<p>That wasn’t your point. You were attempting to claim that the UC’s are outright lying.</p>
<p>If you have a problem with the top 10% metric blame US News, not the UC’s</p>
<p>No, no – If you’ll recall, my original point was that any ranking system that has UC SB, Davis, and Irvine in the top 50 was a joke, namley USNWR. But, alas, we digressed.</p>
<p>CH,
On your post #253, were you ranking the schools by Math 25th percentile?</p>
<p>
</p>
<p>Not plausible. According to the College Board, only 4% of 2008 college-bound seniors who took the SAT were internationals, yet there were approximately 16,000 more 750+ scorers on SAT M than on SAT CR—i.e., about 65% more 750+ M scorers than 750+ CR scorers. It’s likely internationals make up some part of that difference, because on average internationals scored slightly higher than US citizens on SAT M (+ 65 points, on average), and on average internationals scored about 20 points lower on CR than US citizens (remember, “internationals” in this context includes Canadians, Brits, Australians, and others for whom English is their first language, and many more quite proficient in English). With that small a number of internationals taking the test, that big a gulf between CR and M scores, and with international and U.S. mean M and CR scores being as close as they are, I just find it wildly improbably that there could be so many more internationals scoring high in Math to account for the difference.</p>
<p>
</p>
<p>Not true. I count about 90 colleges with a 75th percentile Math score over 700, but only about 50 with a 75th percentile CR score over 700.</p>
<p>I suspect that there are more high math scores because it is more cut-and-dried than interpreting dense prose passages. Even I, as a straight humanities person and heavy reader, scored higher on the math SAT than I did on the CR.</p>
<p>As to why males scores higher in math, I think we need to look at the brain and the way the genders approach social situations. Males tend to be more attuned to a logical ordering of facts, while females try to look deeper for innuendo. A typical female reaction to getting a math answer quickly is, “It can’t be that easy. What did I miss?” This kind of thinking slows her down and makes her question her own competence, thus leading to second guessing. Boys tend to be more confident. BTW, this same kind of thinking makes it difficult for girls as a whole to take standardized tests, no matter what the test measures. </p>
<p>That’s NOT to say that girls don’t get 800s on math, or that boys don’t get 800s on CR. Or that one gender is brighter than the other as a whole. </p>
<p>I’m a big fan of the writing section because it supposedly measures critical thinking and ease of expression under pressure, two skills that the best students possess. In a classroom, I can generally tell which undergraduates get the best grades by their critical thinking skills. Provided that the writing section truly succeeds in measuring these skills, I think it will provide a much better guide than either the M or CR sections.</p>
<p>GoBlue81-
Yes, in post #53 I first ranked on SAT 25th and then on SAT75th if the 25th was tied.</p>
<p>bclintonk-
To identify the colleges with SAT 75th percentile over 700 I used the IPEDS database. What did you use? I excluded community colleges. I excluded the baccalaureate/associates Carnegie Classification. I restricted to 7 Carnegie Classifications among doctoral, masters, and baccalaureate schools. My sense is that math skills are much less common than language skills. I also think that high math skills and high language skills generally go hand-in-hand among native speakers of English but not vice-versa. This is based on personal experience going through school. </p>
<p>If there are more high math scorers on the SAT it could also be that the scaling on the SAT is designed to produce more math high scorers or designed to discriminate better in the lower ranges than in the higher ranges. </p>
<p>I think high math skills are a scarcer commodity than high language skills, SATs notwithstanding.</p>
<p>It makes perfect sense that high Math SAT scores correlate well with high PA scores. PA, for better or for worse, is somewhat tied up with how well a university is perceived as a research institution. Schools with strong research output tend to be science and engineering powerhouses and also tend to have strong PA scores. Those schools also attract students who want to pursue science and engineering. Those students tend to have better SAT math scores.</p>
<p>bclintonk-
I did my search over again of schools with high SAT scores except this time I said greater than OR EQUAL TO 700. This time I searched all Carnegie Classifications. I found 94 schools with math SAT >= 700 and 88 schools with CR SAT >= 700, about the same.</p>
<p>Hey, by the way, this year the survey (at least the one sent to National Universities) also has three separate sections:</p>
<p>“Up and Coming Institutions”
“Commitment to Undergraduate Teaching” (I think I just heard a squee from hawkette)
“Exemplary Undergraduate Programs”</p>
<p>Just to clarify…
I think it is important to evaluate many aspects of a school, not just one. But SATs are a good proxy for several indicators of quality,</p>
<p>Based on my current analysis, for example, the correlations with math SAT 75th percentile were as follows:</p>
<h2>universities</h2>
<p>math SAT 75th and:
.91 SAT CR 25th
.93 SAT CR 75th
.85 peer assessment
.72 retention after one year
.82 graduation percent
-.69 admissions percent
.57 endowment per FTE
.51 research expenditures percent</p>
<h2>LACs</h2>
<p>math SAT 75th and:
.89 SAT CR 25th
.90 SAT CR 75th
.83 peer assessment
.79 retention after 1 year
.77 graduation percent
-.63 admissions percent
.59 endowment per FTE
.70 instructional expenditures percent</p>
<p>An interesting difference between universities and LACs is that SAT is highly correlated with research expenditures in the case of universities but highly correlated with instructional expenditures for LACs. However, many LACs lack a research budget.</p>
<p>hey mom: better not go there (or you could end up like Larry Summers). :)</p>
<p>But, wrt your point on the Writing section: the UC database indicated that the OLD Writing Subject test was the most predictive of Frosh grades (of all the tests), so you may be on to something…</p>
<p>
Note that USNWR did not ask “Excellence in Undergraduate Teaching”… ;)</p>
<p>
</p>
<p>Fwiw, I believe that some of the “new” elements were in last year’s survey --with one noteworthy result in the second place of Up-and-coming National Universities:</p>
<p>
</p>
<p>Do I still have to answer to the question … “What discrepancy?”</p>
<p>
</p>
<p>
</li>
</ol>
<p>^ That’s a typo on Smith’s CR range. The correct range for Smith’s CR is 590-710. Definitely lower than Harvey Mudd, but not nearly by the margin that the typo would have it appear.</p>
<p>Clemson has an outstanding baseball team…</p>
<p>Did they produced multiple versions of their roster … to figure out where they could move things around to make them look best?</p>
<p>
What is the rationale for ranking on SAT 25th (instead of 75th)? Does it correlate better with the university quality?</p>
<p>GoBlue81-
I didn’t have a reason for ranking first on SAT 25th except that it entered my mind first. Actually, SAT 75th correlates slightly better with measures of quality.</p>
<p>
</p>
<p>While the typo was obvious, it does not change anything that Smith had the lowest SAT from the selected list of 26 school listed:</p>
<ol>
<li>Bryn Mawr 620-730</li>
<li>Colby 640-720</li>
<li>Colgate 620-720</li>
<li>Bates 630-710</li>
<li>Smith 590-710</li>
</ol>
<p>Compared to Harvey Mudd, the margins remain 690-760 versus 590-710 for CR, and 740-800 versus 570-680 for Math, or margins of 100/800 and 170/800 at the 25 percentile and 50/800 and 120/800 at the 75 percentile.</p>
<p>And, the typo does not change anything about the validity of the claims of correlation between PA and SAT scores, and in particular, the claim that “Based of the analysis I did of PA, math SATs are the single best predictor of PA. I have long thought that math SATs are the single best index of college quality.”</p>
<p>As far as I am concerned, I maintain that the highest correlation of the PA will be found by combining the rankings of the previous couple of editions, and then adding a solid pinch of the Clemson integrity effect to a good dosis of geographical and historical cronyism – at least for the LACs.</p>