Part I: Ignore the Rankings & Look at the Data (Retention & Graduation Rates)

College Search & Selection
1

<p>Everyone has their own personal misgivings about the USNWR rankings, whether it be the weights that are assigned to various data points, the accuracy of various data points, and even the relevance of some data points. So, let’s focus on the data itself and compare the colleges, category by category, and see how they compare. The reader can decide for him/herself the usefulness of the measure for his/her own purposes. </p>

<p>Following is the data for the USNWR Top 50 National Universities and how they compare on Graduation Rates (6-year and I am adding information on 4-Year rates even though this is not part of the USNWR ranking equation) and Freshmen Retention. If anyone has any additional, related data to provide surrounding this category of comparison, then please provide it. </p>

<p>Ranked by 6-Yr Grad Rate , 6-Year Grad Rate , 4-Year Grad Rate , Freshman Retention Rate , School</p>

<p>1 , 98% , 87% , 98% , Harvard
2 , 96% , 90% , 98% , Notre Dame
2 , 96% , 89% , 98% , Princeton
2 , 96% , 87% , 98% , Yale
5 , 95% , 80% , 98% , Stanford
6 , 94% , 90% , 97% , Duke
6 , 94% , 90% , 97% , Georgetown
6 , 94% , 87% , 98% , U Penn
6 , 94% , 87% , 98% , Dartmouth
6 , 94% , 86% , 98% , Columbia
6 , 94% , 83% , 97% , Brown
12 , 93% , 86% , 97% , Northwestern
12 , 93% , 83% , 96% , J Hopkins
12 , 93% , 82% , 98% , MIT
12 , 93% , 79% , 97% , Rice
16 , 92% , 87% , 96% , Tufts
16 , 92% , 84% , 97% , U Virginia
16 , 92% , 84% , 96% , Cornell
16 , 92% , 84% , 95% , W & M
20 , 91% , 88% , 95% , Boston College
20 , 91% , 83% , 97% , Wash U StL
22 , 90% , 83% , 97% , U Chicago
23 , 89% , 85% , 95% , Vanderbilt
23 , 89% , 80% , 97% , Cal Tech
23 , 89% , 62% , 97% , UCLA
23 , 89% , 61% , 97% , UC Berkeley
27 , 88% , 84% , 95% , Brandeis
27 , 88% , 79% , 94% , Wake Forest
29 , 87% , 82% , 94% , Emory
29 , 87% , 70% , 96% , U Michigan
31 , 86% , 75% , 94% , Lehigh
31 , 86% , 70% , 94% , Carnegie Mellon
31 , 86% , 54% , 94% , UCSD
34 , 85% , na , 93% , Penn State
35 , 84% , na , 95% , USC
35 , 84% , na , 91% , UC S Barbara
35 , 84% , 74% , 92% , NYU
35 , 84% , 71% , 96% , U North Carolina
39 , 82% , na , 93% , Rensselaer
39 , 82% , na , 92% , U Illinois UC
39 , 82% , na , 92% , Syracuse
42 , 81% , 43% , 91% , UC Davis
43 , 80% , na , 94% , U Rochester
44 , 79% , na , 94% , UC Irvine
44 , 79% , 53% , 94% , U Florida
44 , 79% , 46% , 93% , U Wisconsin
47 , 78% , na , 92% , Case Western
48 , 77% , 46% , 93% , U Texas
48 , 77% , 34% , 92% , Georgia Tech
50 , 75% , 48% , 93% , U Washington
51 , 71% , na , 87% , Tulane</p>

2

<p>graduation rates are related to SAT scores except that schools with a large percent engineering have a more difficult overall curriculum and lower-than-expected graduation rates (e.g. Caltech).</p>

3

<p>Top 4-Year Graduation Rates </p>

<p>School Percent
Haverford College (PA) 92%
Williams College (MA) 91%
Duke University (NC) 90%
Georgetown University (DC) 90%
University of Notre Dame (IN) 90%
College of the Holy Cross (MA) 89%
Davidson College (NC) 89%
Princeton University (NJ) 89%
Amherst College (MA) 88%
Boston College 88% </p>

<p>USNews.com:</a> America's Best Colleges 2008: Highest graduation rates</p>

<p>LAC Leaders in Freshman Retention Rate </p>

<p>Freshman Freshman
Institution retention rate enrollment
College of the Holy Cross 96% 754
Davidson College (NC) 96% 464
Haverford College (PA) 96% 314
Swarthmore College (PA) 96% 370
United States Naval Acad 96% 1,228
Wesleyan University (CT) 96% 720
Amherst College (MA) 97% 433
Carleton College (MN) 97% 504
Williams College (MA) 97% 534
Bowdoin College (ME) 98% 474
Pomona College (CA) 99% 378</p>

<p>USNews.com:</a> America's Best Colleges 2008: Liberal Arts Colleges: Schools whose freshmen are least (and most) likely to return</p>

4

<p>The four year graduation rates at Berkeley, UCLA and UCSD caught my eye.
Outside of California, we read that students have trouble at these schools getting access to the courses they need to graduate. Is this the case, or are other things going on?</p>

5

<p>I suspect it also has to do with the number of commuter students at these universitites who are not taking "full loads".</p>

6

<p>
[quote]
graduation rates are related to SAT scores except that schools with a large percent engineering have a more difficult overall curriculum and lower-than-expected graduation rates (e.g. Caltech).

[/quote]
</p>

<p>CH, isn't it interesting that graduation rates could be related to SAT scores? Haven't we be bombarded by quasi-scientific reports that the SAT was not a good predictor for college success? </p>

<p>On a related note, we should remember that USNews does in fact allocate a substantial percentage of its total points to a category labeled Expected Graduation Rates. In fact, this category places such a handicap on highly selective schools (highest ranked schools in selectivity) that it all but negate the "bonus" points for high SAT scores. In fact, schools with lackluster admission standards, easier curriculum, and easier grading policies do benefit in this USNews category. This is perfectly illustrated by comparing schools such Harvey Mudd and the various all-females LACs, or Caltech with numerous schools that have more forgiving grading policies.</p>

7

<p>But then again, about the graduation rates and SAT scores: correlation does not imply causation.</p>

8

<p>
[quote]
But then again, about the graduation rates and SAT scores: correlation does not imply causation.

[/quote]
</p>

<p>Here's goes that unavoidable clich</p>

9

<p>Some programs require more than 4 years (architecture, for example) so it's not really fair to look at the 4-year graduation rate without some qualifiers.</p>

10

<p>xiggi-
Yes, there are some who are convinced the SATs do not predict graduation and retention but the relationship is actually quite strong. I think the misconception comes from trying to predict graduation for individual students within a particular university with SAT scores. This method is problematic because (1) smarter students go into harder majors (2) sometimes smarter students transfer to better schools - upward mobility (3) the decisions to stay or leave are affected by finances, personal issues, and other idiosyncratic factors, (4) graduation for an individual is a dichotomous variable which makes it difficult to determine the magnitude of the relationship with SATs (via logistic regression).</p>

<p>But, when you look at aggregated data, correlating mean SAT or 25th percentile SAT, with graduation rate, the relationship between SATs and graduation is higher.</p>

<p>The relationship between SATs and gpa is even higher than between SATs and graduation.</p>

11

<p>Hawkette:</p>

<p>It might be easier if you just list schools by full-pay percentage. The number of wealthy kids at a school also correlates with graduation rate (and SAT scores). Any thoughts on the lower 4-year numbers for Rice and Wake?</p>

<p>Danas: The UCs, like many publics, are extremely generous with AP/IB credit so graduating in four years is not difficult for kids who are full pay, should they so desire. But, those three campuses have 33% Pell grantees, and UC finaid is not great, so those kids must work part-time; taking a minimum load, therefore requires more than four years.</p>

12

<p>After a student has chosen some appropriate schools, retention and graduation rates might help with a final decision, but I think there are much better starting points for the search.</p>

13

<p>bluebayou,
I think you may have a point as the table below shows the percentage of students at these colleges that did NOT apply for financial aid. While the public universities clearly underperform in 4-yr and 6-yr graduation rates (with the impressive exception of U Virginia), they tended to have the highest percentages of students in search of financial aid (again, U Virginia is the big outlier). </p>

<p>Here is the full breakdown for the USNWR Top 50 National Universities and the percentage of their students that did NOT apply for financial aid.</p>

<p>Rank, % of students who did NOT apply for financial aid , College
1 , 63% , Columbia
2 , 61% , Wash U
3 , 58% , Yale
3 , 58% , U Virginia
5 , 55% , Harvard
6 , 54% , Georgetown
7 , 53% , U Penn
7 , 53% , Duke
7 , 53% , Brown
7 , 53% , W&M
11 , 52% , U Wisconsin (06-07)
11 , 52% , U Florida
13 , 51% , Tufts
13 , 51% , Wake Forest
13 , 51% , Boston Coll
16 , 50% , Emory (06-07)
17 , 49% , Rice
18 , 48% , Princeton
18 , 48% , Johns Hopkins
20 , 47% , USC
20 , 47% , U Texas
22 , 46% , Stanford
22 , 46% , Vanderbilt
22 , 46% , Carnegie Mellon
25 , 45% , U North Carolina
26 , 44% , UCLA
27 , 42% , Lehigh
28 , 41% , UC Berkeley
29 , 40% , Dartmouth
29 , 40% , U Michigan (all 3 U M campuses)
29 , 40% , U Illinois
32 , 39% , Northwestern (06-07)
32 , 39% , Brandeis
32 , 39% , Georgia Tech
35 , 37% , U Chicago
35 , 37% , Cornell
37 , 35% , Notre Dame
37 , 35% , UC Santa Barbara
39 , 34% , UCSD
40 , 31% , Caltech
40 , 31% , NYU (06-07)
40 , 31% , UC Davis
40 , 31% , UC Irvine (03-04)
44 , 29% , MIT
44 , 29% , Syracuse
46 , 26% , U Rochester
47 , 23% , Case Western
48 , 15% , Rensselaer
na , U Washington
na , Penn State
na , Tulane</p>

14

<p>I would strongly downplay the role of 4-year graduation rates. Even at Harvard they are quite low, due to the propensity for students to take time off for various reasons -- health, family, travel, etc.</p>

<p>As for the 6-year graduation rates they may have some utility, but they again fail to consider different academic programs (e.g. 5 year B.Arch, Nursing, or Engineering programs) </p>

<p>Another important point to consider is that a lot of schools suffer from students transferring out to other schools, even though the school is already offering a fantastic education and environment. A great example of this are the SUNYs and the number of students that are able to transfer to Cornell every year. Is SUNY-Geneseo any less off a school if many students are able to transfer to Cornell and graduate within six years?</p>

<p>Finally, I will agree with others that you must control for the quality of "inputs" into this "output" -- namely student quality. Harvard will of course have the highest 6-year graduation rate, as it is widely acknowledged to have one of the most driven, talented, and wealthy student bodies in the country.</p>

<p>I will defer everybody to the Washington Month's analysis of how well colleges graduate students after taking wealth and talent into consideration:</p>

<p><a href="http://www.washingtonmonthly.com/features/2007/0709.natlrankings.pdf%5B/url%5D"&gt;http://www.washingtonmonthly.com/features/2007/0709.natlrankings.pdf&lt;/a&gt;&lt;/p>

<p>It's pretty clear that some schools perform incredibly well in this measure: UCLA, UVA, Texas A&M, Penn State.</p>

<p>Other schools, for all their focus on resources and undergraduate education, do not: Duke, Dartmouth, WashU, etc.</p>

15

<p>Here is Washington Month's analysis of how well "LAC" colleges graduate students after taking wealth and talent into consideration:</p>

<p><a href="http://www.washingtonmonthly.com/features/2007/0709.lacrankings.pdf%5B/url%5D"&gt;http://www.washingtonmonthly.com/features/2007/0709.lacrankings.pdf&lt;/a&gt;&lt;/p>

16

<p>This is probably a stupid question: </p>

<p>Are these graduation rates determined by how many semesters students spend at a school before graduating, or whether they took time off? I mean; if a student takes two years off in between their studies, but graduates after semesters total spent at the school, they will count as having graduated within 6 years, not 4?</p>

17

<p>I don't buy the Washington Monthly rationale.
Why wouldn't a Pell Grant recipient make a special effort to graduate in 4 years and get the heck out there to make a living?
I say this as a parent of a Pell Grant recipient who made darn sure he got in and out in 4 years. Financial aid, at least at the privates, seems to put pressure to do the job in 4 years.
I agree that at schools outside the Washington Monthly list (community colleges, lower tier publics and privates), students face the problem of supporting themselves and their families, and getting through in 4 years is not realistic. But schools in the the Washington Monthly list? Give me a break. More pressure for a kid from a family of modest means to graduate in 4 years, rather than less, IMO.</p>

18

<p>The Washington Monthly is nothing more than an asinine effort to be in the ranking business. The underlying data lacks foundation and seems to be riddled by mistakes. </p>

<p>It's pure garbage.</p>

19

<p>xiggi, I don't see why you believe that the correlation does imply causation in this case. Could you explain why you believe that I misunderstand this?</p>

20

<p>Correlation DOES imply causation. It does not PROVE causation but it does suggest causation. In Social Science, relationships are rarely "proven" because true experiments are almost impossible. "Proof" often consists of building credibility for a hypothesis through repeated correlational studies.</p>

<p>In a true experiment, correlation WOULD prove causation.</p>

<p>True experiments are possible in the physical sciences where correlation does prove causation.</p>

<p>Causality comes from the design of research, not the statistic you use.</p>

<p>Correlation</a> does not imply causation - Wikipedia, the free encyclopedia</p>