<p>No, that's not what you have said to this point. You lose.</p>
<p>In response to that, I have to admit that it is rather impressive in my book for a place like Penn to enroll a student body that is relatively comensurate in quality as a place like Dartmouth where there are 1/2 the students. But, let's be honest here, if Harvard increased its class size to 4000 per grade, it would still be able to enroll one of the most talented student bodies out there.
This also raises an interesting tradeoff. The smaller the student body also means less alumni scattered in the country. So, while school x with only 1000 per grade (i.e., Chicago, Dartmouth) may boast a very impressive student body, it might actually be better off it its student body was a bit larger so there were more alumni around the country.</p>
<p>arbiter213 wrote: "When is that data from? It doesn't seem exactly accurate."</p>
<p>I was on the Princeton Review site today and manually searched on each school's profile.</p>
<p>
[quote]
Yes, the top 1000 at Cornell surpass in SAT scores the top 1000 at the schools you named with the possible exception of Penn, because Penn also has a large student body.
[/quote]
</p>
<p>I don't believe that this can be correct.</p>
<p>1) Let's take the arbitrary 1000 number for a moment.
2) Let's assume for statistical simplicity that the top 1000 students at any given school are evenly distributed amongst all four classes at any given point in time (i.e. frosh, soph, jr, sr classes)
3) This would dictate that there are 250 "top" students for every incoming class at any given school
4) MIT, for example, has roughly 4,000 undergrads (i.e. a nice round number to fit into this scenario - i.e 1,000 incoming freshman --> the top 250 students representing the top 25%).
5) Therefore, what you are saying is that the top 250 incoming students at Cornell next year are statistically superior to the top 250 incoming students at MIT -- or the top 25% of incoming MIT students?</p>
<p>anybody happen to notice the 25th percentile at Caltech scored 780 in math and 1470 in combined verbal/math? that combined is 90 points higher than MIT's, 80 points higher than harvard/yale, and 50 points higher than lauded Harvey Mudd!</p>
<p>
[quote]
xiggi-
My statement about the selectivity of the engineering program at Cornell compared with MIT was based on SAT scores.</p>
<p>The overall SAT scores 25th-75th at MIT are 1380-1560 and the SAT range at Cornell engineering is 1360-1530.
[/quote]
</p>
<p>In that case, you're correct. </p>
<p>Were we to add the missing data for the top 10% and allocate the 50%, 40%, 10% weights, we might even duplicate a selectivity index </p>
<p>
[quote]
anybody happen to notice the 25th percentile at Caltech scored 780 in math and 1470 in combined verbal/math? that combined is 90 points higher than MIT's, 80 points higher than harvard/yale, and 50 points higher than lauded Harvey Mudd!
[/quote]
</p>
<p>Lauded Harvey Mudd? Oh my, Oh my, that's a LAC we are now talking about! Or should I paraphrase Conrad and scream ... the horror, the horror!</p>
<p>random thoughts.</p>
<p>-Penn's admit rate was 16.1%. How that rounds up to 18% I do not know.
-nice mark twain quote
-Penn is indeed twice as awesome as dartmouth for achieving the same selectivity with a ginormous class to fill ;)</p>
<p>arbiter-
I re-thought your point about the worst 6000 at Cornell being worse than the worst 6000 at ND, NU, Vandy, Duke. That is true. And, I am not sure the others would be worse than Cornell if they grew to Cornell's size. The only thing I will say is that the BEST at Cornell is better than the BEST at the other schools, head-to-head. That is usually how schools compete in sports, head-to-head, the best vs the best.</p>
<p>the_prestige-
I have not done the calculation for MIT but, yes, I am saying that the top 250 students at Cornell are very probably statistically superior to the top 250 students at MIT. The only exception would be if there are 250 students at each school with 1600 SAT scores, then they would be the same. It is more certain that the top 1000 would be superior at Cornell because it is unlikely to have 1000 perfect SAT scorers. </p>
<p>If scores approximate a bell curve distribution, then you can estimate the percentage of students who score above a certain point, knowing the 25th-75th percentiles. You can also reverse the calculation and figure out what percentage 6000 of the total student body would be, then calculate the SAT score that would have that percentage of the distribution above it. </p>
<p>There are online calculators to help with that calculation. "Area under the normal curve" calculators. But, the distribution doesn't exactly have to be bell-shaped for this to work. You just have to assume that most students score in the middle and fewer and fewer score near the extremes.</p>
<p>Actually I think you're only as good as your weakest link. The weakest 4000 at Cornell are leagues below the weakest 4000 at Dartmouth. ON AVERAGE, a Dartmouth or Columbia student is stronger than a Cornell student. </p>
<p>Your logic makes no sense to me.</p>
<p>slipper-
Think of it in terms of money. For example, the per capita annual income in DC is about $55K and in California it is $37K. Which is wealthier? It makes more sense to say that California is wealthier than DC even though California's average is lower. Which place would have more millionaires?</p>
<p>All this talk about top 25%, bottom 25%, etc. is only relevant if the student can control with which group they take classes, dorm, eat, and socialize.</p>
<p>If one <em>could</em> do all the above with only the top 25% at a larger school, then these arguments hold. An example I can think of would be a subset of a larger school, such as the Wharton School at Penn, the Honors curriculum at some state schools, the Dept. of engineering at a number of schools. The problem in these cases is that a majority of the contact with the other 75% is inevitable everywhere except the actual classroom, and therefore the CULTURE of the school is still defined by the broader population. This culture is significant -- attitudes about politics, learning, fitness; intellectual curiosity, eccentrism, social commitments, subject matter knowledge, the role that open discourse and exchange of ideas plays in the hallways and around the meal table.</p>
<p>My point is that it is extremely difficult to remove the bottom 75% from a person's experience, and therefore talk of the top 25% doesn't have much relevance.</p>
<p>
[quote]
Warning, very rough calculations (since I'm dividing total endowment over undergraduates)</p>
<p>UC Berkely: $3.5billion/~23,000 undergrads= ~$152,174 per undergraduate.</p>
<p>NU: $6.6 billion*/~7,800 undergrads = ~$846,154 per undergraduate
[/quote]
</p>
<p>Dont you think thats misleading because 1.) universities dont spend the entire profits earned from endowment funds to undergraduate education and 2.) UC Berkeley receives funding from the government aside from the earnings from endowment fund. Unless I see the real/authenticated figures, I would believe that Berkeley has bigger spending for educational purposes than NU has.</p>
<p>
[quote]
I don't know the statistics you're referring to, but it sounds like they were raw placement numbers. Now, if they were, and I'm no expert here, wouldn't it make sense that a school with 3 times as many undergraduates would have numerically more students placed?
[/quote]
No, but how sure are you that most of those Berkeley grads would want to go to grad school? </p>
<p>Lets compare apple to apple. Show me the figures for Berkeley engineering graduates applying to grad school and NU engineering graduates applying to grad school. Show me the figures for Berkeley and NU business and economics majors applying to grad school. Show me the figures for Berkeley and NU premeds applying to medical schools. Of course, cross-majors are not included so not to confuse when we compare schools. General and entirety measurement does not make sense here, because like I said, both are more or less equal.</p>
<p>
[quote]
Let's take Cal Berk as an example. I don't know Cal's exact enrollment figures. Let's say it has 5,000 undergrads per class. I believe Cal's SAT average is around a 1340. Let's assume that Cal has about 750 students per class who score above a 1500 on their SAT. According to the logic above, Cal has a better student body b/c it has more 1500 SAT scorers than does Dartmouth.
[/quote]
</p>
<p>I did check Berkeleys statistics and it says the average SAT scores are 1450, not 1350 as you mentioned and 99% of the students were from the top 10% of their high school. </p>
<p>Now, is that better or worse than Dartmouths student body? I dont know because its easy to manipulate SATs scores. You can take the exam twice and submit the better score and your chances of getting admitted to schools that are SATs craze would rise dramatically. </p>
<p>The only schools that have a clear better student body than Berkeley average-wise are the HYPSM and Caltech and probably Duke. The rest of the top schools though dont have a clear advantage than Berkeley in student quality.</p>
<p>Berkeley also doesn't "superscore" the SATs like some schools do.</p>
<p>^ What is superscore and why would universities do that?</p>
<p>See </p>
<p>The</a> Education Trust - Closing the Achievement Gap </p>
<p>and </p>
<p>The</a> Education Trust - Closing the Achievement Gap </p>
<p>for student-related expenditures per full-time-equivalent for various colleges according to standardized federal data definitions.</p>
<p>
[quote]
Berkeley also doesn't "superscore" the SATs like some schools do.
[/quote]
</p>
<p>Can you show impact from that? Where are the figures that show that that makes much of a difference in comparing colleges?</p>
<p>taken, how reliable are those data? </p>
<p>For example, UCLA's spending was around 30k while Cal's has only 18k. That's a huge gap right there. The endownment fund of UCLA is not any bigger than Cal's.</p>