkeley
24 , 116% , 73% , 85% , Northwestern
26 , 115% , 80% , 92% , U Michigan
27 , 114% , 84% , 96% , UC Irvine
28 , 113% , 85% , 96% , UC Santa Barbara
28 , 113% , 77% , 87% , U Virginia
28 , 113% , 70% , 79% , Brandeis
31 , 112% , 85% , 95% , UC Davis
32 , 98% , 65% , 64% , Wake Forest
33 , 96% , 69% , 66% , NYU
34 , 94% , 70% , 66% , Case Western
35 , 90% , 84% , 76% , U North Carolina
36 , 80% , 80% , 64% , Rensselaer
37 , 73% , 75% , 55% , U Illinois
na W&M
na Boston College
na Carnegie Mellon
na Georgia Tech
na Penn State
na U Florida
na U Rochester
na U Texas
na U Wisconsin</p>

13

<p>I don't understand at all what you are doing with those numbers, why would you do this? Doesn't make sense to me, the first number has no meaning. As I understood it, Descartesz's post was about how most private high schools don't rank, thus, the % public hs figure should not impact the percent in top10% figure much...</p>

14

<p>keefer,
I am struggling to find a way to improve upon the Top 10% number because right now, it's clearly pretty useless when the first 7 national universities are all public and 5 of those are located in California. </p>

<p>As for the data itself, I don't think it is quite as straightforward as you might think. For example, 85% of Northwestern's students indicate a Top 10% rank and 73% of them came from a public high school...but only 46% of them actually submitted a class rank.</p>

15

<p>okay so what? it's not a perfect measure of selectivity, that's why you have other measures. For most of those top schools, scores/sat/class rank won't get you in, it's the essay and recommendation letters that make the difference. I can see exactly how Northwestern would have only 85% in top 10%, and I can see how it can be even lower. be careful when you are trying to randomly mess with two different sets of data to fit your idea of selectivity.</p>

16

<p>


</p>

<p>So your assumption is that the higher the percentage from private HS, the stronger the student body? That strikes me as nothing more than rank prejudice. </p>

<p>That seems to be pretty much what your first column represents, though. Consider a hypothetical Prepster College which has only 20% of its student body coming from public high schools, and a mediocre 50% (of those reporting HS class rank) in the top 10% of their HS class. Then dividing 50% by 20% (which seems to be your column 1 methodology), you'd get a column 1 score of 250%, easily outdistancing Columbia for the #1 position in "student body strength" by your metric. Absurd. Or consider this: Harvard (95%) and Penn (96%) have virtually identical scores in the top 10% ranking, yet in your column 1 ranking of "student body strength" Penn easily outscores Harvard, 185% to 146%. Why? Well, because only 52% of Penn's students come from public high schools, while foolish Harvard dilutes its "student body strength" by taking 65% from public schools. Nonsense.</p>

<p>And by what logic do you think the current top 10% number is defective, just because so many publics score high on that measure? As alememom points out (post #4), ranking in the top 12.5% of one's class is a minimum requirement even to be considered eligible for admission to a school in the UC system; everyone below that is steered toward a Cal State or a community college. Consequently, it shouldn't be surprising that the UCs as a group rank very high in the percentage of their student body ranking in the top 10% of their HS class; it just means in their admissions criteria, unlike most other schools, class rank is given overwhelming weight. Nothing wrong, inaccurate, or misleading about that. Similarly, Michigan scores quite high in the top 10% ranking, but if you look at what they say about their own admissions criteria, they make it abundantly clear that they weigh HS grades/class rank far more heavily than any other factor, with SAT/ACT scores a distinctly secondary factor. Other schools choose to weigh SAT/ACT scores more heavily, and apparently give less weight to HS grades/class rank. They're perfectly entitled to do that, but they shouldn't be surprised when they don't score as high in the top 10% rankings. It doesn't mean the top 10% rankings are somehow defective or misleading. Some schools think (with some empirical evidence to back it up) that high school grades, not SAT scores, are the single best predictor of college performance. Those schools give more emphasis to high school grades, and they'll come out higher on this particular metric, one of many US News uses. That's no more unfair or misleading than measuring SAT scores, which schools like the UCs or Michigan could with just as much justification say are unfair or misleading since they're not the primary factor these schools rely upon to measure their own student body strength.</p>

17

<p>descartes/hawkette:</p>

<p>Less than half of Calif high schools actually rank (according to LA Times). </p>

<p>bclintock: Class rank is NOT(officially) an admissions criteria to any campus, so each UC campus figure is an "estimate." [Yes, top 4% elc can be noted, but that's a separate category.]</p>

<p>The UC target is top 12.5 of the state population, not individual HS.</p>

18

<p>IMO, the single most valuable indicator of predicting college success is a high school student’s performance in college preparatory classes. Not standardized test scores, not GPA and certainly not class rank. Also, I am not alone in that view. The National Association of College Admission Counselors (NACAC) weigh the various factors as follows:</p>

<p>Considerable Weight , Moderate Weight , Limited or No Weight</p>

<p>75.9% , 17.4% , 6.7% , Grades in college prep courses
61.5% , 25.3% , 13.2% , Strength of curriculum
60.4% , 27.9% , 11.7% , Standardized Test scores (SAT, ACT)
51.2% , 36.4% , 12.5% , Grades in all courses
27.9% , 30.6% , 41.5% , Essay and/or writing sample
23.1% , 38.6% , 38.3% , Class rank
21.2% , 40.7% , 38.0% , Counselor recommendation
20.8% , 31.2% , 48.1% , Student's demonstrated interest
19.5% , 41.1% , 39.3% , Teacher recommendation
10.4% , 23.1% , 66.5% , Interview
7.6% , 37.0% , 55.4% , Extra-curricular activities
7.6% , 23.5% , 68.8% , Subject test scores (AP, IB)
6.3% , 13.4% , 80.4% , State graduation exam scores
5.2% , 8.5% , 86.3% , SAT II scores
2.9% , 21.5% , 75.5% , Work</p>

<p>As for the mathematical calculation involving % of public high school students and Top 10% students, I agree that there are many weaknesses. I hope you would agree that, taken on its own, there are also significant weaknesses in the single measurement of Top 10% scorers. I am asking if maybe the mathematical calculation might be better. Perhaps the results are a mathematical accident, but IMO the results look more in line with how most college observers would view the competitiveness of the student bodies of these colleges. </p>

<p>Which Top 20 looks like a better representation of the colleges with the strongest student bodies?</p>

<p>Ratio of % of Public School students/Top 10% scorers, % of Public School students, Top 10% students, College</p>

<p>1 , 192% , 49% , 94% , Columbia
2 , 185% , 52% , 96% , U Penn
3 , 176% , 55% , 97% , Yale
3 , 176% , 51% , 90% , Georgetown
5 , 175% , 55% , 96% , Princeton
6 , 172% , 50% , 86% , Notre Dame
7 , 153% , 60% , 92% , Brown
8 , 151% , 63% , 95% , Wash U
9 , 149% , 61% , 91% , Dartmouth
10 , 147% , 62% , 91% , Stanford
11 , 146% , 65% , 95% , Harvard
12 , 145% , 64% , 93% , Lehigh
13 , 143% , 60% , 86% , USC
14 , 141% , 59% , 83% , U Chicago
15 , 138% , 65% , 90% , Duke
16 , 135% , 65% , 88% , Emory
17 , 134% , 74% , 99% , Caltech
18 , 133% , 60% , 80% , Vanderbilt
18 , 133% , 60% , 80% , Tufts
20 , 129% , 75% , 97% , MIT </p>

<p>Top 10% scorers, College</p>

<ol>
<li>99% Caltech</li>
<li>99% UC Berkeley</li>
<li>99% UC San Diego</li>
<li>97% Yale</li>
<li>97% MIT</li>
<li>97% UCLA</li>
<li>96% Princeton</li>
<li>96% U Penn</li>
<li>96% UC Irvine</li>
<li>96% UC Santa Barbara</li>
<li>95% Harvard</li>
<li>95% Wash U</li>
<li>95% UC Davis</li>
<li>94% Columbia</li>
<li>93% Lehigh</li>
<li>92% Brown</li>
<li>92% Michigan</li>
<li>91% Stanford</li>
<li>91% Dartmouth</li>
<li>90% Duke</li>
<li>90% Georgetown</li>
</ol>

<p>BTW, according to its CDS, U Michigan only assigns a “considered” weighting to class rank while assigning an “important” weighting to standardized test scores (along with GPA, recs and essays). The only factor that ranks “very important” is the rigor of the secondary school record.</p>

19

<p>^ Fair enough. UC eligibility requirements are complicated, and the way I put it was a gorss and probably misleading oversimplification. As I understand it, there are three ways to become UC-eligible. 1) "Eligibility in a statewide context," basically ranking high enough on an objective index of GPA combined with standardized test scores; 2) "eligibility in a local context," i.e., being designated by your HS as one of the top 4% in your class; or 3) "eligibility by examination alone," scoring very high on the ACT or SAT I & II without regard to GPA or class rank. Being eligible doesn't guarantee acceptance at any particular UC campus but the UC system does presently guarantee admission to one of the UC campuses to all UC-eligible California residents. Bottom line, though, HS grades---although not class rank per se (except for "eligibility in the local context")---play a very strong screening role in the UC admissions process. I think it's fair to say that with relatively few exceptions for very high ACT/SAT scores, the UC campuses are making admissions decisions from among an applicant pool that is already heavily selected toward those at the top of their HS class; consequently it's not surprising that the UCs end up with such a high percentage of top 10% stduents.</p>

20

<p>^^ hawkette,
I don't think your ranking is even a "mathematical accident"; I just think it's wrong. No way are the students at Columbia and Penn that much better than Harvard, Stanford, MIT, and Caltech; the only reason you have them so high is that those schools have a higher percentage of private school kids. Nor does Notre Dame beat out Harvard and Stanford (on your index by a wide margin, because so many Notre Dame students went to private, mostly Catholic high schools, some very good, some less so). I just think your index is a mess, fundamentally flawed.</p>

<p>I would never argue that percentage of students in the top 10% of HS class, taken by itself, is a good measure of student strength. I don't think US News claims that, either. That's why they also consider SAT/ACT scores as well as selectivity---and both of those measures taken by themselves also have serious flaws. Combining them, as US News does, doesn't necessarily correct for the defects of each of the components; it may only compound them. Bottom line, I think "strength of student body" is almost impossible to quantify, and casts a serious cloud on this whole ratings nonsense. </p>

<p>I do agree that performance in college prep-level core academic HS courses is the best predictor of college performance. That's essentially what Michigan measures in considering HS GPA: they look only at 10th and 11th grade performance in core academic courses, without weighting, to come up with an objective number. That objective number is combined with SAT/ACT scores and an assessment of the "rigor" of the HS curriculum as well as class rank, if any, to establish a "rating" of "secondary school academic performance," the most important rating in their overall assessment of an applicant's qualifications. It should come as no surprise that there will be a strong correlation of this "rating" with class rank, not because class rank is a big factor in the rating itself, but because grades in core academic subjects will tend to be closely correlated with class rank.</p>