<p>offensive language? Where?</p>
<p>fallenchemist:</p>
<p>What I read from your post #37 and 38 were arguments and criticisms instead of learning and understanding. From the questions and comments you made on post # 38, I believe you didnt get the whole idea of modeling technique and its application to college rankings. It usually takes several years for an engineering PhD to be a mature modeler. Thought I can help on the basic concept for model developing. Alas, let us move on.</p>
<p>alam1:</p>
<p>Prior to 1996, the major ranking categories have remained constant: student selectivity, academic reputation, faculty quality, financial resources, and alumni giving. In 1996, USNews added student retention and graduation rates as a new category to measure a university’s “value added,” which was evaluated by the difference between actual and expected graduation rates, based on the assumption if you let in highly qualified students, your expected graduation rate is high. </p>
<p>Other than the aforementioned change, I can’t think of any other big changes in that year. In 1998, there were many ties for the top 25: five 2-way ties at the top, numbers 3, 7, 17, and 21, one 3-way tie at number 14, and one 5-way tie at number 9, which suggested the needs for adding more categories to help obtain better separation among universities. Two years later, in 2000, “student selectivity” category was set to be further divided into SAT scores of the 25th and 75th percentile of the entering freshman class, their class ranks, the percentage of applicants accepted, and “yield,” the percentage of admitted applicants who enroll.</p>
<p>You are not modeling, you are ranking. Models emulate/predict based on a reality. You model climate change. You model molecular interactions. You don’t create models of what is a best college, you simply create ranking systems. You are the one that doesn’t get the idea, and I doubt you have years of any experience whatsoever, actually. I don’t need to learn and understand, you do.</p>
<p>Good job compiling this data.
I’m not sure how useful college rankings are, it seems like there could be a million variables in the methodology. Do you seriously think HYSPM is any different than PMYHS? (I apologize for the silly acronyms).</p>
<p>No apology needed, your point is correct, and I would take it a bit further. Rankings are only useful if they are pertinent to what you are looking for in a school, and that thing is measurable. So if you are looking for the schools that have the students with the best average SAT scores, that is easy, assuming all schools measure and report this statistic the same way. Which actually shows how hard this can be, because then you have to eliminate SAT optional schools, since students with lower SAT scores will opt not to send them, artificially raising the average at that school. If you are looking for the school with the best athletic spirit, that is nearly impossible to measure. Trying to say what are the “best” schools is obviously unmeasurable, unless you define best.</p>
<p>USNWR tries to define best as some mix of what others think of that school and “hard” statistics, although it is clear that the statistics are not only subject to inconsistent measuring, schools have started to manipulate their processes to get the numbers looking better without outright lying, although that happens too. But even putting all that aside, “best” is a very elusive concept when the goal is to have something that is broadly useful. Since only a very small percentage of students in this country can get into top 50 schools, let alone HYPSM type schools, I think the rankings have created a mentality that is detrimental to the entire process of choosing schools. Add to that the fact there are students that would be accepted to HYPSM schools for whom those schools are clearly not the best choice for them for any number of reasons. Hence my position that this whole thing is a fool’s exercise.</p>
<p>Fallenchemist:</p>
<p>Here you come again repeating your personal opinion:”One cannot model something as subjective as a “best” college”. Whereas, I believe that a well-calibrated model, under appropriate peer-review guidance, should be able to predict and/or help suggest the best colleges within a pre-defined set of selection criteria.</p>
<p>On post #44, the correct statement should be: You don’t simply create ranking systems, you also create models that defend the modeling results (the best colleges), based on the ranking systems created. Your personal statements on me were untrue. I learned and I got the idea. If I didn’t get the idea, how could I receive my PhD within 3.5 years from a top 25 (sometimes top 10) university? If I didn’t learn, I wouldn’t have published more than two dozens of technical papers published under my name on various modeling topics. </p>
<p>Again, you’re entitled to your own opinion but post no more, please. Let us move on.</p>
<p>Quite frankly, Liao, you don’t own this space and decide who posts and how often. If one can do that , then I will tell you to post no more and move on. How does that work for you?</p>
<p>As far as your qualifications, yeah sure. It was good for a laugh, anyway.</p>
<p>Irishevan99:</p>
<p>As summarized on post #6, based on findings presented in post #1, there were five distinguished tiers (levels) among US News favorite national universities (top 25):</p>
<p>Tier 1 (1-5) HYPSM (5 in total)</p>
<p>Tier 2 (6-10) Caltech, Chicago, Columbia, Duke, and Dartmouth (5 in total) </p>
<p>Tier 3 (11-13) Penn, Brown, and Berkeley (3 in total) </p>
<p>Tier 4 (14-20) Cornell, JHU, Northwestern, Rice, UVA, Emory, WUSL (7 in total) </p>
<p>Tier 5 (21-26) Notre Dame, CMU, Michigan, Georgetown, UNC, Vanderbilt (6 in total)</p>
<ol>
<li><p>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. A total of 25 surveys (years) did not suggest greater 5 tiers among top25 universities. More data usually provide greater separation (gap) if criteria are properly chosen. </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>Universities among the same tier should share the same rank since in any given year one university might outrank the others (HYPSM=PHYSM=SYHPM…). </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 two scoring methods to ensure that the cutoff line for each tier was objective. From two- scores comparisons, I found it was very difficult to separate Cornell, JHU, Northwestern, Rice, UVA, Emory, and WUSL into different tiers so the Tier4 was larger than the others, based on USNews annual survey (Top 25) data.</p></li>
</ol>
<p>UVA is rated too high. It is relatively weak in the natural sciences and engineering. Two very big areas of academic prestige.</p>
<p>OK Liao, show me how your model works with the following example. A student with 1920 SAT scores, 3.5 UW/3.8 W GPA, looking for a school with between 4K-8K undergrad population, thinks they want to study Russian culture, likes sports as part of the experience, can afford about $30K per year. Tell me the best school for this student using your “model” and show me exactly how the “model” arrived at its conclusion. Your “model” says HYPSM are “best”. Oh wait, this student can’t get into those schools. Must not be best for him/her. Hmmm, something must be wrong with the “model”. Oh wait! That’s because it isn’t a model. It is a ranking system based on factors that actually have nothing to do with the needs of any individual student other than those extreme few who happen to value the same parameters you chose to use to rank the schools. Never mind.</p>
<p>
That’s because the whole premise is garbage! You prove the point beautifully, assuming your statement is true. Natural sciences and engineering are important to you, therefore there are numerous schools that are “better” than UVa, most not even listed on this ranking of “best” colleges. Case closed.</p>
<p>rjk:</p>
<p>As mentioned on post #20, in 1989, USNWR introduced new empirical formula, which weighed in other important academic criteria e.g., faulty resources, acceptance rate, class size, etc.; and such adjustment had profound impact on top 25 (undergraduate rankings) and top 25 had never been the same ever since. UVAs strong performance in the 1980s and 1990s help secure its ranking at tier 4 (or number 14-20).</p>
<p>In the 1980s: UVA (15-20), Average 18</p>
<p>In the 1990s: UVA (17-22), Average 20</p>
<p>In the 2000s: UVA (20-24), Average 23</p>
<p>From Table 1, we may observe the rising of Non-HYP ivies and falling of big name public universities from 1983 to 2009. In the 1980s, Berkeley (5-24), Michigan (7-25), and UNC (9-23) matched well with Dartmouth (6-10), Brown (7-13), and Cornell (8-14); UVA (15-20), UIUC (8-20), and Wisconsin (13-23) were as competitive as Columbia (8-18) and Penn (15-19) In the 1990s, only UVA (17-22) and Berkeley (13-27) could complete with Columbia (8-15), Brown (8-18), and Penn (6-20). In the 2000s, Non-HYP ivies outscored all the big name public universities. Such trend will continue if current formula (used in 2009) remained unchanged.</p>
<p>Table 1 Overall trend 1: rising of Non-HYP ivies versus falling of big name public universities:</p>
<p>In the 1980s: </p>
<p>Dartmouth (6-10), Brown (7-13), Cornell (8-14), Columbia (8-18), Penn (15-19)
Berkeley (5-24), Michigan (7-25), UNC (9-23), UVA (15-20), UIUC (8-20), Wisconsin (13-23), UCLA (21), UTAustin (25) </p>
<p>In the 1990s: </p>
<p>Dartmouth (7-10), Cornell (6-15), Columbia (8-15), Brown (8-18), Penn (6-20)
UVA (17-22), Berkeley (13-27), Michigan (17-25), UCLA (16-28), UNC (18-27), UIUC (40+), Wisconsin (30+) , UTAustin (40+)</p>
<p>In the 2000s: </p>
<p>Penn (4-7), Columbia (8-11), Dartmouth (9-11), Cornell (10-15), Brown (13-17)
Berkeley (20-21), UVA (20-24), Michigan (22-27), UCLA (24-26), UNC (25-30), UIUC (30±40+), Wisconsin (30+) , UTAustin (40+)</p>
<p>^^^better flow
rjk:</p>
<p>As mentioned on post #20, in 1989, USNWR introduced new empirical formula, which weighed in other important academic criteria e.g., faulty resources, acceptance rate, class size, etc. Such adjustment had profound impact on top 25 and top 25 had never been the same ever since. From Table 1, we may observe the trends on the rising of Non-HYP ivies and falling of big name public universities from 1983 to 2009. </p>
<p>In the 1980s, Berkeley (5-24), Michigan (7-25), and UNC (9-23) matched well with Dartmouth (6-10), Brown (7-13), and Cornell (8-14); UVA (15-20), UIUC (8-20), and Wisconsin (13-23) were as competitive as Columbia (8-18) and Penn (15-19) In the 1990s, only UVA (17-22) and Berkeley (13-27) could complete with Columbia (8-15), Brown (8-18), and Penn (6-20). In the 2000s, Non-HYP ivies outscored all the big name public universities. Such trend will continue if current formula (used in 2009) remained unchanged.</p>
<p>UVA’s strong performance in the 1980s and 1990s help secure its ranking at tier 4 (or number 14-20).</p>
<p>In the 1980s: UVA (15-20), Average 18</p>
<p>In the 1990s: UVA (17-22), Average 20</p>
<p>In the 2000s: UVA (20-24), Average 23</p>
<p>Table 1 Overall trend 1: rising of Non-HYP ivies versus falling of big name public universities:</p>
<p>In the 1980s: </p>
<p>Dartmouth (6-10), Brown (7-13), Cornell (8-14), Columbia (8-18), Penn (15-19)
Berkeley (5-24), Michigan (7-25), UNC (9-23), UVA (15-20), UIUC (8-20), Wisconsin (13-23), UCLA (21), UTAustin (25) </p>
<p>In the 1990s: </p>
<p>Dartmouth (7-10), Cornell (6-15), Columbia (8-15), Brown (8-18), Penn (6-20)
UVA (17-22), Berkeley (13-27), Michigan (17-25), UCLA (16-28), UNC (18-27), UIUC (40+), Wisconsin (30+) , UTAustin (40+)</p>
<p>In the 2000s: </p>
<p>Penn (4-7), Columbia (8-11), Dartmouth (9-11), Cornell (10-15), Brown (13-17)
Berkeley (20-21), UVA (20-24), Michigan (22-27), UCLA (24-26), UNC (25-30), UIUC (30±40+), Wisconsin (30+) , UTAustin (40+)</p>
<p>Hawkette is that you?</p>
<p>Blah, blah, blah. Do you really not see what a complete joke this ranking stuff is on so many levels? Tier 1, tier 4, it makes me laugh very hard.</p>
<p>Old thread but still, this is quite funny. </p>
<p>@fallenchemist, liao is modelling the movements of schools’ rankings based on USNWR’s evaluations of them. You can’t argue that his model is incorrect because you think US News sucks at ranking schools. Your beef is with the wrong person.</p>
<p>You can argue with how the tiers are structured etc in the model but these opinions are all subjective.</p>
<p>modelingLiao wrote:
"As summarized on post #6, based on findings presented in post #1, there were five distinguished tiers (levels) among US News’ favorite national universities (top 25):</p>
<p>Tier 1 (1-5) HYPSM (5 in total)</p>
<p>Tier 2 (6-10) Caltech, Chicago, Columbia, Duke, and Dartmouth (5 in total) </p>
<p>Tier 3 (11-13) Penn, Brown, and Berkeley (3 in total) </p>
<p>Tier 4 (14-20) Cornell, JHU, Northwestern, Rice, UVA, Emory, WUSL (7 in total) </p>
<p>Tier 5 (21-26) Notre Dame, CMU, Michigan, Georgetown, UNC, Vanderbilt (6 in total)"</p>
<p>I know this forum is old and dead now, but I’d like to point out, modelingLiao, that your assessment in post one is not completely accurate. If your ranking for Tier 3 universities includes schools that have ranked in the Top 10 for at least two of the three periods you defined, then Rice should be Tier 3, not Tier 4, since USNWR ranked Rice 9th in 1989 and 10th in 1990, as seen here: [U.S</a>. News Rankings Through the Years](<a href=“http://web.archive.org/web/20070908142457/http://chronicle.com/stats/usnews/]U.S”>U.S. News Rankings Through the Years)</p>
<p>Duncaroo：</p>
<p>In order to observe the dynamics of the top 25 schools, I subdivided the USNews rankings into three periods:</p>
<p>First periods fell between 1983 and 1989(Y1980s),
Second periods between 1990 and 1999(Y1990s), and
Third periods between 2000 and 2009(Y2000s);
And then averaged them to obtain Rank1980s, Rank1990s, and Rank2000s (see post #1s tables for details). </p>
<p>Schools that ranked top 10 for at least two of the three periods were defined as institutions which ranked top 10 for at least two of the three rankings: Rank1980s, Rank1990s, and Rank2000s. </p>
<p>Rice ranked 12, 16, and 17 for Rank1980s, Rank1990s, and Rank2000s, respectively; it belongs to tier 4, a highly competitive level.</p>
<p>^In case it’s not obvious to you, your breakdown has serious flaws. For example:</p>
<ol>
<li><p>Northwestern/Cornell were ranked higher than Berkeley every year in the last two decades yet they are tier 4 while Berkeley is tier 3. </p></li>
<li><p>Northwestern/Cornell were top-10 multiple times across two decades and were always ranked higher than UVA. Northwestern has been consistently ranked 10 spots higher than UVA in the last decade. Yet you lumped UVA into the same tier as NU/Cornell.</p></li>
</ol>
<p>Seems pretty distorted to me.</p>