New US News ranking out. We're number 10

[Duvall]

Just saying, you're only hurting yourself if you don't try to help your ranking.

Really? How?

[cspan37421]

I have some info on Wash U. I can't find the story, but it was clear that Wash U was gaming the system to get a higher percentage of students to accept its offers of admission. It was requesting info on other schools that the applicant had visited and/or applied to, then it was using a "degree of interest" variable to DQ they guys and gals it believed were headed to Harvard or Yale anyway. Total BS and disgraceful. Maybe we should do that too?

sagegrouse

That sounds like yield enhancement, i.e., trying to decrease the denominator in the ratio of students matriculating/students accepted. I don't follow USN&WR rankings at all, but others on college confidential have said that USN&WR no longer uses this measure. But the sense on that website was that multiple schools were doing this.

I don't see yield enhancement strategies as anywhere near as disgraceful as the counting requests for information as completed applications, which is what washu was accused of doing earlier in this thread. Yield enhancement through strategic declines has significant risks for the school ... they could lose otherwise highly desirable students. Plus, if word gets out, highly desirable students may be less likely to apply there. It's somewhat self-correcting, or at least, self-limiting. Lying about how selective you are has no direct downside.

WashU has long had a reputation for having taken shortcuts to achieve selectivity ... in the end, salting the applicant pool with kids they knew they'd never take becomes a self-fulfilling prophecy, for the low accept % attracts more top students, which then allows for more true selectivity, etc. They now attract a very selective group of students, unless they're lying about their students' average ACT/SAT, class rank, etc.

Those interested in alternative, objective rating systems might google the "revealed preference ranking" paper by Caroline Hoxby, et al. This paper demonstrates a ranking system based on applicant preferences when dealing with multiple acceptances. For instance, a student accepted to Duke, Yale, and Stanford who chooses Duke gets coded as a win (1) for Duke and a loss (0) for Yale and Stanford. The method implicitly takes into account financial aid packages, merit or need, which can sometimes make a full ride at Vandy, for example, look more appealing to some than paying full freight at say Cornell or Duke. So it's an interesting ranking, kind of the ultimate popularity contest for colleges, but does not rely on gamesmanship or "peer reputation" and other things that can be easily manipulated.

[CrazyNotCrazie]

These rankings are a way to sell magazines, and the schools are forced to cooperate for fear of dropping in them. US News is virtually non-existent now, and these ranking issues are all that keeps them afloat. As in most rankings, they are directionally helpful but that's it - a school in the top ten is likely more competitive than one in the 70s (and defining competitive is the subject for a totally different discussion), but if a student chooses school # 21 solely because it is ranked ahead of school #22, that is a really lousy reason for choosing a school and shows a lack of judgment. The inferred preference concept mentioned above would be much more accurate, though it is hard to factor out issues like finances, geography, etc.

[johnb]

I work at an Ivy graduate-level institution that has been loathe to make any changes based on US News, and we've watched our ranking slip over the past 20 years. Finally, we asked a peer institution for suggestions, and they said that after the first US News rankings came out, they convened a special committee and hired several full-time experts whose mission was to improve their ranking. Fifteen years ago, our rankings were adjacent; now, despite many improvements on our campus, we're quite separated. Suffice to say, we now have a committee.

While I don't think we want to lie or game the system, we do have to take a look at certain policies. For example, we get credit for selecitivity and average test score. Should we jettison our interest in affirmative action, in atypical applicants, in service-oriented people with a lot of B+'s? Every time we take someone without a 3.9 and high test scores, we are hurting our ranking. And when we hurt our ranking, we lose out on applicants who go to places ahead of us. One can argue admissions policies, but I'm not sure the decision should be made based on US News.

[cspan37421]

Should we jettison our interest in affirmative action, in atypical applicants, in service-oriented people with a lot of B+'s? Every time we take someone without a 3.9 and high test scores, we are hurting our ranking. And when we hurt our ranking, we lose out on applicants who go to places ahead of us. One can argue admissions policies, but I'm not sure the decision should be made based on US News.

So if I understand you correctly, it sounds like a conundrum: by seeking a certain mix of students, you fall a little short on certain measures ... which leads to lower rankings, which leads to fewer top students being interested in your school, which in turn drops you more in the ratings, etc. Almost reminds me of an insurance rate spiral. Well, if you have too few top students attracted, perhaps you have to raise the bar on the service oriented people. Or, let more top students in ED and fill the rest as RD. I'm no expert in this - sounds tricky. But I wonder, are the schools getting ahead of you just getting grade grubbing automatons? In a way, so much the worse for them.

crazynotcrazie, you're right about factoring out geographic preference etc. Not sure if you meant it this way, but the revealed preference ranking implicitly takes into account all factors that go into choosing one school over another, assuming you were accepted to both. This includes cost, geography, vibe, prestige, quality of dorms and food, religious preferences, student life, you name it. None of it is teased out separately* - it just shows the net effect of kids' decisions. Duke was 19th in this (published 2005), behind Williams and Rice, behind all Ivies, LACs like Swarthmore, Wellesley, Amherst, plus Georgetown, Notre Dame, MIT, Stanford. You may find it interesting that the list began 1. Harvard 2. Caltech.

Keep in mind that this is an aggregate of student opinion, wins and losses vs. other schools, when admitted to both. So if you were admitted to Wellesley, you probably weren't neutral about single-sex education. If you applied to Georgetown and Notre Dame, you may have preferred a Catholic school. MIT and Caltech - if you applied there and Duke, you probably were seeking an engineering degree and it's not surprising if those schools win out more than Duke does in a head-to-head. But financial aid could be a big factor. All things are NOT equal in these head-to-head comparisons ... and that's the point. For whatever reason, the 18 schools ahead of us edged us (from barely to frequently) in decisions made by kids admitted to both.

Pretty fascinating stuff.

* - the factors are not teased out. But they do break down revealed preferences by some sub-categories (by region, by intended major in the humanities vs. sciences)

[Bluedog]

I personally agree that the school should seek to improve itself on measures it sees most important first and foremost rather than the measures US News deems as important. Obviously, improving in the ranking certainly doesn't hurt, however, so doing things to help Duke's standing without compromising values should be done. Some points in this regard:

1.) Yield is not included in the ranking, but does have some significance to students when they appeal as to the desirability of the school. However, it is not the end all be all. BYU has a sky high yield and it doesn't rank particularly high. Duke's yield of ~44% is quite a bit behind all the Ivies except Cornell where it's closer. One could argue that one reason Duke's yield is on the relatively lower end is that Duke is the only top 15 private school (save Stanford, but California has 4x the population of NC) with a top 10 public school in the same state. NC has the second most applicants to Duke (behind California) and I'd imagine a lot of those in-staters choose UNC over Duke if they have to pay full price. UPenn, Dartmouth, and Brown all are 52%+, with Penn having a particularly high yield of 66% (8th in the nation and just 2% behind Princeton) by admitting 50% of its class under ED. Duke "only" accepts about 1/3 of its class through ED.

• However, Duke cannot employ the UPenn strategy without significantly decreasing the quality of its students most likely. Why? Well, UPenn gets nearly 4,600 ED applications, Duke only gets 2,200. That's right, more than twice as many people apply to Penn early than Duke. So, in the end while Penn fills a very sizable portion of its class with early applications which allow for 100% yield, the acceptance rate for ED to both schools is roughly the same. Penn's is actually even a bit lower at 26% vs. Duke's 29%.

• Obviously, certain schools have a priority to improve their yield - WashU seeks to improve its yield by utilizing two techniques:
  
  
  • The first is to not admit student applicants that they believe are using WashU as a backup and waitlist them (AKA waitlist candidates unlikely to attend). WashU is basically the only top school that doesn't release its common data set (CDS) to the public and refuses to release all its acceptance information to the public. Thus, this is largely anecdotal since statistics are kept under wraps. Duke has also been accused of this previously (not releasing everything), but revealed waitlist numbers and CDS figures the past few years. But, from what I've seen, the very very top students are OFTEN waitlisted at WashU (rarely rejected) and accepted to some of the Ivies, Duke, Stanford, MIT, etc. The near the top students (who are still very high quality) that apply to WashU are more likely to accept an offer of admission. If you then profess great interest after getting on the waitlist, then WashU may take you, again effectively increasing their yield since they don't take you off the waitlist until you indicate your high level of interest and likelihood to attend.
  • The other strategy WashU uses (which they explicitly state on their website) is to keep track of and consider interest. WashU keeps track of how many times you visit, write, call, etc. They state directly on their admissions website that interest is taken into account in decisions. Thus, they are more likely to accept you if you've visited and called multiple times as you will more likely be excited to attend. Duke explicitly states that they do NOT consider interest when evaluating a student, except they do consider ED interest a boost as opposed to RD. But all RD candidates are evaluated on their merits regardless of how many times they've visited, called, etc. Duke does not keep track of that information on file for each candidate. I like this as Duke should seek to have the highest quality class possible, while ED is enough to have plenty of students who are really excited to attend the school.

2.) Acceptance Rate is included in the ranking, but only accounts for 1.5% of it (10% of student selectivity, which is 15%). Thus, getting unqualified applicants to apply just so you can reject them helps very little in the ranking. So, while WashU may employ this strategy (by sending obscene amounts of mail to students who have no chance of admission), it doesn't have much affect on ranking. However, it does improve their selectivity which perhaps makes it more appealing to certain students. I visited WashU between the summer of my sophomore and junior year in high school (and greatly liked the school). Big mistake. At least, big mistake for my mail carrier's sake. I literally got 100 pieces of mail over the next two years and they had no idea how qualified or not I was for the school since I didn't send them any of that information. They have the hardest working mail staff in the nation.

• The Tulane strategy for reducing acceptance rate is also interesting, but hasn't really improved their ranking that much, but has certainly improved the quality of the students that attend. They also have a very hard working mail department, but they take it up a notch and offer to waive the application fee for a large number of students. Thus, many many more students end up applying to Tulane. Why not? It's free! Might as well see what happens. The application itself is incredibly easy, especially if just utilizing the common app. No additional work or money is required. Has this worked? Yes. Tulane has the single highest number of applications of any private school in the nation last year. However, their yield is abysmal - at something like 15%. So, they have to accept a lot more people than ultimate enroll. I don't personally think Duke should utilize this strategy.

• The other technique that Tulane uses, as well as USC in recent years, is the abundance of merit scholarships. In USC's case, this has significantly helped the quality of students it attracts (in addition to their great fundraising efforts). Tulane offers academic merit scholarships to a LOT of candidates, which obviously encourages students to attend. USC guarantees money for national merit scholars. Duke, on the other hand, gives nothing to national merit scholars. Duke has a fairly significant percentage of its student body that is national merit scholars, etc. so I don't think they should go ahead and give them all money. I think that Duke strikes a nice balance with offering about 60 highly competitive academic merit awards per year. This perhaps convinces a very small percentage of students to choose Duke over the likes of Harvard and Yale, but doesn't have a significant impact. However, these programs are great and the students accomplish a lot with the additional mentoring and research opportunities, leading to a very large number of AB Duke awarded prestigious nationwide scholarships like the Rhodes which ultimately boosts Duke's appeal, so these merit scholarships have undoubtedly improved Duke's stature of being a nurturing place for students where they can succeed at the highest level. I wouldn't like Duke to expand the scholarships too much a la USC. Duke already has much better financial aid so needy students should get a hefty discount anyways.

3.) Retention - this encompasses 20% of the total score, which is quite a large percentage. This is more than student selectivity (which includes test scores, top 10%, acceptance rate). Six-year graduation rate is 80% of it and the other 20% is freshman retention. Having students enjoy their Duke experience obviously should be a priority and luckily Duke has a low transfer rate. The 6-year grad rate is also high at 95%, but not as high as Harvard and Yale (both 98%). Still, Duke is near the top so there's not that much room for improvement. In addition, frankly I'd think a school could simply get the highest percentage as possible by making the curriculum extremely flexible and as easy as possible. Cal Tech, for example, has an extremely rigid and demanding schedule. They require ALL students to take several difficult math and science courses. I'm sure some can't handle it and end up not graduating, whereas they'd be fine at another institution that allows them to take more forgiving classes. Does this make Cal Tech a worse school? Not in my opinion, but I guess in the opinion of US News it does. If Cal Tech wants to improve its graduation rate, it could easily move to the Brown model (i.e. basically no requirements whatsoever and students can change a course to pass/fail right before their final exam if they think they're going to get a C or D.)

• Getting back to CalTech, and I don't want to put this in PPB land, but CalTech is basically the only school that I know of that doesn't really have preferential admission practices for underrepresented minorities, legacies, etc. They evaluate the candidates simply based on their accomplishments. This leads to a very high quality class from a SAT/GPA standpoint. However, the school is all Asian and white. There were three African American freshman at Caltech in 2006. There was only one black freshman there in 2004 and 2005. That's right - one in the entire class! Three years over the past 13 years, there have been zero black students entering. There certainly are more qualified African American candidates than those numbers suggest, but I'd imagine black students don't want to apply/attend CalTech since there are so few others. When there's only 6 black people on the entire campus, it's hard to feel at home. MIT, on the other hand, certainly makes efforts to increase diversity. Diversity is not taken into consideration by US News (although they do have a list, but it's not part of the ranking), but having a diverse class does encourage students of various backgrounds to apply in larger numbers and subsequently increases desirability and selectivity.

4.) Faculty Resources make up 20% of the ranking, which includes classes under 20 students, class over 50, faculty salary, % of faculty with terminal degrees, student/faculty ratio, and full-time faculty. Class size is 40% of this score (8% of total, or more than 5 times as important as acceptance rate) and Duke already caps a lot of classes at 18. I'm sure a lot of schools do this. Seminars are capped at 12, last I recall. Having several classes in the 60 student range hurts your numbers, whereas it'd be better if you simply had two 300+ person lectures instead of ten 60-person classes, according to US News. I don't think that that's actually the case that a school with fewer 300+ person lectures is better than a school with several more 60 person lectures, but I guess the size at that point doesn't matter all that much. (By the way, I find the numbers quite misleading. I'd say the majority of my classes were greater than 50 students as a BME major, while Duke has something like 70% of classes under 20. However, the large classes affect more people, so if you have one class with 15 and one with 500, you're at 50% classes under 20, but obviously the 500 person class affects 97% of the theoretical student body of 515, while the 15 person class is only applicable to 3%).

• UChicago has improved class size numbers in recent years, and along with Harvard, Yale, and Columbia, is one of four schools in the top 10 in this category that is also ranked in the top 10 overall. Duke could perhaps game this a bit more, but I think they already do that to a certain extent and it certainly requires more money. Faculty salary makes up the next largest portion at 35%. Duke had financial constraints and can't easily increase this. A school like Harvard which has basically unlimited resources cannot be competed with.

5.) Ok, this is really long already, so I'm going to get the one point that I think schools like Duke could easily "improve" on if their sole purpose is to improve USN&WR rank with minimal effort/change. And that is percentage of students in the top 10% of the high school class. It makes up 40% of the student selectivity score, and this is something UPenn (Morse's alma mater) basically sees as gospel. Their top 10% percentage is 98% last I recall. If you're not in the top 10%, you basically have no shot at UPenn.

• A school like Stanford, which is certainly more competitive admissions-wise than UPenn, has a lower percentage. How is that possible? Well, they take a more holistic approach to admissions. Duke rejects 60% of valedictorians each year. They could easily have 100% of the class from the top 10% if they wanted to. They have nearly 30,000 applicants. And you don't think 3,500 of them (which is how many students Duke accepts about), or 12% of the applicants, come from the top 10% of their high school class? Of course they do. Obviously, not all high schools rank however so the 12% figure isn't completely accurate. In addition, why would being at 15% in rank at a highly competitive magnet school be necessarily worse than an applicant in top 10% at a weak public school?

• In addition, Duke (and Stanford) are the only top 10 schools that have top notch athletics at the DI level. About 10% of Duke's student body is varsity athletes. That is a sizable portion and obviously athletes are given preferential admissions (which I'm cool with). Stanford's holistic approach to admissions and relying on "special qualities" as opposed to simply the numbers perhaps harms its overall score. It's continually ranked behind HYP, even though it's on their level in the eyes of most people. Stanford can make up for this, however, with their sky high (equivalent to HYP) peer assessment scores. Duke's PA, on the other hand, which is the single largest component, is a bit lower, so can't make it up as well.

• If Duke wanted to improve its US News score, I see this as the single easiest way to do it. Duke is certainly relatively high (90%+), but not UPenn high (98%). Simply have a cutoff of top 10% like UPenn does would improve our score...but then athletics could potentially suffer in some sports. It's possible the difference between a raw score of 92.4 and 92.5 was the difference between 5th and 10th in the ranking this year. (I obviously have no idea if Duke scored 92.4 or if one of the fifth place schools scored 92.5; it's simply for illustrative purposes that little changes can make a big difference due to rounding). I personally see this criteria as worthless for top schools as they could all be at 100% top 10% if they wanted to. It makes sense to differentiate perhaps for schools ranked lower, but for the top 25 schools, it's basically a decision by admissions to go simply by numbers or to take other qualities into more consideration that could compensate for perhaps a lower class rank.

If you finished reading this, congratulations! I'm impressed. Sources below:

http://www.usnews.com/education/arti...l-universities
http://www.usnews.com/education/best...-2012?PageNr=4
http://www.jbhe.com/preview/autumn06preview.html
http://www.usnews.com/education/best...allest-classes
http://www.usnews.com/education/best...d-weights-2012

[cspan37421]

I'm glad you're impressed! Thanks for the post.

a few brief comments:

- yield: I'm not sure many applicants pay any attention to this except at the extremes, where it might indicate safety schools or dream schools.

- What did you think of WashU, besides the fact that they come across as a stalker school or desperate?

- I do not understand why retention rates should be 13x more important as acceptance rates.

- faculty resources: I hope they adjust salaries for cost of living. Also, as far as class size, in my experience at Duke (math/psych/Judaic Studies), smaller classes were not better - more often, they were the worst among the few bad ones. But it had nothing to do with class size anyway. It had to do with the quality of instruction, which for me was best in medium sized classes.

- the top 10% is also a poorly chosen stat IMO, because Duke draws a decent % from selective and academically rigorous private schools. The numbers just aren't comparable across all schools unless you know that breakdown.

[Bluedog]

What did you think of WashU, besides the fact that they come across as a stalker school or desperate?

I think it's a great school actually. Just their admissions policies are a bit wacky, in my mind. I actually was at WashU this summer and it's a really beautiful campus. The cohesive architecture almost reminds me of Duke, although obviously it's a much different style. It seems like a great fit for somebody who perhaps might be interested in LACs, but also wants top notch research and a medical center nearby. It's in a nice area of St. Louis and Forest Park is convenient. People I know who have attended WashU have nothing but good things to say about it, and the campus food is supposed to be one of best in the nation. So, yeah, I think it's a great school once your there, but admissions certainly rubs me the wrong way...I ended up not applying after receiving 100+ pieces of mail.

faculty resources: I hope they adjust salaries for cost of living.

They do. "Faculty salary is ... adjusted for regional differences in the cost of living using indexes from the consulting firm Runzheimer International." I'm not sure how city-specific these regional differences are.

http://www.usnews.com/education/best...-2012?PageNr=3

Oh, and I thought of one more thing schools do to inflate their numbers. Excluding spring admits or admits on a different campus. UC-Berkeley, for example, has a fairly significant number of students who are only admitted starting in the spring term and their enrollment is effectively delayed a semester. These students can (according to US News rules) be excluded from the figures, reducing the school's acceptance rate, and most likely increasing the stats of the admitted class since spring admits are typically weaker. I'm sure Berkeley's rationale for this, however, is not to game the system but rather they have a need for more students in the spring. But if they're allowed to report numbers that look more favorable, why not? Middlebury has a bunch of spring admits too. There are others schools that do this as well.

Emory does the "hide weaker students from the numbers" game more effectively. They have a college in Oxford that has a 58% acceptance rate with an average GPA of 3.5 and an SAT range of 1160-1340 compared to the Atlanta campus which has an average 3.84 GPA and about a 1400 SAT. I'd understand having the numbers separate if the students actually finished at Oxford College, but students "graduate" from the Oxford College after four semesters and then become juniors at Emory College in Atlanta and become graduates of Emory two years later. Students admitted to Oxford College are not included in the US News statistics, even though all of them become Emory graduates. Oxford has about 900 students, so it's not an insignificant number and would certainly affect the numbers. Again, I doubt Emory's purpose of having this college is to improve its US News rank, but if it's allowed to exclude those students from their figures, it might as well.

Georgetown is penalized for having diplomats and those with great foreign policy experience teach courses simply because they don't have PhDs. I'd think somebody with real world experience in these areas would be a great person to learn from, but they don't have a terminal degree in their field so they must not be as good, right?

[sagegrouse]

• In addition, Duke (and Stanford) are the only top 10 schools that have top notch athletics at the DI level. About 10% of Duke's student body is varsity athletes. That is a sizable portion and obviously athletes are given preferential admissions (which I'm cool with). Stanford's holistic approach to admissions and relying on "special qualities" as opposed to simply the numbers perhaps harms its overall score. It's continually ranked behind HYP, even though it's on their level in the eyes of most people. Stanford can make up for this, however, with their sky high (equivalent to HYP) peer assessment scores. Duke's PA, on the other hand, which is the single largest component, is a bit lower, so can't make it up as well.

This is an incredibly informative post. Thank you. One quibble is the use of stats on number of athletes. The total scholarships permitted by the NCAA in the sports Duke offers seem to be 289. I think -- but can't find the reference -- that Duke gives a bit fewer. This is, at most, 72 per class (and remember that many footballers use up a five-year scholarship through red-shirting). The size of the freshman class is 1665. This figure is 4.3%. Moreover, I remember when the basketball team had three players who were #1 or #2 in their HS class: Hurley, T. Hill and Lang IIRC. Now I know that some athletes get accepted who are not on scholarship but receive an admission preference. But still, I think 10 percent is way too high.

And then there is the competition in the -- pardon -- Hypocritical Ivy League. They give huge preferences to athletes and give them need-based scholarships out of the general scholarship fund but do no reporting. So the gap between Duke and Stanford, on the one hand, and HYP or Penn or Cornell on the other is not so great.

I am prepared, however, to concede that premier schools with Division III teams like Wash U, Chicago, MIT, and CalTech are likely totally innocent of preference for athletes.

sagegrouse

[Bluedog]

This is an incredibly informative post. Thank you. One quibble is the use of stats on number of athletes. The total scholarships permitted by the NCAA in the sports Duke offers seem to be 289. I think -- but can't find the reference -- that Duke gives a bit fewer. This is, at most, 72 per class (and remember that many footballers use up a five-year scholarship through red-shirting). The size of the freshman class is 1665. This figure is 4.3%.

No problem, glad you found it informative. With the 10% figure, I meant to suggest varsity athletes in general and not necessarily varsity scholarship athletes. According to 2009 Department of Education data, the percent of varsity athletes in Duke's student body is 10.1%. I'd imagine not all of the non-scholarship varsity athletes are recruited, though, but I can't find those percentages. Here's a link: http://www.forbes.com/colleges/duke-university/

Athletic Divison
NCAA Division I-A
Percent of Varsity Athletes in Student Body
10.1%

Sports numbers from 2009 Department of Education data

And I didn't mean to suggest Duke's athletes are not great students as well. In fact, many of the athletes are Duke are exceptional students and I was incredibly impressed with many by their amazing time management skills. I couldn't imagine having the time commitment they do and still being able to stay on top of my coursework. However, from a stats perspective, obviously Duke coaches have some sway over the admissions with Coach K and Cutcliffe (and Coach P) obviously having the most say, but the others coaches can certainly push the acceptance of a possibly lesser academic candidate if they've exhibited extreme athletic abilities and success. And with a sizable percentage of athletes (albeit the number is smaller for those that actually receive scholarships), it can have an influence on the numbers. (I'm not opposed to this mind you as I think it leads to a better experience for all Duke students, but US News doesn't consider sports team support in their rankings.)

And, yeah, definitely the Ivy League gives preferential admissions to athletes, but they still have an index that the team average so to be above - and it's higher than that at Duke. If you are a "normal" student with a 3.9 GPA and a 1450 on the SAT Math + Verbal, your chance of acceptance at Harvard is still very small (probably like 10%), but if you're a coveted varsity athlete, your acceptance rate with those numbers shoots up a lot. And with their great financial aid, they'll likely get money too even though it's technically not an athletic scholarship. Either way, both students are qualified and can do Harvard coursework, but the supreme athlete has an extra special skill that makes them more desirable. I'm sure a lot of Harvard athletes fall into that category. Same with Duke, but Duke's athletes are even more cream of the crop from an athletic standpoint, so it's difficult to find individuals that are tops in both so sometimes you dig a little deeper on the academic front to field as quality of a sports team as possible. But I agree with the general theme in your post.

[duke79]

I'm glad you're impressed! Thanks for the post.

a few brief comments:

- yield: I'm not sure many applicants pay any attention to this except at the extremes, where it might indicate safety schools or dream schools.

- What did you think of WashU, besides the fact that they come across as a stalker school or desperate?

- I do not understand why retention rates should be 13x more important as acceptance rates.

- faculty resources: I hope they adjust salaries for cost of living. Also, as far as class size, in my experience at Duke (math/psych/Judaic Studies), smaller classes were not better - more often, they were the worst among the few bad ones. But it had nothing to do with class size anyway. It had to do with the quality of instruction, which for me was best in medium sized classes.

- the top 10% is also a poorly chosen stat IMO, because Duke draws a decent % from selective and academically rigorous private schools. The numbers just aren't comparable across all schools unless you know that breakdown.

I agree with you on the "10%" stat issue. Using high school rank as a factor in assessing quality of incoming student body to a college assumes ALL high schools are of a similar quality. I think most people would recognize the fallacy in this assumption, especially for schools like Duke or the Ivies which draw a substantial percentage of their applications from students who attend either academically rigorous private or parochial schools or competitive public high schools. A student who is at the top 25% level at Phillips Andover Academy might easily be the top 10% of the class at Podunk Regional high school. I suppose this is where the SAT scores are supposed to equalize applicants.

[cspan37421]

A student who is at the top 25% level at Phillips Andover Academy might easily be the top 10% of the class at Podunk Regional high school.

I actually went to Podunk Regional; the 25% PAA types would be in the top 1-2% at Podunk. But watch out, once at Duke, some rested on their laurels and let Podunk kids outhustle them!

[cspan37421]

I'm slow to thank you for your super-informative, thoughtful post. You've obviously got a good understanding of the way admissions departments make sausage.

I always tell my DD that the only ranking that should matter is her own.

For me, when I applied, geography was the main factor, so I applied to Duke and UNC-CH, and would have to WFU if I hadn't already received the big packet from Duke ED. Looking back, it's not the most comprehensive deliberation I've ever undertaken, but it worked out OK.

[Bluedog]

A student who is at the top 25% level at Phillips Andover Academy might easily be the top 10% of the class at Podunk Regional high school. I suppose this is where the SAT scores are supposed to equalize applicants.

I actually went to Podunk Regional; the 25% PAA types would be in the top 1-2% at Podunk. But watch out, once at Duke, some rested on their laurels and let Podunk kids outhustle them!

Andover and Exeter (as well as many other highly competitive private boarding schools) don't report class rank for this reason. They're smart enough to know that top colleges nowadays want students that are ranked near the top of their class and probably wouldn't take as many students if class rank was reported. The top colleges are also smart enough to know that somebody who is in the top 30% at Exeter is still a very high quality applicant, but may be more reluctant to admit that individual if it affects their numbers. So, eliminating the numbers all together is a relatively easy solution to get around it.

This is a trend among high schools. My highly competitive public high school reported only decile rankings when I went (i.e. gave only the range to the nearest 10%, so if you were fourth decile, you knew you were 30-40% top of the class). However, they have since gotten rid of rank all together as the state school wouldn't take people below a certain percentage even though they were high quality students from a test score/GPA perspective. This trend is only going to increase, and I'd think US News at some point may have to re-examine their approach when such a large percentage of high schoolers don't receive any class rank. But perhaps even if it's only 50% receiving a rank, that's still a large enough sample to be representative. I'm not sure...

[duke79]

Andover and Exeter (as well as many other highly competitive private boarding schools) don't report class rank for this reason. They're smart enough to know that top colleges nowadays want students that are ranked near the top of their class and probably wouldn't take as many students if class rank was reported. The top colleges are also smart enough to know that somebody who is in the top 30% at Exeter is still a very high quality applicant, but may be more reluctant to admit that individual if it affects their numbers. So, eliminating the numbers all together is a relatively easy solution to get around it.

This is a trend among high schools. My highly competitive public high school reported only decile rankings when I went (i.e. gave only the range to the nearest 10%, so if you were fourth decile, you knew you were 30-40% top of the class). However, they have since gotten rid of rank all together as the state school wouldn't take people below a certain percentage even though they were high quality students from a test score/GPA perspective. This trend is only going to increase, and I'd think US News at some point may have to re-examine their approach when such a large percentage of high schoolers don't receive any class rank. But perhaps even if it's only 50% receiving a rank, that's still a large enough sample to be representative. I'm not sure...

Yea, I attended a small private high school back in the mid-1970's and, at some point during my high school career, the school did away with class rank (or at least what they were reporting to colleges) for the very reason that they thought that reporting class rank was hurting the college admission chances for some students. I believe that they would just give broad percentiles of where a grade point average fell for the entire grade (which was only about 50 students at the time). I'm not sure that this really helped the students, though, because I'm sure many college admission offices had a hard time figuring out exactly how an applicant had done at the school relative to his or her peers and I'm guessing that some colleges just declined the student as a matter of course. I'm sure that because of rankings like US News and World Report and others, there is a lot of pressure on the college admissions office at any highly selective colleges to accept as many people from the top percentile of their high school class, no matter how good or bad the quality of the high school. Frankly, it is probably a consideration that the parents of a smart middle school child should consider before spending $30,000 to $50,000 per year to send their child to a top private high school. It may NOT actually help your child get into that top college.

[duke79]

I actually went to Podunk Regional; the 25% PAA types would be in the top 1-2% at Podunk. But watch out, once at Duke, some rested on their laurels and let Podunk kids outhustle them!

Absolutely true. Any advantage a student has in attending a top private high school is probably dissipated after a year or so of college. The same can be said of grad school as well. When I attended law school at Duke, many of the students who did the best came from second, third, and fourth tier colleges and universities but just worked harder than the Ivy and little Ivy and Duke and Stanford grads.

[duke09hms]

Just saying, you're only hurting yourself if you don't try to help your ranking.

Really? How?

Because as you drop in the rankings, the best students will no longer choose to accept your offer of admission over other top schools. Trust me, the vast majority of high school students are very much still maturing and rankings unfortunately matter A LOT to them. I sit on the Duke admissions committee, and as a young alumni who hears the real talk from applicants and accepted students, rankings are a huge part of their decision. Students who are international or places where Duke isn't as well known like the West Coast make their decisions almost entirely based on rankings.

If it were 2005 again, and Duke was ranked outside of the top 10 instead of at #5, would I have chosen Duke over acceptances from harvard, yale, princeton, and MIT? To be perfectly honest, no I wouldnt have, and I probably would have had Northwestern on my list before Duke in that case (Chicago > Durham) - and this is coming from someone who now loves and bleeds Duke blue.

I'm sure there are also downstream effects affecting alumni and corporate donations, overall reputation. People want to give to a winner, and what other way do we have to "keep score?"

Fact is, the rankings matter, and if you can play the game to win with minimal or nonexistent sacrifices (like the dishonest/desperate methods described above employed by upenn/washu/columbia), you should. You are only hurting yourself.

[cspan37421]

Duke09hms:

Sorry to have to repeat myself, but again: can you provide a source for the specific claims you made about WashU?

You wrote, "WashU sends info mailings to thousands of HS students and counts every returned request to join the info mailing list as a complete application, thus making them appear more "competitive" and increasing their ranking. This is a dishonest method, and Duke shouldn't do this."

I'd like to know how you really know this.

When you say you sit on the Duke admissions committee, are you a volunteer interviewer on the AAAC or do you work in the admissions office as part of Christoph's team? I am a local chair of the AAAC and thus am particularly interested in your opinions on this subject and what info you came across in forming them.

Allow me to make a counter argument to the notion that playing the USN&WR game helps us get the "better students." I would consider a student who knows not to put much stock in USN&WR college rankings as a "better student" - they have some critical thinking, and that's all too rare no matter what your age.

[cspan37421]

Because as you drop in the rankings, the best students will no longer choose to accept your offer of admission over other top schools.

By the way, a bit of indirect evidence against this happening is this: We all know that WashU has been playing the USN&WR game for some time. They are currently ranked #14 according to the USN&WR website. I think they've been top 20 for some time. Anyway, in the "revealed preference" paper by Avery, Glickman, Hoxby, and Mettrick, they rank 65th. The paper was published in 2005; I don't think WashU's place in the USN&WR survey has changed much since then.

That means that if a student was accepted to WashU and any of the 64 schools ranked above it, an applicant was statistically more likely to choose the other top-64 school over WashU. The greater likelihood ranges from obviously very high (Harvard vs. WashU) to only slightly greater (Emory vs. WashU). So playing the game does not appear to have actually helped them in the way you imply.

I concede it leaves open the question of whether not playing the game hurts you, but it seems reasonable to infer this disconnect works both ways.

I'll leave it to others with more knowledge of what other colleges and universities do in terms of playing the USN&WR game or not playing it. If anyone wishes to post a list here of schools known to have fallen in the USN&WR rankings for not playing their game, I'll look them up in the Revealed Preferences report and perhaps we can infer whether it actually hurt them or not. I'll just close this with a great quote from Reed's college president, "by far the most important consequence of sitting out the rankings game, however, is the freedom to pursue our own educational philosophy, not that of some newsmagazine." [wikipedia]

Whatever other pros and cons there are to Reed College, I give them a +1 for this stance; I wish more schools had the self-confidence and ethical backbone to do the same.

[sagegrouse]

If it were 2005 again, and Duke was ranked outside of the top 10 instead of at #5, would I have chosen Duke over acceptances from harvard, yale, princeton, and MIT? To be perfectly honest, no I wouldnt have, and I probably would have had Northwestern on my list before Duke in that case (Chicago > Durham) - and this is coming from someone who now loves and bleeds Duke blue.

...............

Fact is, the rankings matter, and if you can play the game to win with minimal or nonexistent sacrifices (like the dishonest/desperate methods described above employed by upenn/washu/columbia), you should. You are only hurting yourself.

Well, at least your reckless choice of Duke apparently didn't hurt your chances of getting into Harvard Med.

sagegrouse