2008 vs 1999: What’s changed in the USNWR data? Who’s hot and who’s not?

College Search & Selection
241

<p>Alexandre,
I share some of your concerns about the comparisons of public universities vs privates, but mostly because of the typically much larger size, their typically more limited faculty & financial resources and their certainly different mandate/responsibility to the citizens of their states. These differences can lead to some good things (among other things, likely broader assortment of areas of study occur with larger size) and some bad things (among other things, lower overall student quality, larger class sizes). </p>

<p>I go back and forth on the issue of whether publics and privates should be ranked together. One danger with separate rankings would be the potential message that public schools are inferior to privates and can’t compete. But the way that the USNWR now does its rankings, some statistics clearly work to the benefit of publics (high weight assigned to top 10%, low weight to acceptance rate, very high weight to PA scores) while others are negative and not particularly relevant (Alumni Giving). </p>

<p>For your comments about SATs and their use by publics and privates, I suspect you are correct that publics assign it less weight, but the privates likewise are not solely guided by this number. For example, U Penn only accepted 26% of those who scored 750 or above on their Critical Reading and only 21% of those who scored 750 or above on the Math section. Thus, they REJECTED 74% of those at this level in CR and 79% in Math. Clearly, if the primary determinant for private school admissions was SAT scores, a highly competitive college like U Penn has ample scope (and probably greater than all publics) for boosting these numbers. </p>

<p>(As an aside re your comments about U Michigan and its SAT policies, you might want to check with hoedown and/or others at the school. My understanding is that, while the admissions policies do not superscore, the reporting on the Common Data Set does and these are the numbers that are reported publicly.)</p>

<p>For the public universities, in terms of selectivity, I did some calculations for 2008 vs 1999 to see if their position relative to one another had changed much. Using the same formula that USNWR now uses to calculate this (50% SAT scores, 40% Top 10%, 10% Acceptance Rate), here are the results:</p>

<p>1999
1. UC Berkeley (score of .99)
2. UCLA (.96)
3. U Virginia (.91)
4. U North Carolina (.82)
5. U Michigan (.77)</p>

<p>2008
1. UC Berkeley (score of 1.00)
2. UCLA (.97)
3. U Virginia (.92)
4. U Michigan (.91)
5. U North Carolina (.87)</p>

<p>Based on this, one can see that the numbers and order change only marginally for most of the schools, but U Michigan is the exception as it made large statistical gains over the last ten years. (Note: My calculations differ slightly from USNWR as the benchmark in each category was the leading public and thus the measurements were made against that rather than against the top university (Harvard) for selectivity.) Congratulations to U Michigan.</p>

242

<p>Or else Penn is such a TTT that it realizes that it has to reject any genuinely strong students to yield protect.</p>

243

<p>That's an interesting theory on U Penn that, given its very high yield, is not completely implausible. I'll be interested to see if some of the U Penn folks can provide more insight on that.</p>

244

<p>I don't have any data about their UG admissions, but the institution is known for gaming the numbers, and they certainly do it in the law school. From LawSchoolNumbers, which isn't perfect, but has no reason to be inaccurate in this case, here's how people with 3.6-3.9, 177-179 (range around my numbers, if it seems arbitrary) did:
Harvard: 22 accepted, 1 rejected, 2 waitlisted, 1 deferred <a href="http://www.lawschoolnumbers.com/search_schools.php?action=search&school_code=0008&lsat1=177&lsat2=179&fee_waiver=0&status=0&gpa_lsdas1=3.6&gpa_lsdas2=3.9&attending=0&cycle=4&gpa_degree1=&gpa_degree2=&withdrawn=0&program=1&index1=&index2=&state=0&application_type=0&scholarship1=&scholarship2=&sex=0&multiple_lsat=0&urm=0&x=31&y=8&international=0%5B/url%5D"&gt;http://www.lawschoolnumbers.com/search_schools.php?action=search&school_code=0008&lsat1=177&lsat2=179&fee_waiver=0&status=0&gpa_lsdas1=3.6&gpa_lsdas2=3.9&attending=0&cycle=4&gpa_degree1=&gpa_degree2=&withdrawn=0&program=1&index1=&index2=&state=0&application_type=0&scholarship1=&scholarship2=&sex=0&multiple_lsat=0&urm=0&x=31&y=8&international=0&lt;/a>
Penn: 4 accepted, 4 waitlisted, 4 deferred
<a href="http://www.lawschoolnumbers.com/search_schools.php?action=search&school_code=0027&lsat1=177&lsat2=179&fee_waiver=0&status=0&gpa_lsdas1=3.6&gpa_lsdas2=3.9&attending=0&cycle=4&gpa_degree1=&gpa_degree2=&withdrawn=0&program=1&index1=&index2=&state=0&application_type=0&scholarship1=&scholarship2=&sex=0&multiple_lsat=0&urm=0&x=52&y=9&international=0%5B/url%5D"&gt;http://www.lawschoolnumbers.com/search_schools.php?action=search&school_code=0027&lsat1=177&lsat2=179&fee_waiver=0&status=0&gpa_lsdas1=3.6&gpa_lsdas2=3.9&attending=0&cycle=4&gpa_degree1=&gpa_degree2=&withdrawn=0&program=1&index1=&index2=&state=0&application_type=0&scholarship1=&scholarship2=&sex=0&multiple_lsat=0&urm=0&x=52&y=9&international=0&lt;/a&gt;&lt;/p>

<p>If Penn thinks you will get into Harvard, whether law school or undergrad, it will think twice about accepting you.</p>

<p>Of course, it's worth mentioning that the statistics you posted don't prove that they are rejecting anyone particularly qualified. The rejects could have low scores on the other section and/or low GPAs (I know many people with 800 math, 600 verbal and 3.2 GPAs or the like).</p>

245

<p>My brother and most every other kid I know who got into a non-Wharton school at Penn had far lower than average scores for Penn and its peer schools (lower Ivies and little Ivies).</p>

246

<p>Based on statistics, the theory that Penn yield-protects in the 750-800 SAT range is highly doubtful, given that:</p>

<ol>
<li><p>the 75th percentiles for ENROLLED (as opposed to accepted) students in Penn's Class of 2010 were 740 on the SAT CR and 770 on the SAT M;</p></li>
<li><p>the distribution of acceptance percentages for Penn's Class of 2011 were as follows:</p></li>
</ol>

<p>(SAT range/% accepted with CR score in that range/% accepted with M score in that range/% accepted with W score in that range)</p>

<p>750-800 26% 21% 27%
700-740 20% 18% 21%
650-690 16% 15% 15%
600-640 14% 14% 12%
550-590 8% 10% 8%
500-540 4% 4% 4%
Under 500 1% 1% 1%</p>

<p>If Penn were truly yield-protecting in the 750-800 range, there would be lower acceptance percentages in that range, instead of the steadily increasing percentages that actually occur as the ranges increase.</p>

<p>Obviously, as these stats show, there are some applicants accepted who have scores lower than Penn's (or other top schools') averages or medians. However, this is true at most top schools, which take a "holistic" apporach to admissions and don't just admit based on SAT scores.</p>

<p>The bottom line, though, is that the middle-50% range of SAT scores for students ENROLLED at Penn are comparable to those at similarly ranked schools (non-HYP Ivies, etc.). If Penn were yield-protecting any more than, e.g., Columbia, Brown, etc., Penn's ranges would be lower.</p>

247

<p>Hawkette, my point was that there isn't an appreciable difference in the quality of the students or resources between elite publics and elite private universities. Obviously, Cal, Michigan and UVa do not have student bodies comparable to Caltech, Harvard or MIT. However, I do not see a significant difference between the student body at Cornell or Northwestern and Cal or Michigan. And class sizes do not vary that much between most elite private universities and the elite publics. Statistically, the way universities chose to report SAT scores varries. A student takes the SAT twice, getting a 650 M and 740 CR the first sitting and a 730 M and 640 CR in the second sitting. A private university would report a 730 M and 740 CR (1470 combined) in both the Common Data Set and to the USNWR for that student. A public university would report an a 650 M and 740 CR (1390 combined) in both the Common Data Set and to the USNWR. Bottom line, state schools really do not care about SATs. It is deeply rooted in their admissions philosophy that SATs are not a good indicator of ability. Whether one choses to agree with that philosophy is another matter altogether.</p>

248

<p>
[quote]
It is deeply rooted in their admissions philosophy that SATs are not a good indicator of ability. Whether one choses to agree with that philosophy is another matter altogether.

[/quote]
</p>

<p>Alexandre, while it's true that "Statistics are always fun to look at, but to draw any conclusion from those statistics can often be misleading," I also believe it's misleading to attempt to draw any conclusion about the philosophy of admissions by public schools. One has only to spend some time studying the changes in admission policies at Michigan, UT-Texas, and especially at the UC system to understand that there is little philosophy, little logic, little intelligence, but a hige amount of politics involved. While we can excuse them for being hostage from the stated purposes of serving their state constitiency, glorifying their admission policies with terms such as "deep rooted" philosophies flies in the face of the truth. </p>

<p>As far as valuing the SAT, you may remember that, even in the darkest days of the "dispute" between Atkinson and Caperton, dropping the SAT was never, ever a possibility at the much misguided UC. The reality is that, if the SAT is not a good indicator of "abilities," its combination with other metrics make those same metrics ... stonger. For the overwhleming majority of institutions, the use of the SAT will remain as prevalent as ever, short of simplistic policies such as the "top-ten" percent rule in Texas. </p>

<p>For all the talk about holistic approaches, we will always have schools that are more or less forced to make admission decisions in a "paint-by-the-numbers" format. You know which ones fall in that group! We are watching them on national TV ... right now.</p>

249

<p>Alexandre,
We don't agree on the student quality of even the top publics vs the top privates. If one objectively looks at the data, there is a subset that statistically overlaps, but having an overalapping of students is not an equivalence of student quality. If one wants to conclude that there is no appreciable statistical difference between a Northwestern and a U Michigan, then one must also conclude that there are not appreciable differences between U Michigan and schools ranked far below it (Clemson, SMU, Northeastern, etc). IMO neither comparison holds water. </p>

<p>Also, I believe you are wrong in how the SAT is reported by public institutions in the CDS (it is superscored) although I agree that many state colleges will not use this in admissions. As for their admissions policies and standardized test scores, my understanding is that it varies by state and certainly there is not a state school philosophy that these are not a good indicator of ability. To my knowledge, none have dropped standardized test scores as part of their application process. In fact, the vast majority that I am familiar with list standardized test scores in their CDS as either "Very Important" or "Important" and not as "Considered" or "Not Considered."</p>

250

<p>A blanket statement that "state schools really do not care about SATs" is simply incorrect. There are a variety of scholarships--tuition reductions, if you prefer--at my state's various public institutions based on nothing other than standardized test scores.</p>

<p>Furthermore, I personally know quite a few public university administrators, at a variety of state schools, and I have yet to find one who thinks there is no predictive value to SAT/ACT scores. Much depends on the major field of study.</p>

<p>I believe the reason many state schools do not superscore is because they see little value in encouraging multiple test taking. It isn't that they don't value the information, it is because they believe the information is debased by giving advantages to those who can afford multiple retakes. The private schools should follow suit. I think we can all guess why they do not.</p>

251

<p>Xiggi, yes the UCs use SAT scores. They use them in a similar way as Amherst uses SAT scores. If you come from a privileged background, you better have high SAT scores. If you don't come from a privileged background, other parts of the application can help you. </p>

<p>The way Michigan has used test scores is exactly the way Alexandre has stated. This came up during the AA trial.</p>

<p>Using average SAT scores as the measure of a student body, what do you think Xiggi? If all those students used the Xiggi method (which I think works :)) the SAT scores would overstate those students intelligence. :) Or do you think all SAT scores are the same no matter how much tutoring a person has?</p>

<p>Anyway, the average SAT scores as a measure of the student body is not my favorite method. ;) For one, it doesn't tell you how many strong students there are at a school. I read the list of schools with the strongest students and the schools with the largest amount of strongest students are left off the list.
I wonder what happens to these strongest students. A school with 8,000, 10,000 , 12,000, more than 12,000 students (as measured by SAT scores). I guess according to some, they are going to schools with a large group of weak students. So what happens to these strong students at these schools. Are the curriculums watered down? Or are these strongest students challenged? If it is the latter, than the argument that smart students should go to the schools with the highest average SAT scores doesn't hold water. </p>

<p>This argument that poor people don't work hard enough or aren't intelligent is rather ignorant and sickening to me. What do you think Xiggi? One only needs to look at what is happening in China and India to see that argument as weak.</p>

<p>Which gets me back to average SAT scores. China's per capita wealth is not going to equal the United State per capita wealth for a long time. If China's per capita wealth were SAT scores, China's average SAT scores are lower than the US. This doesn't mean China is not going to be a wealthier nation than the US or that China is not going to have more wealthy people than the US. If the US focuses on average or per capita wealth of China, it is going to find itself owned and controlled by the wealthy Chinese. ;)</p>

252

<p>hawkette:</p>

<p>There is only one spot on the UC app for one SAT Reasoning Test score to be reported so its likely their CDS report is not superscored.</p>

253

<p>I said earlier that Peer Assessment scores agree with other data in US News in the sense that Peer Assessment can be shown to be a sum of other US News data, weighted in a certain way. There are lots of ways to demonstrate this. I used:
faculty resources rank
classes over 50
SAT 75th percentile
financial rank
actual graduation rate</p>

<p>The exact formula is derived mathematically and will probably seem strange to everybody, but it is the result of "multiple regression analysis". (The "R-square" was .85, in case anybody is interested.) The asterisk means multiplication. It doesn't agree exactly but it is pretty close.</p>

<p>The formula:
peer assessment prediction = 26.49134-(.42506<em>freshman retention)+(.05936</em>classes over 50)-(.00224<em>financial rank)
+(.00001881</em>freshman retention cubed)-(.0000377<em>classes over 50 cubed)+(.000000000245426</em>sat 75th percentile cubed)
+(.000000749913*actual graduation rate cubed) </p>

<p>Here are the results of using this formula:
school abbrev, predicted peer assessment score, actual peer assesment score</p>

<pre><code> 1 Harvard 4.9 4.9
2 Massachu 4.8 4.9
3 Princeto 4.7 4.9
4 Yale Uni 4.6 4.8
5 Stanford 4.6 4.9
6 Notre Da 4.5 3.9
7 Columbia 4.5 4.6
8 Dartmout 4.5 4.3
9 Cornell 4.5 4.6
10 Pennsylv 4.5 4.5
11 Brown Un 4.5 4.4
12 Cal—Los 4.5 4.2
13 Rice Uni 4.4 4.0
14 Washingt 4.4 4.1
15 Virginia 4.4 4.3
16 Cal Inst 4.4 4.7
17 Cal—Berk 4.3 4.8
18 Johns Ho 4.3 4.6
19 Northwes 4.3 4.3
20 Duke Uni 4.3 4.4
21 Michigan 4.2 4.5
22 Georgeto 4.2 4.0
23 Chicago 4.1 4.6
24 Southern 4.0 4.0
25 North Ca 4.0 4.2
26 Florida 3.9 3.6
27 Tufts Un 3.9 3.6
28 Brandeis 3.9 3.6
29 Cal—San 3.9 3.8
30 Carnegie 3.9 4.2
31 Vanderbi 3.9 4.0
32 Boston C 3.9 3.6
33 Georgia 3.8 4.0
34 Cal—Irvi 3.8 3.6
35 Wisconsi 3.8 4.1
36 Pennsylv 3.8 3.8
37 Rocheste 3.8 3.4
38 Texas—Au 3.8 4.1
39 Illinois 3.8 4.0
40 Emory Un 3.7 4.0
41 New York 3.7 3.8
42 College 3.7 3.7
43 Washingt 3.7 3.9
44 Maryland 3.7 3.6
45 Renssela 3.7 3.5
46 Case Wes 3.7 3.5
47 Cal—Davi 3.7 3.8
48 Cal—Sant 3.6 3.5
49 Texas A& 3.6 3.6
50 George W 3.5 3.4
51 Rutgers- 3.5 3.4
52 Kentucky 3.5 3.0
53 Pittsbur 3.5 3.4
54 Ohio Sta 3.5 3.7
55 Wake For 3.4 3.5
56 Lehigh U 3.4 3.2
57 Virginia 3.4 3.4
58 Boston U 3.4 3.4
59 Connecti 3.4 3.2
60 Michigan 3.4 3.5
61 Delaware 3.4 3.1
62 Arizona 3.4 3.6
63 Clarkson 3.4 2.6
64 Worceste 3.4 2.8
65 Cal—Sant 3.3 3.2
</code></pre>

254

<p>I also calculated a formula for replicating the US News rank and score without using peer assessment. You don't really need peer assessment to arrive at a ranking that is almost the same.</p>

<p>I used:
freshman retention
classes over 50
financial rank
freshman retention
sat 75th percentile
actual graduation rate</p>

<p>The "R-square" was .97.</p>

<p>The formula:
predicted score/rank = 243.56107-(.05009<em>faculty resources rank)+(.39242</em>classes over 50)-(.2524<em>sat 75th percentile)
+(.00000005527644</em>sat 75th percentile cubed)-(.05519<em>financial rank)+(.00003687</em>actual graduation rate cubed) </p>

<p>school, rank, US News score, score predicted by formula (excluding peer assessment) </p>

<pre><code> 1 Harvard 2 99 104
2 Princeto 1 100 100
3 Yale Uni 3 98 98
4 Massachu 7 93 94
5 Stanford 4 95 91
6 Dartmout 11 89 91
7 Cal Inst 5 94 90
8 Columbia 9 90 90
9 Duke Uni 8 92 89
10 Pennsylv 5 94 89
11 Brown Un 14 86 88
12 Rice Uni 17 80 88
13 Washingt 12 87 86
14 Notre Da 19 79 85
15 Cornell 12 87 84
16 Northwes 14 86 83
17 Chicago 9 90 83
18 Johns Ho 14 86 83
19 Georgeto 23 74 80
20 Tufts Un 28 70 77
21 Carnegie 22 77 75
22 Vanderbi 19 79 75
23 Virginia 23 74 74
24 Cal—Berk 21 78 73
25 Emory Un 17 80 73
26 Cal—Los 25 73 73
27 Brandeis 31 66 72
28 Southern 27 72 70
29 Cal—San 38 62 69
30 Coll Wm Mary 33 65 68
31 Boston C 35 63 67
32 Michigan 25 73 66
33 New York 34 64 66
34 Wake For 30 69 66
35 Rocheste 35 63 64
36 Lehigh U 31 66 63
37 Case Wes 41 61 63
38 North Ca 28 70 63
39 Renssela 44 58 63
40 Georgia 35 63 61
41 Cal—Davi 42 59 59
42 Cal—Sant 44 58 59
43 Illinois 38 62 59
44 Yeshiva 52 54 58
45 Boston U 57 52 57
46 Wisconsi 38 62 57
47 Florida 49 56 57
48 Cal—Irvi 44 58 57
49 George W 54 53 56
50 Tulane U 50 55 56
51 Texas—Au 44 58 55
52 Worceste 62 50 55
53 Pennsylv 48 57 55
54 Pepperdi 54 53 54
55 Syracuse 50 55 54
56 Rutgers- 59 51 53
57 Maryland 54 53 53
58 Miami (F 52 54 52
59 Washingt 42 59 52
60 Pittsbur 59 51 51
61 Texas A& 62 50 50
62 Stevens 75 46 50
63 Southern 67 48 49
64 Connecti 64 49 49
65 Delaware 71 47 49
</code></pre>

255

<p>We're forgetting in this discussion that students are only part of what makes a great academic community. Professors and graduate student instructors are the other half. PA scores may relfect this component as well.</p>

256

<p>
[quote]
We're forgetting in this discussion that students are only part of what makes a great academic community. Professors and graduate student instructors are the other half. PA scores may relfect this component as well.

[/quote]
</p>

<p>Fwiw, I think that the most vocal critics of the PA do think that the PA does in fact ignore that second part to a great extent, especially the lower physical presence of the professors with great reputation in the UNDERGRADUATE classes and the rather large presence but unequal contribution of Teaching Assistants who are hardly qualified, prepared, and trained. </p>

<p>Too much is made of the reputation of professors who hardly ever teach undergarduates and too little is made of the ones who really do spend time educating undergraduates.</p>

257

<p>quote]This argument that poor people don't work hard enough or aren't intelligent is rather ignorant and sickening to me. What do you think Xiggi?

[/quote]
</p>

<p>Dstark, I hope you don't think I have ever made this argument. </p>

<p>
[quote]
Or do you think all SAT scores are the same no matter how much tutoring a person has?

[/quote]
</p>

<p>DS, that all depends on the form of the tutoring. I believe it's a mistake to think that tutoring is a based on a passive relation between the tutor and the tutee. I happen to think that the SAT scores are a function of natural abilities, the amount of time one devotes to maximize those abilities, and the adherence to winning methods. When the tutoring is based on removing self-imposed obtacles and spending practice to overcome weakness, I do believe that the score will move --upwards usually. </p>

<p>And, as you know, except for special cases, I do believe that students can accomplish much with a couple books borrowed from the local library and the desire and ambition to do well.</p>

258

<p>
[quote]
Too much is made of the reputation of professors who hardly ever teach undergarduates...

[/quote]
</p>

<p>That's a common refrain from those that support LACs. But, I'd be interested to hear, xiggi or hawkette, of any actual statistics that support that position. Do we really KNOW that the nobelists/research super stars at Chicago, Michigan, Texas or Cal do not teach? Do not hold office hours? Do not show up in labs?</p>

259

<p>BB, it is indeed a common refrain, but hardly confined to people who support a LAC education. Evidence of the validity of this claim is only as distant as one's keyboard and a modicum of googlin' experience. This said, would any amount of evidence change the opinion of people who support and like schools that use TA and TF to their fullest potential and accomodate professors' love for research and give them all liberties in exchange of landing the mighty dollar?</p>

<p>Here's the first hit of a FIVE second search:</p>

<p><a href="http://query.nytimes.com/gst/fullpage.html?res=9F0CE1D81031F93AA15752C1A965958260&sec=&spon=&pagewanted=print%5B/url%5D"&gt;http://query.nytimes.com/gst/fullpage.html?res=9F0CE1D81031F93AA15752C1A965958260&sec=&spon=&pagewanted=print&lt;/a&gt;&lt;/p>

<p>Universities Seek Ways to Get Their Stars to Teach </p>

<p>The article being written in 1993 shows that it's not an issue that started with the dwindling budget cuts on the past ten years.</p>

260

<p>The only thing that collegehelps' regressions prove, while little, is that Chicago is incredibly overrated and doesn't deserve to be a top 10 school over any other top 20 school. Sure, it might have great research and great reputation? So do 20+ schools?</p>