The most selective colleges have a higher density of strong students and test takers (selected for in undergraduate admissions), so it is no surprise that they have lots of students who can earn high GPAs and LSAT scores.
Attendance at Yale undergraduate mostly says that the student was a strong student in high school, perhaps with some exceptional characteristic (which may have been unearned by the student, like relation to alumni, donor, VIP, etc.). Attendance at Yale law school mostly says that the student was a strong student as an undergraduate.
Of course they did. As any AP Stat student knows, correlation does not equal causation.
(Hint: “elite schools” select out poor test takers, by definition. In other words, they only accept excellent test takers. And someone who can ace the SAT/ACT can also ace the LSAT, a very learnable test. It’s just that most kids at Podunk State are not excellent test takers, so they could never pull the 173 mean for HLS.)
And yes, there is no doubt that academia, like Wall Street, is one place where pedigree matters for jobs. But I would submit most high schoolers who pining for HYPS have little/no intention of entering academia.
You can view more specific stats of Harvard Law School admits and rejects at http://harvard.lawschoolnumbers.com/stats/1415 . Note that the green admits closely form a line based on LSAT and GPA. The overwhelming majority of applicants who are above the LSAT + GPA admit threshold line are accepted, regardless of whether they attended “Podunk State” for undergrad or HY for undergrad. And the anomalies who are accepted with weaker stats are almost without exception self identified URMs, rather than students who attended HY for undergrad…
Agreed, the formulas are not so complex they are hard to understand.
Nevertheless, to move the ranking needle very far in your favor, a college needs to make changes - probably multiple changes - that arguably amount to more than “gaming”. Suppose a college starts a mass marketing campaign to increase the number of applications, reduces average class sizes from 25 to 19, and raises faculty salaries 10%. Its USNWR ranking very likely will improve. But I think it’s debatable whether these changes only amount to “gaming”, or if they are in fact substantive improvements driven by market competition.
There may well be other changes with bigger impact on academic quality, yet lower impact on rankings.
A college could resist grade inflation, assign heavier reading loads, assign more writing assignments, and create new incentives (or training programs) to improve undergraduate teaching. Those changes might improve academic quality significantly, yet not move the USNWR ranking needle at all (not for a long time, anyway). A few colleges (Reed maybe, UChicago formerly) have had reputations for real academic excellence without necessarily being rewarded with commensurate prestige or T10 rankings.
I’d prefer to see substantive improvements in student-faculty (or student-student) engagement, such as the NSSE surveys measure. Absent that, I’ll settle for smaller classes, higher faculty salaries, and improvements against some of the other USNWR metrics, as the next best widely available proxies for academic quality. I think on balance we are better off for having those measurements (as long as we understand their limitations). They can be used to expose a data-driven list of 50, 75, or 100 colleges that appear to have similar features to the most famous/prestigious/selective schools, which are out of reach even for many very good HS students. That’s no substitute for serious research, but probably better than depending on lunch room gossip to build a good reach-match-safety list.
Gaming the class size can be done without reducing the average class sizes. For example, if the college has 100 24-student classes, but changes them to be 50 29-student classes and 50 19-student classes, the average class size is still 24, but the distribution looks better for the USNWR rankings, since half of the classes are now in the next better range (1-19), while the other half are in the same range (20-29). Indeed, compared to the previous situation, 1450 of the students are now in a large class (29 instead of 24), but only 950 are in a smaller class (19 instead of 24).
^ I think that would be a fairly trivial change and would have little significant and lasting impact on the rankings. It is one thing to suggest gaming moves that could be made, another thing to demonstrate that they are in fact being made often and with results that give some colleges (and not others) a significant lasting advantage.
At Columbia, almost 83% of classes have fewer than 20 students (per US News, that is). At UMichigan, less than 60% do; at Berkeley, it’s about 52%. Private universities ranked in the T10 have much higher levels of instructional spending per student than Public Ivies ranked in the 20s and 30s. That difference in resource allocation makes a far bigger difference … I think … than gaming the class size numbers.
But maybe I’m wrong. Or maybe I’m not wrong, but some of us don’t think the class size differences matter much and should not be an important ranking factor. What are some better metrics, ones USNWR does not use and that are more resistant to gaming, that would show very different results from the USNWR/Kiplinger/Forbes T20? Or, are some of us just opposed to the very idea of performance measurements?
If you look at Northeastern’s class schedule on line, you will see that 19 is an extremely common class size cap, but 17, 18, 20, and 21 are much less common.
How likely is that if class sizes were determined without regard for ranking criteria?
“The student-faculty ratio at University of California–Berkeley is 18:1, and the school has 52.3 percent of its classes with fewer than 20 students.”
“The student-faculty ratio at Stanford is 4:1, and the school has 70.1 percent of its classes with fewer than 20 students.”
Does anyone really believe these numbers? If you do, I urge you to come to the bay area and sit in on classes at these fine institutions and see how many of them have less than 20 and how many of them have less than 4 or 18 as implied by the numbers. I know that s/f ratio is not the same as class size but that’s what USNWR is saying when they look at these two numbers for the rankings. These numbers are gamed, no doubt it.
^ You can also work the top end of the class size game. Let’s say you find you have 30 classes that regularly enroll 50-60 students. You get no credit for those in the US News formula, but they do count in the denominator used to calculate fractional shares of classes in each of the ranges for which you do get credit. So the large-ish classes actually hurt you in the ranking formula. So you cap enrollment in each of those classes at 49. Same total number of classes, but now you’re getting some credit for classes in the 40-49 range. And you can do the same thing at each of the break points.
Does that make things better for students? There’s little functional difference between a class with 49 students and a class with 55 students—except that some students may be closed out of a class they want.
Or you can take a meat-axe approach and simply eliminate some popular classes that regularly draw 50+ students. Or combine two 55-student sections of the same or similar classes into one 110-student class. These changes alone won’t move the needle very much, but in combination with hard caps of 19 on all classes currently in the 20-29 range, and similar changes on up the scale, they can make a difference. Students might end up with less choice, but US News doesn’t punish you for that.
The numbers are believable, but they do not mean what many readers think they mean.
The average class size students will encounter on average is larger, because more students are in large classes. There are lots of small classes, but they are small because few students choose them. As many or more students may enroll in a large introduction to computer science course as enroll in 50 or 100 small classes, for example.
@privatebanker Nixon grew up in Whittier, CA, graduated from Whittier High School, and then attended Whittier College. Other members of his immediate family suffered from serious illness, and he stayed in his hometown for college so that he would be available to help out. Otherwise he probably would have gone somewhere other than Whittier College. He ultimately went to Duke for law school, and one of his younger brothers later attended Duke as well.
I was a Classics major. There was one- exactly one- course at my university in the Classics department which had an enrollment of over 200. A highly popular survey literature course which compared the great literary works of Classical Greece and Rome to both the English Classic novels (think Tom Jones) and contemporary literature. You did not need ANY Latin or Greek- all works were read in English.
Every other class required language study or proficiency, and even the Freshman intro classes had 10-12 students.
it is really hard to pack a large lecture hall with a seminar on Ovid, read in the original, or a course on the foundations of democracy where proficiency in Greek is a requirement! And that holds whether you are at a tiny LAC or gigantic university!
I have 3 degrees from Stanford and have taken many Stanford classes in a wide variety of fields. I do believe the numbers. However, the class size is highly variable, particularly dependent on class level and major. For example, any popular intro math/science class, such as most pre-med requirements, will be quite large… often with hundreds of students. However, if you take upper level major specific classes in a less popular major, practically every class will have less than 20, and many will have less than 4. The larger classes generally break up in to smaller sections of less than 20, which are run by TAs.
Math 51 (common starting point): 300+ students
Math 61DM (advanced proof based starting point): 22 students
Math 116 (complex analysis): 14 students
Math 272 (topics in partial differential equations): 4 students
SLAVLANG 1 (first year Russian ): 5 students
SLAVLANG 5 (Russian for native speakers): 0 students
SLAVLANG 181 (fifth year Russian): 2 students
“Knowledgeable people understand that getting into the top grads schools is much more likely from H/S/P.”
Top law school admissions has a LOT to do with having high LSAT scores. Getting into HYPS undergrad has a LOT to do with having high SAT/ACT scores. Since HYPS undergrad are brimming with good standardized test takers, you would of course expect them to put a lot of kids into top law schools.
But that has zero to do with actually attending HYPS for undergrad. The kind of kid who can get into Harvard College will tend to be a good applicant for Harvard Law School. Regardless of whether the kid actually attend Harvard College or not. But most kids who get into HC end up enrolling at HC. Correlation not causation.
FYI, most recent HLS class had 562 kids coming from 182 different colleges. My T14 law class had a decent number of HYPS kids, but also plenty of kids (like me) whose undergrad college was meh or worse.
“If you look at Northeastern’s class schedule on line, you will see that 19 is an extremely common class size cap, but 17, 18, 20, and 21 are much less common.”
^ So it would take 16+ classes of 0-19 students to equal the number of students in just that one class of 300+. More, if the average size of the small classes is lower than 19, or if the number in the large class significantly exceeds 300, or both.
In the examples you gave in post #112, the average enrollment in the small classes is 5. At that rate it would take 60 small classes to equal the enrollment in just one class of 300.
US News doesn’t seem to grasp this point, but it’s really the number of large classes that determines how much time students are spending in large classes—not so much the number of small classes, which is misleading. Many people are fooled by this. They see that figure that says 70% of the classes have less than 20 students and they think, “Oh, that’s great, I’ll be spending 70% of my time in small classes.” Nope. It doesn’t work that way. At Stanford and almost all other major research universities, undergrads are, on average, spending as much or more of their time in large classes than in small ones, despite the gaudy percentage of small classes. Do the math yourself if you don’t believe me. The large classes wouldn’t be large unless an awful lot of students are enrolled in them.
Not so at the better LACs. At my daughter’s LAC there were no classes with more than 50 students, so no one spent any time in large classes as defined by US News,
I also think US News does students a disservice by lumping all classes of 50+ together, regardless of whether the class enrolls 50 or 500 (or 1300 as at Cornell where and intro Psych class reportedly enrolls that many). It’s a qualitatively different experience, In a class of 50 I can call on each student at least 2 or 3 times over the course of a semester and get to know every student at least by name and face, or better if they come to my office hours or join in non-mandatory social hours. Not so in a class of 500. Not to say all mega-classes are bad; some are quite popular. But a class of 50 is not terribly different from a class with 45 or even 40. A class of 500 is a different kind of experience entirely.
Here is an examples of what I was talking about. The numbers are far more meaningful if you do the math based on class sizes. As always H/S/Y dominate.
The Hersch paper is the one that predefines tiers of prestige that do not necessarily make sense:
Tier 1: includes Harvard and Yale, but also Saint Louis and Syracuse
Tier 2: includes Amherst and WIlliams, but also Westminster ¶, Puget Sound, and others
Tier 3: includes UCB, Michigan, UNC-CH, UVA, but also Wayne State, UAB, and others
Tier 4: all others, including Harvey Mudd that is not listed in the other tiers
say: I hope that you did notice that 5 of the top 10 schools in teh USNews listing of colleges for members of Congress were (gasp!) publics and one college has a heavy religious influence (BYU)? (
H/S/Y are overrperesented in some desirable fields (not all) for a variety of reasons. There also various desirable fields that display other patterns. For examples, the Hsu and Chronicle articles you linked focus on a paper about hiring at “elite” investment banking and consulting firms. While hiring at elite investment baking and consulting firms can focus on college prestige, they are the exception, not the rule. Hiring in the overwhelming majority of other industries is completely different and places a tremendously lesser degree of emphasis in prestige. In the employment survey I linked earlier, college employers as a whole indicated college reputation was the least influential factor in their hiring decisions for new grads. Consistent with this, if you look tech companies that college graduates as a whole seem to think are especially desirable (Google, Apple, Facebook, …), Harvard and Yale are not among the colleges with the most hires. The number of these especially desirable tech employers attending H/Y career fairs seems lower than numerous other colleges, including quite a few publics.
Or looking at members of congress, the link you listed mentions that a little under 5% of members of congress attended H/Y. That is an overrepresentation, but >95% did not attend H/Y; and much of that 95+% attended publics. If >95% did not attend H/Y, it certainly is not a requirement, but there is some degree of correlation. Correlation is not the same as causation. I doubt that the primary reason those <5% were elected was because they attended H/Y. Instead I expect the reasons for they are correlated because the two groups, H/Y graduates and members of congress, are more likely to share various common traits – good student, smart, leader, HS president, … Coming from a wealthy family can also be a big asset in politics, and wealthy families are overrpersented at H/Y, some of which have special admissions advantages, probably more so back when most members of congress were applying. There is also a selection effect. For example, was Bill Gates primarily successful because people were wowed by Harvard’s prestige, or was it more that Harvard admitted a really special kid with unique traits who was likely to go to do amazing things, regardless of how others felt about him having attending Harvard for 2 years.