somedumbnoob

You can't judge "best placement" in these fields unless you take the differing student bodies into account. For example, Georgetown's student body is generally known to be far more preprofessional than Uchicago's. Therefore, Gtown doing better than Chicago in the number of students it sends to top law/med/mba schools is not unsurprising. But that doesn't necessarily mean that a Georgetown degree is better than a Uchicago degree as the term "best placement" implies because fewer Chicago kids aimed for those professional schools in the first place.

Similarly, Reed college is ranked at the top for sending kids to phd programs but doesn't surface in your data... does that mean that they can't send kids to law/business/med school or does it mean that the students there simply aren't interested in going to professional schools?

It would be better to compare how popular preprofessional programs are at these colleges before comparing them. But I wouldn't know the means of obtaining that data.

davida1

tealover, Caltech does not have the strongest placement numbers at professional schools.

Here is the data for Caltech:

Percentage: 2.2%
Number: 5

davida1

abl doesn't understand the fact that a study can be imperfect and still indicative and insightful. WSJ's study was also imperfect, but it's not like this data is easy to come by. Schools tend to not publicize this information, but I think that my study gets at information that normally would not be available to high school seniors.

abl also doesn't understand the fact that the numbers would not change in any meaningful way with the adjustments that he/she is pushing for. I see no evidence of your Yale/Cornell numbers being accurate.

hawkette

I think that DunninLA got it right early on in this thread with his conclusion that the data presented is pretty irrelevant and about as insightful as saying that the colleges with the highest selectivity will likely have a higher number going on to the highest ranked Law/Med/MBA school. No surprise there as it is quality and quality out. The undergraduate college probably has very little to do with it.

I should also like to introduce two ideas:

1. The data for MBA schools should be taken with a HUGE grain of salt. Students go back to business school after an average of 3-7 years in the workforce. At the top of every application is going to be the experience and the achievement in the real world. MBA adcomms are not looking at your grade in Film Studies from sophomore year and deciding if you have what it takes to join their incoming class. You need a good GPA and/or a good GMAT to show that you have some academic ability, but what really counts in the MBA admissions process is how you perform in the post-undergraduate world. This often has next to nothing to do with the undergraduate college you attended.

This delayed matriculation to graduate business schools is also increasing at the Law Schools and even the Med Schools. While certainly far less than the grad business schools, the law and med schools seem to be realizing that the applied intellectual and maturity level of its applicants is far better understood and evaluated after some time out of the college environment. I would expect (hope!) that this trend will continue.

2. There is a cultural effect for graduate school attendance. Some colleges are just wired for this and their culture “expects” many of their students to go to graduate schools (not just of law/med/business, but across many disciplines). This campus predisposition leads to outsized numbers coming from certain colleges and not from others where such a culture is less dominant or not existent.

I would not automatically conclude from this that ABC College is better than XYZ University at grad school placement. To make a true determination, you need to understand who is applying and seeing admissions data from each college. The data that WSJ and Facebook provide may be some indication of which colleges are most actively represented at these grad schools, but they very well may not reflect the quality and number of applicants from a given undergraduate college. Matriculation patterns?-I’d accept that. Placement strength?-I think that is a poorly-informed guess.

davida1

Actually if there are certain schools that have particularly strong placement at MBA programs that would suggest (according to your line of reasoning) that they are producing alumni that are successful in the workplace because they are earning spots at top b-schools (and great work experience is more important than academic factors). Your conclusion doesn't follow.

To pretend like there is no significant relationship though between "matriculation patterns" and "placement strength" is laughable.

And schools with high selectivity do not always perform well (that is sort of the point of the WSJ study and my study, actually). US News selectivity rating, as a proxy for selectivity, does not equal strong professional school placement. WUSTL's selectivity ranking is very high relative to its performance as is the case with other schools like: Caltech, Emory, U Notre Dame, Tufts, Vanderbilt, Harvey Mudd, Haverford, Claremont McKenna, Bowdoin, Davidson, W&L, Carleton, Hamilton, and Carnegie Mellon U.

To lump all of those under-performing schools and use the excuse that they are just interested in pursuing other routes needs evidence to support it not just an assumption. You shouldn't just assume self-selection without some evidence to suggest that that is occurring. It seems more reasonable to consider these schools not doing as great of a job in getting students admitted to top professional schools.

There are places where people post their stats and their admissions decisions to professional schools (I won't point them out) but a quick screening shows that schools of equal selectivity do not in many cases give graduates the same chances of admission to top JD/MBA/MD programs given that they have the same stats.

hawkette

davida1,
I think your perspective is naïve, biased and not representative of what grad school admissions data mean. But maybe I’m reading you wrong and I don’t’ fully understand your perspectives. Would you please explain a few things?

First, what do you believe is the connection between one’s college’s brand and one’s application to “top” MBA schools? How do you interpret this as a reflection of the differentiated quality of the undergraduate education from that college?

Second, do you know the numbers of applicants from each of the schools, highly ranked or not, and their admissions decisions and if there is or is not a connection between campus interest, matriculation patterns, and placement strength? Do the conclusions differ as applied to law schools, to med schools, to business schools?

Third, can you comment on how you see law school students positioning themselves for law careers in various regions of the country, including in major states like Texas and how that is factored into your rankings?

1of42

I think it's also interesting how well Princeton does. Yes, the school is prestigious, but we have no professional schools, and many of our grads go straight into finance, making it even more impressive.

tomslawsky

One thing the study doesn't and possibly can't correct for is institutional bias. For example, I go to U-Florida (MBA) and I know there are many people at this school who have the stats to go to top 15 professional schools but due a UF bias, they decide to stay in Gainesville for their entire academic career. Maybe they are going to med school and theit parents are alumni? For example, look at the MBA GMAT scores for UF- they encroach on top 15 territory. In fact, I'm not so sure that if a UF professional school grad intend to stay in the state, the value added of going to a school in the top 15 (especially at the lower end of the 15) would make a difference 5 or 10 years out of school.

ejr1

I am not a statistics-oriented person. I want to know facts, and what I see here are lots of questions which make me question the validity of the numbers. Such as noted above, nothing has been done about taking into account the type of student body as in U v. LAC, etc. What I would really like to know if the percentages of students from each school who applied to these programs and were accepted. That would tell me a lot more than these. If school A had only 15 apply, and all 15 were accepted, that would mean more to me than school B had 500 apply, and 25 made it. B would beat out A in these numbers, but A would beat out B in my book.

abl

"abl doesn't understand the fact that a study can be imperfect and still indicative and insightful. WSJ's study was also imperfect, but it's not like this data is easy to come by."

Yes, I would say that the WSJ's study is a perfect example of a study that is imperfect while being indicative and insightful. Yours, however, is far to imperfect to be indicative or insightful. A somewhat parallel situation would be if you ranked undergraduate college's selectivity based on how many students were in a "I got a 1600 on the SAT" facebook group. That so-called "study" of selectivity would obviously not be indicative or insightful of anything save perhaps the propensity for students to brag at a particular school.

"abl also doesn't understand the fact that the numbers would not change in any meaningful way with the adjustments that he/she is pushing for. I see no evidence of your Yale/Cornell numbers being accurate."

You're almost right. You don't understand whether the numbers would change in a meaningful way or not. It may be a fact that the numbers would change, and it might be a fact that they would not. You don't know. I don't know. This is why you need to adjust for your selection bias. Whether you end up with similar data or very different data, your data is meaningless before making that adjustment. Go back and read through your stats book--it should help make things more clear.

This "study" of yours wouldn't pass at any of the colleges on your list. I doubt it would pass at pretty much any college in the country, for that matter; it sure as heck wouldn't pass at the high school that I am currently teaching at (which is by no means an especially good high school). The fact is, your ranking is shoddy work. I don't know if you're a high school student, a college student, or even a college graduate, but you must understand that it's important to at least aim for accuracy, especially in a forum like these boards where high schoolers may be easily misled.

Your problem isn't so much that you've compiled a list of what you believe the top schools in the country are. The problem is that you have attached a modicum of science to what is not scientific. In fact, if I were to compile a list or ranking of what I believe the top 20 schools in the country were, it would look pretty similar to your per-capita adjusted list. However, I would not--as you have done--mislead anyone about the scientific validity of that list.

Look, I know you've invested a lot of time into this ranking. As I've said before, you can still salvage your results, get something useful out of this. Don't be stubborn--make the adjustments!! If it's a time issue, I am sure you can recruit other posters here to help you. If it's not, well, what is it?

bluedevilmike

Tag .

afan

Davida1,

You are discovering that anything posted that purports to be a ranking will be flamed by those whose favorite colleges are not at the top.

Of course, methodological issues aside, a PERFECT run at this question would yield the same complaints. If you did have complete data on the undergraduate institutions of every student at all top 15 professional schools, some colleges would have higher percent representation than others. It is silly to claim otherwise.

The more interesting question is what to do with such perfect data?

The answer would indicate how much, and what kind of, attention to pay to this incomplete attempt at estimating what the complete data set would show. That is, for a given student does (s)he improve chances of attending a top professional school by going to College A vs College B?

I am firmly with those who would say "No. The differences to be found in the perfect data set would reflect the ability of the students who enroll in each college (the strongest effect), their goals and orientation (explaining most of the rest), and to a small extent socialization towards or away from professional school that occurs during the college experience."

If these are the factors that dominate the professional school enrollment of students, then ranking the colleges becomes more a description of where their students end up, rather than an indication of relative quality.

Look, the vast majority of successful people in the US did not go to professional school at all, and only a small fraction of, say, doctors, attended a top 15 medical school. To "grade" colleges by how many of their graduates did this is silly.

Very few Caltech students go to law school. Does that make it a bad college? A much larger proportion of Caltech than Harvard students get PhD's. Does that make it a much better college than Harvard? Of course neither is true. Both are great places to get an undergrad education FOR SOME STUDENTS. A description of what are typical career paths for Caltech vs Harvard students could be very useful for high school students considering college. A ranking that claimed "going to professional school is good, but going to grad school, or entering the workforce, are bad" is useless.

Here's one. Do a ranking of the proportion of college grads who are military officers two years out. The service academies will dominate the top of the list. Does this mean students should forget about HYPSMC and all apply to West Point? Well, those who want careers in the Army should certainly consider West Point. Those who do not want to join the military probably should stay away.

You cannot apply a single criterion to rank the appropriateness of a set of colleges for everyone. Different people have different goals, so they SHOULD attend colleges that emphasize, and enroll students who are interested in, different things.

abl

afan-

I think a perfect iteration of this ranking would be helpful. While Caltech may not do well on these sorts of rankings, I doubt that will discourage any student from applying--students interested in Caltech understand that it's somewhat of a math/science/engineering/CS specialty school. However, a plurality of students at most schools are (at least upon entering college--these things obviously do change) interested in attending Med/Law/Business school. Consequently, it is of interest how many students from different undergraduate schools end up at these top pre-professional schools. Obviously different schools are better at different things; for a student looking for a Philosophy PhD or an Engineering Masters, these rankings don't yield much interesting.

However, I believe that when combined with a selectivity ranking, these sorts of rankings can yield us all sorts of really interesting information. If College B is the 18th most selective in the country, but has the 7th most students attending top Med/Law/Business schools...well, that does tell us something about the school. Maybe students at that school are more Med/Law/Business oriented than average, or possibly the school has a significant added value. We can make pretty educated guesses about which of these conclusions is most likely to be valid based on the populations of the schools--say, if Amherst, Williams, Dartmouth, and Middlebury significantly over-performed or under-performed compared to each other, we could probably assume that the difference was due to some sort of added value in that school, given the similarity of these schools' student bodies. I believe the student bodies at most of the top schools (most of the Ivies, most of the Nescacs, Swarthmore, Duke, Stanford...etc,etc) are similar enough that over-performance or under-performance could reflect some sort of added value on the part of the school.

aquafoam

Yay wellesley

afan

As I reported in the other thread, a simple correlation of median SAT vs % of student body at top 15 professional schools yields a r of 0.78. In other words, 61% of the variance among these 20 undergraduate institutions is explained by this one number.

I do not believe it is reasonable to assume that students at Amherst, Williams, Dartmouth, and Middlebury are homogeneous in their goals. It is particularly difficult to contend this among those students who are most likely to be able to attend the top professional schools if that is what they want. These would be the students who had the strongest records coming out of high school, and who therefore had the most choices about where to go to college.

If you want to isolate treatment effects you would need to look at residuals on the SAT vs % top prof school regression, look at qualifications of students accepted to the top schools, and the relationship between gpa and exam score and admissions probabilities to these schools. If you find a college where the students enroll at top professional schools at rates disproportionate to their high school performance then you have a treatment effect. If they enroll at rates disproportionate to their college performance then you have a preferential admissions effect.

But remember, you only have 39% of the variance to play with.

they look at the normative careers for students at various colleges