<p>Hawkette, of the 15 top employers of Duke grads, 10 of them are on the same few blocks in New York, one in Seattle (Microsoft), one in DC (Accenture). I don't think when discussing top schools employers are that biased towards their own region - they want top talent, not just from schools in the area.</p>
<p>
[quote]
But it still means something .
[/quote]
Very funny sakky... saying stuff like this doesn't make you any smarter.. of course it means something, just possibly what you don't think it means.. everybody knows this.</p>
<p>The RP is a fairly small sample from kids who attend Northeastern schools.. I will admit that it is more likely that a kid from the Southeast is going to pick Duke over an Ivy than someone from the Northeast.</p>
<p>
[quote]
I'm not particularly convinced that a clear Northeast bias exists that could not be explained by the undeniable fact that the Northeast just happens to have a disproportionate number of the best schools in the country.
[/quote]
Also it isn't fair to say, "LOL! THIS SURVEY IS FAIR SINCE ALL THE SMART PEOPLE ARE IN THE NORTHEAST ANYWAY!!!" in the South, it's not like everyone there is stupid... last year at the ARML competition, math teams from states such as Virginia (does TJ School of Science ring a bell?), North Carolina (NC School of Science and Mathematics? Anyone?), and Georgia were taking top prizes.</p>
<p>Also you must remember that it is also a small sample.
I know, that those who fight against Duke hate to hear personal anecdotes, but I know of a hell lot of people who turn down non-HYP Ivies to come to Duke just as I know a lot of people who turn down Duke for non-HYP Ivies.</p>
<p>The RP is somewhat flawed since it doesn't deal with kids all around the country.. of course, when you come to it (due to the fact that a lot of kids like to stay close to home).. places like Brown, Dartmouth, and Penn will have a high preference.</p>
<p>It was never meant to be a ranking of schools..</p>
<p>The article has no focus. I don't see what the writer is trying to do. Is he advocating that Ivy-caliber students to go to Stanford, Duke, Chicago, Northwestern, etc...? If so, then those cost just as much as an Ivy League school. Ditto for state schools if an applicant is out-of-state.</p>
<p>Is he saying that Ivies aren't undergrad-focused enough? Then he better throw out just about any school that isn't an LAC (and those are expensive, too).</p>
<p>Grad schools don't favor Ivy grads? Yes they do. The author takes an incredibly simplistic view of the process, and doesn't account for student interest. Obviously the kid at Reed or Oberlin or Swarthmore will be more "academic" than the average Penn or Cornell kid (due to their large pre-professional programs) and thus be more inclined to go to grad school. I seriously doubt that Ivy kids are not favored over almost every other college.</p>
<p>Rankings of specific depts. don't favor the Ivy League? Firstly, as a whole, they do. And secondly, the only other schools with a presence in those rankings are schools like Chicago, Duke, Stanford, Michigan, and Berkeley...hardly easy to get into or cheap (OOS).</p>
<p>The Ivy League scares kids to death? Guess what -- those kids don't apply.</p>
<p>All in all, this article is clearly just another article in a long list of many to help parents rationalize the fact that their kid won't get into an Ivy, or to help overly-ambitious kids deal with that same fact.</p>
<p>
[quote]
the undeniable fact that the Northeast just happens to have a disproportionate number of the best schools in the country.
[/quote]
</p>
<p>sakky, on this I totally agree. And, you might add, that the differential between the publics available in the NE and the top tier privates available in the NE is HUGE, unlike highly ranked publics available in other states, particularly Calif, Tx, Mich, NC, and Va. The end result is an extreme demand for private (Ivy) education in the NE corridor by NE residents. And, we know statistically, that most kids (and/or their parents) prefer to attend college within 250-400 miles of home, or a day's drive. Thus, most NE residents would prefer a NE college over Stanford for example, if for no other reason than it is closer to home for NE residents.</p>
<p>Even Homer Simpson could figure that out!</p>
<p>
[quote]
not what NE prep school students (10% of which probably apply to top 10 schools anyways) think
[/quote]
</p>
<p>
[quote]
The RP is a fairly small sample from kids who attend Northeastern schools..
[/quote]
</p>
<p>
[quote]
The RP is somewhat flawed since it doesn't deal with kids all around the country
[/quote]
</p>
<p>Uh, I don't understand what you guys are complaining at. First off, I don't see what this necessary has to do with NE prep school students. It seems to me that public schools were indeed included within the data, as can be evinced from the following quote:</p>
<p>"Counselors at public schools selected students at random
from the top 10 percent of the senior class,</p>
<p>Now, granted, it is true that NE prep school data was indeed included, and perhaps overrepresented. But shouldn't it be? After all, NE prep schools tend to be among the best high schools in the country, with many of the top students in the country. The RP student explicitly states that it is attempting to measure the RP's of only the top students. Hence, NE prep schools should probably be overrepresented. But by no means do they represent ALL of the data.</p>
<p>Secondly, I don't see any evidence to indicate that the study looked only at students at Northeastern schools. In fact, the authors explicitly state the opposite:</p>
<p>"The final sample contains students from 43 states plus the District of Columbia."</p>
<p>Of the 7 missing states, they don't seem to be particularly concentrated in one region, except for N and S Dakota (but come on, these aren't exactly the most populous states in the country). </p>
<p>Now, unless you want to accuse the authors of lying, I think you have to agree that the claim that the study looked only at students who attend Northeastern schools is flat-out wrong. I said it before, I'll say it again: * Please read the paper before you complain about it *. </p>
<p>Now, to the contention that the statistical sample size is small - of course it is true. But like I've always asked, does anybody know of another RP student that has a larger sample size? Or any other ranking system that can be said to demonstrate statistical significance? I don't think so. So for all its flaws, the RP is still better than the other rankings out there. </p>
<p>But besides, to your point that students prefer schools within their region, all you have to do is turn to page 46 of the survey where RP's are broken down by home region of the student.</p>
<p>
[quote]
Also it isn't fair to say, "LOL! THIS SURVEY IS FAIR SINCE ALL THE SMART PEOPLE ARE IN THE NORTHEAST ANYWAY!!!" in the South, it's not like everyone there is stupid... last year at the ARML competition, math teams from states such as Virginia (does TJ School of Science ring a bell?), North Carolina (NC School of Science and Mathematics? Anyone?), and Georgia were taking top prizes
[/quote]
</p>
<p>And who has here has ever said that the South doesn't have any smart students? Of course it does. That's not the point. The point is, the Northeast has a * disproportionate number * of the top colleges. It's an undeniable fact that cannot be ignored. </p>
<p>But I'll say it again. If you prefer a more regional approach, just turn to p.46 of the RP to see the preferences broken down by student home region.</p>
<p>thoughtprocess, there is one thing that you keep ignoring. Duke's yield is eh around 40 percent each year. Penn's is consistently around 60-65, right near Princeton and other places of that caliber. If Duke is such a hot school, why do so many kids turn it down? That suggests something, doesn't it.</p>
<p>^Penn has many specialized programs such as Wharton and Huntsman which have near 100% yield.. try finding out Penn's yield with only CAS & Engineering.. you will have some number higher than 40-45, but you will get something lower than 60%.</p>
<p>Yowza, I think this whole argument sums up what the author of the article is talking about. I was fortunate enough to get into Princeton. The school has its faults, and after my first year it many of them became apparent to me, but its just amazing how many things I can study. However, I honestly think I would be just as happy at the University of Chicago, Carleton College, or UC-Santa Cruz. I went to a Mid-Atlantic school, and my teacher and friend really thought I would be a good fit at Princeton. The rest is history. I know it seems futile but I notice the hysteria when I talk to high schoolers about college, and its just a shame that everyone is under so much pressure from their parents, their peers, and misconceptions of the meaning of "success" that weigh people down. I've even talked to one student who told me not getting into an Ivy would be his personal failure, and his achievements would be "wasted". I'm not saying anyone is to blame but the colleges themselves, but I think the article really does ask us all the important question of why this 'panic' is occurring and what can we do to stop it.</p>
<p>Uh, Columbahopeful - the students that enroll at Duke are just as strong (and stronger in some areas) than Penns - that tells you something too, doesn't it. Look at SATs, National Merit Scholars, and feeder rates. Of course, you won't look at those because it proves your point of view is void of factual information while the only thing you repeat, over and over again, is yield rates.</p>
<p>A lower yield certainly doesn't mean that students prefer it less than a school with a higher yield - it may mean that it cross-admits in high rates with other selective universities (including HYPSM) and loses them. It may also mean less are accepted Early Decision. I mean, yield, acceptance rates, and matriculation rates are also a factor of geography, how many students apply, and how many major cities are nearby. All of this points to the fact that when NUGrad and Columbiahopeful's entire argument is based on Duke having a lower yield...there is something weak about that argument.</p>
<p>I mean, since when does the quality of students who turned down a school take more importance than the quality of students who enroll? The ones who enroll are the ones who count, and Duke's student body is extremely strong, comparable to every other top 10 school in every way, though still behind HYPSM. </p>
<p>The fact that Duke gets students with higher test scores, more national merit scholars, and sends those kids to top Ivy grad schools at similar rates says something...but, as usual, this is consistently ignored by people from the NE who refuse to acknowledge facts (such as Columbiahopeful and NUGrad). Its a joke to say Duke isn't as good as non-HYP Ivies, in any way. </p>
<p>Sakky - my point is, the sample size is small, the data is 10 years old, and doesn't actually reflect the quality of the schools themselves, at least when talking about Duke which enrolls students as or more competitive than its peers (non-HYP Ivies). Of course the RP means "something" - I meant relevant to arguing how strong Duke is compared to Penn. The perceptions of high school seniors a decade ago with an average SAT of 1300 (100 some points below accepted student range) doesn't reflect how strong a university is.</p>
<p>Columbiahopeful and Nugrad - I've noticed that the only evidence you have of Duke not being up to par with the Ivy League is its lower yield. You ignore the Duke Ugrad office stating (without argument by its competitor schools) that it splits applicants evenly with Penn, Columbia, Dartmouth, etc. You ignore the fact that students who attend Duke are empirically as strong or stronger. Why is that? Just out of convenience?</p>
<p>So, you put more importance on the number of students who attend out of those who are accepted, rather than the quality of the students who attend. </p>
<p>What if a school rejected top applicants and got 100% of mediocre applicants? Would that make it better - by using yield as the ONLY part of your argument, thats what you are saying. NUGrad or Columbiahopeful have only cited yield to backup their argument that Penn is better than Duke, whereas I've cited everything else to say that Duke and Penn are even.</p>
<p>And, in note, the researchers in RP themselves pointed out that the high SAT scores and the fact that there were national merit scholars in their sample showed how strong the student sample was, and why their study should be considered an attempt to gauge top students.</p>
<p>
[quote]
And if he indeed got into Columbia and Dartmouth and Penn, to me, he didn't make the best decision. To him, that could be the case. However, we are all entitled to our opinions, aren't we?
[/quote]
</p>
<p>Haha.. you admitted here that when it comes to the non HYP-Ivies and Duke.. it all matters by opinion! :D :D That's it!!!!! </p>
<p>Also.. I don't have a lot of proof for this.. but I know that schools such as Columbia and Brown are infamous for wl'ing/r'ing HYPSM acceptees.. Duke doesn't play that game.. may say a lot about the yield rates.</p>
<p>Duke gets a few in through merit scholarships and a few just through the fact that Duke seems like the better fit..</p>
<p>My opinions:</p>
<p>1.) Revealed preference from employers would make more sense.</p>
<p>2.) The study does not seem to have a Northeast bias.</p>
<p>3.) Duke almost certainly does yield protect, as does basically everybody, Princeton (apparently) included.</p>
<p>4.) The RP study needs to be updated. It's old.</p>
<p>5.) The RP study is more powerful than actual cross-admit data because of exactly what sakky has said: you have to account for the students who don't even bother to apply.</p>
<p>6.) The RP study is a popularity contest, but that is useful information.</p>
<p>sakky:</p>
<p>where in the study does it show the actual numbers of students by state/region? Is the west region RP based on a sample of 5, 50, 500, or 5,000? </p>
<p>(Personally, I find the lack of such a data rather surprising, and the cynic in me suggests that there is an must be a reason why.)</p>
<p>
[quote]
where in the study does it show the actual numbers of students by state/region? Is the west region RP based on a sample of 5, 50, 500, or 5,000?
[/quote]
</p>
<p>It doesn't show the data. Just like no study/ranking shows you all the raw data. Few studies out there will actually show you all the raw data. Where is the raw data for USNews?</p>
<p>If you want the raw data for the RP, you can probably ask for it from the authors.</p>
<p>"no study shows...raw data", even if possibly relevant?</p>
<p>You gotta be kidding, yes? What classes at Cal did you take? (You must have missed Purves & Pisani.)</p>
<p>Would you want NEJM to post datasets including fifty variables on 4000 patients?</p>
<p>The NEJM, and JAMA for that matter, DO discuss patient population, country of origin, states, medical centers included in the study, etc., to provide the background for conclusions. Indeed, how do you think medical outcome studies are performed? How does one find that, say, heart bypass surgery is much higher in a certain region of the country than another? Or that hip replacement is 2x more likely in xx state?</p>
<p>btw: an early draft of the RP study DID include a list of high schoools included, but it got dropped out of the final edits. (Bad me for not saving it.)</p>
<p>One of the key criteria of the scientific method (taught beginning in middle school) is data Description, so the study can be peer reviewed, and replicated by others. Even the authors (and sakky) admit RP is based on a small sample, so how can other researchers add to the knowledge base if they don't know the data inputs?</p>
<p>I don't have an issue with the modeling, which I think does bring new insight to the rankings game. But, we know several facts, and, I think a logical conclusion:</p>
<p>Facts:
1) The authors started out to survey the preferences of students attending highly selective schools.
2) (Even sakky admits that) the majority of highly selective schools are in the NE.
3) The majority of kids attending NE schools are from the NE (including mid-Atlantic), altho Calif is well-represented at most.</p>
<p>My conclusion/speculation:</p>
<p>Given numbers 2&3, the authors HAD to oversample NE high schools to obtain data for #1, their primary goal. How could they not?</p>
<p>1.) Yes, but nobody posts their data in "raw data" form, which would be a massive spreadsheet that would take up 80% of the journal. I agree that most studies certainly do post some background characteristics, but this is not raw data by any stretch of the imagination.</p>
<p>2.) Again, so it's a popularity contest among highly-selective schools. If your question is Northeast-dominated, then for the survey to be Northeast dominated is not a "Northeast bias", it's an accurate means of answering the question.</p>
<p>
[quote]
"no study shows...raw data", even if possibly relevant?</p>
<p>You gotta be kidding, yes? What classes at Cal did you take? (You must have missed Purves & Pisani.)
[/quote]
</p>
<p>
[quote]
The NEJM, and JAMA for that matter, DO discuss patient population, country of origin, states, medical centers included in the study, etc., to provide the background for conclusions. Indeed, how do you think medical outcome studies are performed? How does one find that, say, heart bypass surgery is much higher in a certain region of the country than another? Or that hip replacement is 2x more likely in xx state?
[/quote]
</p>
<p>Uh, like bluedevilmike said, no article posts their "raw" data. If you want it, you can ask for it. This is a well-established academic protocol of form. </p>
<p>Besides, I don't know why you're invoking the NEJM or JAMA anyway. This is not a medical study. This study is likely to end up being published in an economics, sociology, or public policy journal and hence the study should be judged as to whether it comforms to the format that those kinds of journals demand. Just pick up a copy of, say, the American Economic Review or Adminstrative Science Quarterly and notice how many (probably most) of the papers do not include significantly more information regarding their dataset than what the RP study included. </p>
<p>The notion of replicability is not relevant for this study or any statistical preference study for that matter. After all, think about what that would mean. To truly replicate a statistical preference study, you would have to find the * exact same * subjects and ask them the * exact same questions * as the researcher did. And even if you were somehow able to do that, the answers they would inevitably be different because, if for no other reason, they've already been asked those questions before and hence have had time to think about it and perhaps change their mind. No statistical preference study is truly replicable. </p>
<p>Replicability is reserved for pure scientific experiments where you can actually go into the lab and repeat the precise steps to get the results that the researcher described. Atoms are atoms, molecules are molecules, gravity is gravity, etc. so if somebody claims that you can perform cold fusion in your kitchen and provides you with the mechanism, you should be able to reproduce the experiment exactly the way it was described. But not only is every person in the world is different, every person changes with time. In contrast, a water molecule doesn't care what experiments it has been subjected to in the past. It has no path dependence. </p>
<p>Regarding the notion of peer review, one accepted protocol is to simply provide all of the raw data * to the reviewers *. But not within the published article itself, for, as bluedevilmike said, raw data would then take over the vast majority of the journal. </p>
<p>And besides, bluebayou, like I have always said, you are using criteria that you are not applying to any of the other rankings out there. Where is the raw data for USNews? Where is the raw data for Gourman? Why aren't you asking for that? Why single out RP? </p>
<p>
[quote]
1) The authors started out to survey the preferences of students attending highly selective schools.
2) (Even sakky admits that) the majority of highly selective schools are in the NE.
3) The majority of kids attending NE schools are from the NE (including mid-Atlantic), altho Calif is well-represented at most
[/quote]
</p>
<p>Uh, I never said anything about 'majorities'. I said possible over-representation. But I didn't say anything about a 'majority'. </p>
<p>But like I said, isn't overrepresentation to be expected? After all, the study explicitly examines top students. And let's face it, NE schools tend to have a disproportionate percentage of top students. Similarly, I would not be surprised to find that a study that examined tall people would disproportionately include professional basketball players.</p>
<p>
[quote]
Besides, I don't know why you're invoking the NEJM or JAMA anyway.
[/quote]
</p>
<p>Perhaps you should follow your own advice. READ the posts. Bluedevilmike invoked NEJM. </p>
<p>your argument about replicability is specious.</p>
<p>I am NOT applying any criteria to this ranking that that I don't apply to others. I've posted frequently that they are ALL crap. But, you seem to think RP is best of breed.</p>