what are the new top25 universities as a senior for class 2009?

College Search & Selection
161

<p>Hello everyone:</p>

<p>Thank you for your participation in this thread’s mini modeling exercise. I’m not just come here to defend my modeling results. Instead, I would like to briefly share with you my observations, from a modeler viewpoint. </p>

<p>First of all, there isn’t a list which can fit everybody needs. Not even USNews or NRC can achieve the goal even though they keep updating their lists on regular basis. I didn’t intend to make such a list. </p>

<p>As posted in #131, data availability is always the key issue for the model development. A reliable model needs appropriate data support. My biggest challenge was to get more seniors involved in this mini-study or research. With only a few survey participants (SP, which is less than 10), rigorously speaking the modeling results represented only their opinions. The model will be deemed acceptable/credible as long as the majority of the SP agrees upon the modeling results. Eventually, I will let my SP to defend the model for me.</p>

<p>During the questions and comments periods, my interests were focused on SP’s comments on technical issues e.g., scoring of the survey data, evaluation of undergraduate/graduate strengths data and handling of missing data instead of taking too much weight-in on professional peer-assessments. Upon re-examining missing data as per suggestion from SP members, model has an up-to-date ranking (same as post #158), which is the final draft, based on SP of this mini-modeling exercise. Brief discussions of how to collect defensible data and how to obtain/develop a credible model of college ranking from a modeler’s viewpoint can be found on post #112 and post #75, my reply to tk21769. </p>

<p>SP members, please express your final opinions/thoughts on this exercise.</p>

<p>Quote from tk21769:
“This thread took shape as a modeling exercise guided by the OPs interests and skills. My own approach was to take independent rankings of one or more measurable features, then merge them to see if the outcome would yield reasonable approximations to participants’ intuitions. That process does not necessarily reflect how I’d choose a school to attend. But one thing I’d like to explore is how to break away from over-reliance on holistic, professional peer-assessments that may be in thrall to a “halo effect”.”</p>

<p>Tk, well put! That was an excellent modeling statement. You could become a good modeler. </p>

<h2>Strength Institution Ranking</h2>

<p>2 Harvard 1
2 Princeton 1
3 Stanford 1
3 Yale 1</p>

<h2>3 MIT 1</h2>

<p>7 Caltech 2
8 Columbia 2
9 Chicago 2
10 Penn 2</p>

<h2>10 Duke 2</h2>

<p>13 Cornell 3
14 Berkeley 3
15 Brown 3
15 Dartmouth 3
16 Northwestern 3
17 Johns Hopkins 3
18 Michigan 3
18 Rice 3</p>

<h2>18 WUSTL 3</h2>

<p>21 UCLA 4
21 UVA 4
22 Emory 4
23 Carnegie Mellon 4
23 Vanderbilt 4
24 Georgetown 4</p>

<h2>24 UNC 4</h2>

<p>27 Notre Dame 5
27 NYU 5
27 Wisconsin 5
28 UIUC 5
29 USC 5
30 Georgia Tech 5</p>

162

<p>Quote from ModestMelody:</p>

<p>"Modeling, you may want to include these numbers:
Rankings of Undergraduate Universities by Prestige</p>

<p>People are listing schools in tiers and there is a broader set of people responding. More data = better model. "</p>

<p>Modest,</p>

<p>Good suggestions. Scoring won’t be a problem as long as their listings were in tiers. It would be interesting to see the model reactions to the new group. Stay tune.</p>

<p>Before doing that, my only question is “Do I need to get the permission to access or use the data from other threads?” If yes, who is the contact person?</p>

163

<p>Luckily, I don’t believe that any information on here is something you need permission to reuse unless you’re reproducing it off-site. I do know that the moderators don’t delete threads/accounts even at direct requests most of the time because they view maintaining this forum as an archive of these discussions as an important aspect of the CC community. If you’re concerned at all, I’d send a Private Message to tokenadult or one of the other moderators, but really I think that there should be no issue here.</p>

164

<p>Modest, data ownership oftentimes goes side-by-side with intellectual properties. As a modeler, I have been through this all the time especially for FED and state agencies. It would be better to get the green light before I start it. Would you please email tokenadult for me in this regard?</p>

165

<p>wow, I’m not even going to read through this thread but can somebody give me a quick summary of what you are attempting to do?</p>

166

<p>pierre0913:</p>

<p>If you understand what “liability” issue means, then you will know it. No need to go through the entire thread, just read post#112. Hope this helps.</p>

167

<p>Yes, indeed. This tread has been dragging too long. I will try to conclude this mini-study soon (According to my original plan,we only completed two thirds of it but I will condense it.). I hope I didn’t waste your time.</p>

168

<p>An additional 22 tiered ranking data were incorporated into the model: four from this thread and 18 from another thread suggested by ModestMelody. Based on all 31 survey data, model has predicted the final top 25 universities rankings (Please see attached table for details). </p>

<h2>Order Ranking Name Score Tiered Rankings</h2>

<p>1 1 Princeton 100.0 1
2 1 Yale 100.0 1
3 1 Harvard 100.0 1
4 4 Stanford 99.4 1</p>

<h2>5 5 MIT 96.8 1</h2>

<p>6 6 Columbia 80.6 2
7 7 Caltech 76.8 2
8 8 Cornell 75.5 2
9 8 Duke 75.5 2
10 10 Penn 74.2 2
11 11 Chicago 71.6 2
12 12 Berkeley 67.1 2
13 12 Brown 67.1 2
14 12 Dartmouth 67.1 2
15 12 Johns Hopkins 67.1 2</p>

<h2>16 16 Northwestern 63.2 2</h2>

<p>17 17 Michigan 54.8 3
18 18 UVA 49.7 3
19 19 WUSTL 49.0 3
20 20 Rice 47.7 3</p>

<h2>21 21 Georgetown 47.1 3</h2>

<p>22 22 Vanderbilt 39.4 4
23 23 UCLA 37.4 4
24 24 Carnegie Mellon 31.6 4
25 25 Emory 31.0 4
26 26 UNC 27.7 4</p>

<h2>27 27 Notre Dame 24.5 4</h2>

<p>28 28 NYU 17.4 5
29 29 USC 13.5 5
30 30 Wisconsin 11.0 5
31 31 Georgia Tech 7.7 5
32 31 UIUC 7.7 5
33 31 Tufts 7.7 5
34 34 Texas 5.8 5
35 34 W&M 5.8 5
36 36 Brandies 3.2 5
37 36 Tulane 3.2 5
38 36 UCSD 3.2 5
39 39 Washington 2.6 5
40 40 BC 1.3 5</p>

169

<p>

</p>

<p>tk21760 -</p>

<p>If I’m not mistaken, the NRC data you cite is from 1995 - and based on a ranking of the 1995 NRC program rankings by average of nonzero scores - Duke places 20th, behind schools like Northwestern, Illinois, Penn, UDub, UT, UCLA, UM, etc.</p>

<p>In terms of a ranking based on all scores - Duke is not in the top 15.</p>

<p>[United</a> States National Research Council rankings - Wikipedia, the free encyclopedia](<a href=“http://en.wikipedia.org/wiki/United_States_National_Research_Council_Rankings]United”>United States National Research Council rankings - Wikipedia)</p>

<p>Reportedly, there should be an updated NRC rankings later this year and we shall see how things shake out.</p>

170

<ol>
<li><p>Using tiered format for your survey: A 5-tiered format was used for realistic reasons because it should be easier to tell if a given university belongs to a certain tier rather than guess its exact rankings. For top25 universities, between 25 and 40 best universities should be included in case of tie situations. </p></li>
<li><p>Tiers are relative terms and should have their well defined sample size: In our case, we identified several (5) tiers to separate top 25 national universities, all of these top 25 would certainly be on the top tier list of the original 250ish universities defined by Carnegie Institute. Depending on sample size, if 5-tiered format are using each tier may have 5 institutions for top 25 and 50 institutions for top 250. </p></li>
<li><p>If I were to analyze the ranking among top5, I would take similar approach and make sure everyone knows that we focus either on undergraduate or graduate education but not both in order to avoid unnecessary confusions. HYPSM may be very good in undergraduate education but it will be a different story when it comes to graduate schools or professional schools (law, medicine, business, education, etc). Many sources identify universities e.g., Berkeley (graduate school), Michigan/UVA (law), JHU (medicine), Penn (business), Columbia (education) have distinguished roles in their respective fields.</p></li>
<li><p>Use appropriate number of tiers to separate universities: Based on survey data, more than one scoring methods were applied to verify the “’gaps” between the universities. In our case, I used three scoring methods to ensure that the cutoff line for each tier was objective. From three- scores comparisons, I found it is very difficult to separate Brown, Columbia, Cornell, Dartmouth, and Penn (5 ivies), Berkeley, Chicago, Caltech, Duke, and JHU into different tiers so the Tier2 is larger than the others, based on survey data.</p></li>
<li><p>USNews08 and NRC95A are two excellent sources for undergraduate- and graduate- strength ratings. The former is a good reference for undergraduate institutions. The latter, a comprehensive summary/analysis of NRC95 database by Dr. H. J. Newman of TexasA&M, is an excellent reference for graduate institutions. Upon deriving the ranking from the survey data, I compared the calculated rankings (Ranking from post#169) with other sources e.g. USNews08 and NRC95 to identify surprises. As a result, the derived ranking had a good balance from both undergraduate (USNews) and graduate (NRC95A) strengths of the institutions, which were reflected and/or highlighted by high positive or negative numbers by their respective influences from DUSNews (Berkeley, Michigan, and Notre Dame) and DNRC95 (Berkeley, Brown, and Dartmouth). </p></li>
<li><p>S-Ranking (2nd column from post #169’s table) is a numerical ranking of top25 universities generated by statistical averaging process (I am not a big fan of this though.). Statistics discerns the differences between polls which may be a minor thing we didn’t pay attention to, but believe it or not people usually take the results very seriously. Usually the larger or more diversified the sample size (participants of polls) is and the bigger differences the results would be.</p></li>
</ol>

171

<p>I am glad NYU is increasing in ranking. If it makes top 25 I would be ecstatic. </p>

<p>Also, I agree with ModestMelody, Brown University is severely underrated. I dislike how some people group Brown with Cornell. Here is my ranking of the ivies:</p>

<p>1) HYP</p>

<p>(small gap)</p>

<p>2) UPenn/Columbia</p>

<p>( small gap)</p>

<p>3) Dartmouth/ Brown</p>

<p>( noticeable gap)</p>

<p>4) Cornell</p>

<p>Brown had an 8% acceptance rate this year RD! Geez!</p>

172

<p>IMHO, Cornell and Brown are in the same league as Columbia and Dartmouth. UPenn is overrated compared to Cornell and Brown for undergrad :slight_smile: I have high respect for Penn med, law, and business professional graduate schools though. UChicago is overrated for undergraduate. Berkeley and Georgetown are both highly underrated in CC forums :-(</p>

173

<p>Phead, if you are referring to the ranking in #169, I disagree that Chicago is over-rated for its undergraduate programs. Or at least, its tier placement is not too high. I’m biased because I’m a Chicago alum, but I think I can back up my opinions with data.</p>

<ol>
<li>USNWR’s #8 ranking is consistent (tier-wise) with several other studies of overall undergraduate experience or post-graduate outcomes. Examples:
It performs better than any other national liberal arts university in the HEDS studies of baccalaureate origins of Ph.D.s. In fact it is the only school in the overall top 10 that is not a LAC or engineering/technical institute.
The college ranks 14th in the Wall Street Journal’s ranking of top feeder colleges to fifteen elite business, law, and medical schools.
In 2008, Forbes ranked the college 4th best in the country after Harvard, Yale, and Princeton.
In 2007, Princeton Review ranked the University of Chicago #1 in the country for “best overall academic experience for undergraduates”.<br>
It ranks first among colleges with fewer than 5,000 students for number of Peace Corps volunteers.</li>
</ol>

<p>For a school its size, the college has educated a remarkable number of prominent people in government, higher education, journalism and the arts ([List</a> of University of Chicago people - Wikipedia, the free encyclopedia](<a href=“http://en.wikipedia.org/wiki/List_of_University_of_Chicago_people]List”>List of University of Chicago people - Wikipedia))</p>

<ol>
<li><p>For average undergraduate class size, among national universities by % of classes under 20, it ranks 9th. By % of classes not larger than 50, it is tied for 3rd. On another thread, Hawkette developed a composite class size metric that showed Chicago tied for 1st.</p></li>
<li><p>It not only is famous for the strength of its faculty (more affiliated Nobels than any other US university), but many of the star professors actually teach undergraduates. Chicago has what may be the oldest award for undergraduate teaching excellence of any university, the Quantrell Award. This student-nominated award has been conferred on many nationally prominent faculty members including Pulitzer and Nobel recipients.</p></li>
<li><p>Student quality is high, too. By 75th% SAT scores, among national universities it is tied for 10th in a very close pack.</p></li>
<li><p>Among all library systems in the United States, Chicago’s is the 19th largest. Only 11 universities own more volumes, and they are all much larger schools than Chicago. The quality of its research facilities overall is superb. </p></li>
</ol>

<p>It’s a mistake to think of the school’s college as being overshadowed by outstanding graduate and professional schools. If anything, in the past 10 years the role of the College has been strengthened relative to other divisions. But the university has a long history of quality and innovation in undergraduate instruction.</p>

<p>Phead says Georgetown (1 Tier lower than Chicago) is highly under rated. I’ve taken many classes there, too, and it’s a good school. However, in my opinion it is not in the same league as Chicago for overall faculty quality, instructional approach, facilities, or class size.</p>

174

<p>

</p>

<p>K&S, yes, that data is from 1995. It’s the most up-to-date we have (as far as I know) unless you go to something like the citation-density study I’ve cited elsewhere.</p>

<p>And yes, Duke places about 20th, if you use the arithmetic mean of non-zero scores. That’s not how I averaged. I assigned the lowest score (L) plus 1 to the departmental ranking of a school that had no score in that field. This penalized schools that lacked breadth, compared to schools that achieved some score in many fields. I thought this was appropriate for departmental quality relative to undergraduate needs. And, I took the harmonic mean of all individual departmental scores to get my composite for each university.</p>

<p>Here are the ranks I get using the arithmetic means of NRC-95 departmental ranks (but still assigning L+1 to each non-scoring department):
Rank School
1 Stanford
2 Harvard
2 Yale
4 Chicago
5 Berkeley
6 Michigan
6 Wisconsin
8 Columbia
8 Cornell
8 Penn
8 UCLA
12 Duke
12 Washington
14 Hopkins
15 CalTech
15 Illinois
15 UNC
18 Princeton
19 MIT
20 Virginia
21 Northwestern
22 Brown
23 NYU
24 Cal SB
24 WUSTL
26 Vanderbilt
27 Emory
27 Rice
29 Carnegie Mellon
30 Georgetown
31 Notre Dame</p>

<p>As stockbrokers like to say, these results are deemed reliable but not guaranteed.</p>

175

<p>Modeling, I like your style.</p>

176

<p>Princeton is the #18 school? Carnegie Mellon below Vanderbilt? Accurate using the data given, perhaps; but quite a useless ranking.</p>

177

<p>Baelor, you need to read through the thread a little. Nobody’s seriously claiming Princeton is the #18 undergraduate school.</p>

178

<p>

</p>

<p>Looks perfect except that I would replace Cornell with Berkeley. :)</p>

179

<p>^^I did – this is why I included the caveat in my statement.</p>

180

<p>Baelor, a ranking like that highlights schools that have quality research programs across a broad range of arts and science fields, with relatively few weak areas among the surveyed departments. Princeton lacks breadth in some areas of biological science research compared to larger universities that have medical schools; that’s one reason it comes out lower than expected in that particular list. Compare stateuniversity.com, where UPenn comes out as #66. You do get these anomalies in some rankings.</p>