Top Law Schools

Unemployed wrote:This is the fatal flaw. Even if Yale's median is 172, and Cooley's is 145 (I am pulling these #'s out of thin air), Cooley will win the "median LSAT" battle because 172x700 < 145x3000.

Cooley's full-time enrollment is only 560 (the other ~2,500 are part-timers). Also, Yale is #25, and Cooley is #189 (unsurprising since they scored close to dead last on all but one factor).

Splitt3r wrote:Harvard doesn't give merit money, so how did they get such a low score on the grants thing? Is it because they're particularly generous due to their large endowment? Doesn't really seem fair to downgrade someone because they are more responsible about helping out people who can't afford the university than other schools. It also seems to be biased somewhat against private schools, as they as a rule are more expensive and thus are more likely to have people attending who qualify for need-based aid.

I think this is just another way of saying that, all else equal, people prefer to attend the cheaper school. That's sort of the point -- if a school is having to give away lots of large grants in order to close the deal with the same class of applicants as another, cheaper school that is equal in every other factor, then that just means that the first school's high cost was a relative turnoff.

If two schools were equally appealing in every way and one was more "responsible" in giving out financial aid, then we would expect that school to do better on all the other factors here -- they would make fewer offers, would recruit a generally higher scoring class, and would have fewer transfers out. So we would expect that they in fact wind up ranking higher (i.e. appearing more desirable) compared with the stingy school that candidates might otherwise prefer just as well.

In as much as they're generally more expensive, then it probably does have some bias against private schools. But it also certainly has some bias against schools in places that have crappy weather, schools in ridiculously expensive cities, and schools with any number of other unpleasant factors. Again, that's sort of the point. It doesn't pretend to be a measure of the "best" schools. It's just an attempt to identify schools that candidates, on the whole, most want (i.e. need the least encouragement) to attend.

snotrocket wrote:

Unemployed wrote:This is the fatal flaw. Even if Yale's median is 172, and Cooley's is 145 (I am pulling these #'s out of thin air), Cooley will win the "median LSAT" battle because 172x700 < 145x3000.

Cooley's full-time enrollment is only 560 (the other ~2,500 are part-timers). Also, Yale is #25, and Cooley is #189 (unsurprising since they scored close to dead last on all but one factor).

Am I reading something wrong? It appears that Cooley does beat Yale in your rankings (with respect to the LSAT score)

doyleoil wrote:

lex talionis wrote:

doyleoil wrote:

lex talionis wrote:Doylie, that aussie is not that cute.

you are wrong - please see my swimming thread for proof

no thanks.

And if I'm fucking a swimmer, it's going to be Ryan Lochte.

just for you, doll:



broken_image
SO HOT. Dumb as a rock, but SO HOT.
broken_image

Unemployed wrote:Am I reading something wrong? It appears that Cooley does beat Yale in your rankings (with respect to the LSAT score)

Well, 0.596 > 0.450, so maybe I just was not clear in how I explained it. It's a quick and dirty formula, and there are schools that score ahead of Yale, for more or less the reason that you're talking about. But Cooley is not one of those schools. There are other approaches, like sorting by LSAT and then by class size, and using the resulting rank as the measure for that factor, and that might seem less anomalous.

snotrocket wrote:

Unemployed wrote:Am I reading something wrong? It appears that Cooley does beat Yale in your rankings (with respect to the LSAT score)

Well, 0.596 > 0.450, so maybe I just was not clear in how I explained it. It's a quick and dirty formula, and there are schools that score ahead of Yale, for more or less the reason that you're talking about. But Cooley is not one of those schools. There are other approaches, like sorting by LSAT and then by class size, and using the resulting rank as the measure for that factor, and that might seem less anomalous.

So do you truly believe that the University of South Dakota, Suffolk University, William Mitchell, South Texas, and St. Mary's are all more desirable law schools than Yale?

snotrocket wrote:

Splitt3r wrote:Harvard doesn't give merit money, so how did they get such a low score on the grants thing? Is it because they're particularly generous due to their large endowment? Doesn't really seem fair to downgrade someone because they are more responsible about helping out people who can't afford the university than other schools. It also seems to be biased somewhat against private schools, as they as a rule are more expensive and thus are more likely to have people attending who qualify for need-based aid.

I think this is just another way of saying that, all else equal, people prefer to attend the cheaper school. That's sort of the point -- if a school is having to give away lots of large grants in order to close the deal with the same class of applicants as another, cheaper school that is equal in every other factor, then that just means that the first school's high cost was a relative turnoff.

If two schools were equally appealing in every way and one was more "responsible" in giving out financial aid, then we would expect that school to do better on all the other factors here -- they would make fewer offers, would recruit a generally higher scoring class, and would have fewer transfers out. So we would expect that they in fact wind up ranking higher (i.e. appearing more desirable) compared with the stingy school that candidates might otherwise prefer just as well.

In as much as they're generally more expensive, then it probably does have some bias against private schools. But it also certainly has some bias against schools in places that have crappy weather, schools in ridiculously expensive cities, and schools with any number of other unpleasant factors. Again, that's sort of the point. It doesn't pretend to be a measure of the "best" schools. It's just an attempt to identify schools that candidates, on the whole, most want (i.e. need the least encouragement) to attend.

That's not at all a way of saying that. If we're talking about merit-aid only, sure you could say that. When we're talking about need-based money it doesn't work anymore.

This argument makes sense, except that some schools beat those above them in every category except the $$$ one, and are still ranked below these other schools. You'd need to play with the weights of categories for it to be at all valid.

I think it's fair to say that when 4 TTTs are top 20 schools, the factors are a poor measure of desirability.

snotrocket wrote:

Unemployed wrote:Am I reading something wrong? It appears that Cooley does beat Yale in your rankings (with respect to the LSAT score)

Well, 0.596 > 0.450, so maybe I just was not clear in how I explained it. It's a quick and dirty formula, and there are schools that score ahead of Yale, for more or less the reason that you're talking about. But Cooley is not one of those schools. There are other approaches, like sorting by LSAT and then by class size, and using the resulting rank as the measure for that factor, and that might seem less anomalous.

Ah, so I did read it wrong. My apologies.

But yes - something's wrong when Florida gets a .973 and Yale gets a .596...

Splitt3r wrote:I think it's fair to say that when 4 TTTs are top 20 schools, the factors are a poor measure of desirability.

Now you're just saying that if it departs too much from U.S. News it must be wrong. I'm not saying there's nothing wrong with with, but that's not a convincing reason to conclude that there is. Like most alternative ratings of schools, it's probably most useful -- if it's useful at all -- for comparing them to each other within a given tier (i.e. taking just the T14/T1/T2/T3/T4 schools each as a group, and see how they sort out within their tier based on this). I meant to add by-tier rankings, but just haven't gotten around to them yet.

Unemployed wrote:But yes - something's wrong when Florida gets a .973 and Yale gets a .596...

Maybe. Florida had (2008 numbers) 1,364 students, and 682 of them scored at the 78th percentile or higher on the LSAT. Yale had 576, and half of them (288) scored at the 99th or above (assuming relatively uniform statistics for all class years enrolled). Which of those recruiting outcomes reflects the school that has an easier time attracting law school candidates who have a range of choices in what school to attend? I'm not really sure. Keep in mind though that it's not so much meant to measure "who you would rather get with" as it is "who gets with better, more often, with less effort." It's a little like comparing the guy who has bagged 5 supermodels in 5 months and the guy who goes home with a different 8 every night of the week. To some extent those just aren't commensurable outcomes. They're different audiences and techniques, obviously. Is one more impressive than the other? Maybe. Or maybe not.

I've tried, as an alternative, ranking by LSAT, then class size, and using that rank as a measure. It changes less than you would imagine. Yale moves from 25 to 9, South Texas goes from 7 to 44, and Cooley falls to dead last. Florida, Harvard, and UNC still take the top 3 spots. Some other schools move around as well obviously. The issue then is that it basically neglects the magnitude of difference between a school with 50 people who scored over 172 and a school with 500. That's a huge difference in terms of actual recruiting outcomes. I'm not sure if there's a good way to account for that difference though without over-accounting for it. Of others that I played with, the ones that didn't rely on sorting by LSAT and FTE all amounted to the same scoring as just using the product of LSAT*FTE, and that was the simplest formula to explain.

snotrocket wrote:

Unemployed wrote:But yes - something's wrong when Florida gets a .973 and Yale gets a .596...

Maybe. Florida had (2008 numbers) 1,364 students, and 682 of them scored at the 78th percentile or higher on the LSAT. Yale had 576, and half of them (288) scored at the 99th or above (assuming relatively uniform statistics for all class years enrolled). Which of those recruiting outcomes reflects the school that has an easier time attracting law school candidates? I'm not really sure. Keep in mind though that it's not so much meant to measure "who you would rather get with" as it is "who gets with better, more often, with less effort." It's a little like comparing the guy who has bagged 5 supermodels in 5 months and the guy who goes home with a different 8 every night of the week. To some extent those just aren't commensurable outcomes. They're different audiences and techniques, obviously. Is one more impressive than the other? Maybe. Or maybe not.

I've tried, as an alternative, ranking by LSAT, then class size, and using that rank as a measure. It changes less than you would imagine. Yale moves from 25 to 9, South Texas goes from 7 to 44, and Cooley falls to dead last. Florida, Harvard, and UNC still take the top 3 spots. Some other schools move around as well obviously. The issue then is that it basically neglects the magnitude of difference between a school with 50 people who scored over 172 and a school with 500. That's a huge difference in terms of actual recruiting outcomes. I'm not sure if there's a good way to account for that difference though without over-accounting for it. Of others that I played with, the ones that didn't rely on sorting by LSAT and FTE all amounted to the same scoring as just using the product of LSAT*FTE, and that was the simplest formula to explain.

What super model?

deleted

Her?

deleted

Or her? Big difference you know.

Unemployed wrote:But yes - something's wrong when Florida gets a .973 and Yale gets a .596...

Went back and recalculated with a different approach, estimating a sort of "market share" for each school based on median LSAT. Taking the percentile for the school's median LSAT, I multiplied out how many candidates per 100,000 would get that score, then figured the ratio of 1/2 the full-time enrollment to that "market." This seems like it sensibly accounts for class size, without over-accounting for it the way simple multiplication might. For the new top 30, this is how it works out:

X "Market Share" ratio
M Number per 100,000 candidates who score at or above this school's median LSAT
E Half of the full-time enrollment

X M E
0.032 21,600 682 University of Florida (Levin)
0.027 13,200 356 University of North Carolina-Chapel Hill
0.716 1,200 860 Harvard University (MA)
0.030 11,100 335 University of Georgia
0.513 1,200 615 Columbia University (NY)
0.121 2,200 267 Stanford University (CA)
0.025 9,100 230 Brigham Young University (Clark) (UT)
0.320 900 288 Yale University (CT)
0.106 6,200 657 University of Texas-Austin
0.016 21,600 353 University of South Carolina
0.153 3,700 565 University of Michigan-Ann Arbor
0.018 24,600 432 Seattle University
0.022 18,300 408 University of Houston
0.080 7,400 593 Fordham University (NY)
0.424 1,700 721 New York University
0.021 13,200 278 Rutgers, the State Univ. of New Jersey-Camden
0.198 2,900 573 University of Virginia
0.012 27,800 328 Indiana University-Indianapolis
0.176 1,700 300 University of Chicago
0.002 47,400 116 University of South Dakota
0.010 21,600 227 Georgia State University
0.026 15,700 409 University of Wisconsin-Madison
0.011 43,200 457 South Texas College of Law
0.005 35,200 171 Campbell University (Wiggins) (NC)
0.047 13,200 620 University of California (Hastings)
0.025 11,100 277 Southern Methodist University (TX)
0.096 7,400 714 George Washington University (DC)
0.025 13,200 337 University of Maryland
0.003 43,200 121 University of Montana

So Yale draws 32% of people who scored in the 99th percentile. Florida nabbed 3.2% of people who scored in the 78th and higher. Harvard will draw a whopping 72 out of every 100 people who scored 172 or above (based on an estimated 100,000 candidates).

I think that this is definitely a useful endeavor, but still probably needs some tweaks. It sees a little strange to be adding up unweighted percentiles when there are such large outliers in certain categories. For example, Harvard's market share ratio is 22 times that of Florida, but Florida gets about 83% as many "desirability points" as Harvard does from this category. I'd suggest giving scores based on z-scores (subtract off mean, divide by standard deviation) rather than percentiles if these things are at all close to normally distributed.

jco wrote:I think that this is definitely a useful endeavor, but still probably needs some tweaks. It sees a little strange to be adding up unweighted percentiles when there are such large outliers in certain categories. For example, Harvard's market share ratio is 22 times that of Florida, but Florida gets about 83% as many "desirability points" as Harvard does from this category. I'd suggest giving scores based on z-scores rather than percentiles if these things are at all close to normally distributed.

I'll take a look at using z-scores, but I'm not sure it's going to change the ordering by much, if at all.

i will add my contribution:

broken_image

Mr. Matlock wrote:

doyleoil wrote:it's been said before, possibly even tonight, but TTIWWOP

God I'm in love with this woman!

Um. who is this

ace0260 wrote:Um. who is this

Allison Stokke.

You do not know who she is? Well, either you are not a man or you have no internet.

Here's some more Delicious:

jco wrote:I think that this is definitely a useful endeavor, but still probably needs some tweaks. It sees a little strange to be adding up unweighted percentiles when there are such large outliers in certain categories. For example, Harvard's market share ratio is 22 times that of Florida, but Florida gets about 83% as many "desirability points" as Harvard does from this category. I'd suggest giving scores based on z-scores (subtract off mean, divide by standard deviation) rather than percentiles if these things are at all close to normally distributed.

The market share values are exponential rather than normal (not surprising, considering the nature of the data). The others seem more or less normal.

broken_images

broken_images

Clearly a joint OperaSoprano/Mormon conspiracy to put BYU and Fordham in the T14. Such a plot shall never work!

SoftBoiledLife wrote:Clearly a joint OperaSoprano/Mormon conspiracy to put BYU and Fordham in the T14. Such a plot shall never work!

Curses. Foiled again.

lawfool wrote:Who is this?

Sara Carbonero


Leryn Franco:


another pole vaulter ftw

T-14, BC, ND, GW