ATL 2016 Rankings are Out: Discuss Forum

[guynourmin]

They seem to address the HYS glitch by giving 5% to federal judges and 5% to SCOTUS clerks.

Yep--both of which are totally ridiculous bases for a ranking. (I think fed clerkships would probably be a factor in an ideal law school ranking. I can't think of any advantages--and I can think of many disadvantages--of using just SCOTUS clerks.)

They've addressed this before - whether or not you like their answer is obviously a different story. Said using SCOTUS clerkship only affects the rankings at the highest level. That is, it's 5% of the ranking, but if every school from the top 10 down has 0 SCOTUS clerks in the last 5 years then it has no bearing on their standing compared to one another. Basically, it's their way of giving some bump to y. (Like someone above said, they probably went through a number of variants before coming up with a list they liked.)

[cron1834]

This list is literally as stupid as USNWR. ATL has seriously gone downhill since closing comments, too. SMH.

[jbagelboy]

Chicago ascendant, Harvard/Columbia/NYU in perilous decline. New world order.

Also rpupkin call-in on the savage Return of the UPENN.

These rankings are basically Cooley-level.

[jbagelboy]

Actually, now that I have some time to read it right... are they completely ignoring LRAP in order to make schools with higher PI numbers and high tuition costs seem like they're somehow worse investments?

I know that any ranking list is basically bullshit, but if I'm reading that correctly, this one is objectively idiotic.

That's how I read it as well, which means that Yale (with a great loan assistance program and high average cost proportionately borne more by students who can afford it) is penalized vs BYU (without much of a loan assistance program and a lower average cost borne more evenly by everyone). This is a pretty minor point, though, as most law schools have pretty similar costs. (Also, I'm not sure the data's available to incorporate things like LRAP and PI and even how aid gets distributed into a ranking.) It's also a pretty minor point because average debt is meaningless as debt at graduation is pretty much the only factor that you can individually evaluate with a high degree of certainty.

There are lots of nonsensical things in ATL's rankings, though. I highly doubt that anyone at ATL thinks that this is a particularly valid measure of anything at all. I suspect that this is ATL's attempt to create a methodology that results in a ranking that's similar enough to USNWR to look legit but different enough to generate page views / discussion. My guess is that the first dozen or so ranking methodologies that they tried either resulted in lists that were basically the same as USNWR or were too different to be taken seriously by anyone (e.g., didn't have HYS in the top 10, had a totally random school at 4, etc).

True. I bet they just fuck with it to find a surprising but not entirely absurd result. (It's actually pretty high quality trolling in that respect.) Although US News does the same thing to a certain extent. Including the random surveys of faculty and judges basically just reinforces what everyone *expects* to be right, creating arbitrary distinctions between a handful of spots.

My biggest critique of ATL is that the inclusion of cost is totally asinine, since what student when choosing a school (their stated objective) looks at the average cost of attendance or some theoretical cost that is not their own? Everyone compares the opportunities offered to their individual cost of attendance in choosing between schools.

[abl]

They seem to address the HYS glitch by giving 5% to federal judges and 5% to SCOTUS clerks.

Yep--both of which are totally ridiculous bases for a ranking. (I think fed clerkships would probably be a factor in an ideal law school ranking. I can't think of any advantages--and I can think of many disadvantages--of using just SCOTUS clerks.)

They've addressed this before - whether or not you like their answer is obviously a different story. Said using SCOTUS clerkship only affects the rankings at the highest level. That is, it's 5% of the ranking, but if every school from the top 10 down has 0 SCOTUS clerks in the last 5 years then it has no bearing on their standing compared to one another. Basically, it's their way of giving some bump to y. (Like someone above said, they probably went through a number of variants before coming up with a list they liked.)

Sure: they've addressed most of the criticisms. Most of their responses are absurd. (Including the one you repeat re SCOTUS clerkships. Federal clerkships can fulfill the same function, but do so with far higher numbers and stability--and thus greater validity. I'm actually not sure why they'd use that/judgeships over federal clerkships--as I'd imagine that giving federal clerkships a 10-15% weight would also give HYS a pretty big bump--other than because they seem some value in using factors that seem less obvious.)

[Calbears123]

ND went up to 20, I'm ok with that.

Can anyone explain Southern California's drop from the top 50?

[luckenmeister]

If based on employment, Columbia at 11 makes no sense. NYU did have weaker employment numbers this year, but nothing that warrants the 15th spot. Debt at these schools being in NYC could likely make them worse choices for many students, but i dont think that's how rankings should be assessed.

But other than these two absurdities, the list probably makes more sense than USNWR, and fewer people seem to whine about that. If USNWR uses a school's library as a ranking factor, I think we can give SCOTUS clerkships a pass (Btw, wouldnt scotus clerkships give a fair assessment on how a school competes for top legal jobs?) Alas, it's just one more useless law ranking publication.

[rpupkin]

Can't argue with any of these rankings, frankly.

[rpupkin]

Chicago ascendant, Harvard/Columbia/NYU in perilous decline. New world order.

Also rpupkin call-in on the savage Return of the UPENN.

LOL@Columbia. You guys got stomped by Penn again. One year might've been a fluke, but this repeat performance suggests a new Ivy League Order in which Penn is ascendant and CLS is reeling. You guys can't even stay ahead of Cornell. Just LOL.

TPTB should just move Columbia to the Patriot League or whatever it is they call that thing that Fordham is part of. Very sad.

[jbagelboy]

Chicago ascendant, Harvard/Columbia/NYU in perilous decline. New world order.

Also rpupkin call-in on the savage Return of the UPENN.

LOL@Columbia. You guys got stomped by Penn again. One year might've been a fluke, but this repeat performance suggests a new Ivy League Order in which Penn is ascendant and CLS is reeling. You guys can't even stay ahead of Cornell. Just LOL.

TPTB should just move Columbia to the Patriot League or whatever it is they call that thing that Fordham is part of. Very sad.

CLS was 4th last year ahead of penn and cornell so not exactly a repeat performance, but yea point taken I agree. Its an unmitigated disaster for new york city. Probably too many New York Values. Also Boston/Cambridge. After Belicheck's ruse there was bound to be blowback in the legal world, just as Penn and Sandunsky several years ago.

[rpupkin]

CLS was 4th last year ahead of penn and cornell so not exactly a repeat performance, but yea point taken I agree. Its an unmitigated disaster for new york city. Probably too many New York Values.

According to the OP, CLS was 8th last year, not 4th. Why does Lyin' Bagel not own up to his law school's mediocrity? Sad!

[mornincounselor]

Aren't federal clerkships also included in the "Quality Jobs" 30% metric? That would allow federal clerkships to play a part in more than a third of the criteria. I think 5% for SCOTUS is reasonable, too. Obviously it shouldn't play a part in the vast majority of decisions, but everyone is interested in where the unicorn jobs come from.

I'm a fan of having 10% dedicated to the debt vs jobs, although I'm not sure exactly how those two metrics are calculated.

Overall I think the list is better than US News. I mean c'mon Michigan went up several spots in US News this year . . .

[jbagelboy]

CLS was 4th last year ahead of penn and cornell so not exactly a repeat performance, but yea point taken I agree. Its an unmitigated disaster for new york city. Probably too many New York Values.

According to the OP, CLS was 8th last year, not 4th. Why does Lyin' Bagel not own up to his law school's mediocrity? Sad!

Lol i forgot 2015 was a thing

[Johann]

Of the rankings criteria

15% is education cost
5% is debt per job
5% is salary to debt ratio


25% of the criteria is solely or largely about how much the school costs. Obviously debt is a big part of the equation but isnt this a tad redundant.

for a cost benefit analysis, 100% would seem more appropriate than 25%.

[raven1231]

Akron will no longer be able to claim it's a "top 50 law school"....Or maybe they will just cite the prior rankings

[abl]

Aren't federal clerkships also included in the "Quality Jobs" 30% metric? That would allow federal clerkships to play a part in more than a third of the criteria. I think 5% for SCOTUS is reasonable, too. Obviously it shouldn't play a part in the vast majority of decisions, but everyone is interested in where the unicorn jobs come from.

I'm a fan of having 10% dedicated to the debt vs jobs, although I'm not sure exactly how those two metrics are calculated.

Overall I think the list is better than US News. I mean c'mon Michigan went up several spots in US News this year . . .

Federal clerkships are, but they're weighted equally to biglaw. If the point is that we're trying to differentiate good from great outcomes (in part to distinguish between the tippy top schools and in part because ), you need some sort of more-competitive and more-desirable input. Ideally, your input would be: (1) something that was widely desired across many law schools; (2) something that's highly competitive--enough that you'll see distinctions between T14 schools that are not just based on self-selection; and (3) something that is reported. SCOTUS and federal clerkships both meet all three requirements.

The problem with SCOTUS clerkships is the small numbers. The difference between UVA getting one SCOTUS clerkship in a year or two is meaningless from a statistical standpoint. Yet if we're using SCOTUS clerkship figures as a rankings factor, that counts as a 100% huge improvement in UVA's figures. Now, if we used SCOTUS data from enough years, this problem would decrease. But then you're faced with the problem of basically just having one factor that doesn't change from year-to-year or account for relative changes between the law schools: a fifteen-year running average of SCOTUS clerkships. And, because few schools besides HYS Chicago and UVA place more than a small handful of students in SCOTUS clerkships, you still end up with a small numbers problem when using a fifteen-year running average: there might be real things to learn about the relative differences between, say, Yale and Harvard and Columbia, but the difference between Georgetown and Cornell is still basically noise that can be extraordinarily impacted by a single additional all star student. To top it all off, because SCOTUS clerkships are so particularly connection-based, it is incredibly easy for just one school's connection with just one feeder judge to make a pretty big impact on the numbers. Is UVA a better law school than Penn because Wilkinson likes UVA kids and therefore UVA lands an additional SCOTUS clerkship every other year as compared with Penn? I think the answer is pretty obviously no. But SCOTUS clerkship numbers are incredibly prone to this sort of bias--especially once you drop out of basically the HYS range. As a consequence, really the only thing that SCOTUS clerkship numbers tell you is that HYS are better for ultra-elite outcomes than other law schools.

Using federal clerkship rates instead of SCOTUS numbers basically fix all of these problems. Instead of talking about 1-2 students a year, you're generally talking about double digits. So, Penn having one particularly good student in a given year ends up making a 4-5% difference in that year's ratings, as opposed to a 50% or 100% difference. I could go on and on, but there's really nothing that you get via SCOTUS clerkship numbers that you wouldn't get in a much better form via federal clerkship numbers. Probably most importantly, though, federal clerkship numbers allow you to meaningfully distinguish between schools besides just HYS.

[abl]

So now I'm on a roll. If you were going to design a ranking system, what would it look like?

Mine would probably be a mishmash of ATL's and USNWR's.


I also think that real law jobs matter. I'd probably make that around 50% of the total ranking (probably half and half between real jobs and something similar to ATL's quality jobs).

I think that there are far better outcomes than biglaw. There's no perfect way to account for these, but I think fed clerkship placement isn't bad: these jobs are often or even usually going to be open to federal clerks and will rarely be open to those without federal clerkships. I'd make this maybe 20% of the total. This isn't going to impact schools outside of the T14 much for better or for worse: it'll mostly be a differentiating factor between T14 schools. The only other thing I'd consider is trying to do just fed COA clerkships or some mishmash of uber-prestigious outcomes on a three- or five-year running average: SCOTUS + Bristow + Skadden, or something similar. But I'm pretty skeptical that this captures "great" outcomes better than just using the overall fed clerkship %.

I think resources are worth considering. How many (and how good) speakers a school can bring in, what sort of clinics that school operates, funding to go to conferences, etc, are all things that'll have a material impact on your experience. I'd make this maybe 10% of the total and use funding-per-student or something similar.

I do think that LSAT scores and UG GPA matter. The quality of students in school with you will make a noticeable impact on your experience, and I'm not sure how better to measure this. I'd make them maybe 10% of the total ranking.

Finally, I do think that reputation matters. I'm not a huge fan of USNWR way of doing this, but I'm not sure of a better one either. I'd make this 10% of the total.

I don't think average educational cost or debt matters at all. The only thing that matters is your COA. The most useful rankings are ones that you can use with your COA to make a decision. I would not give any weight to this factor. Ditto alumni satisfaction (also it's too easily manipulated), library size, etc.

---

tl;dr: my ranking would be:
25% Employment Score (similar to ATL)
25% Quality Jobs Score (similar to ATL)
20% Federal Clerkship Score
10% Student Quality Score (LSAT + GPA: similar to USNWR)
10% Reputation Score (similar to USNWR)
10% Resources Score (per-student funding)

---

*70% outputs (jobs)
*10% inputs (student quality)
*10% reputation
*10% resources

[Hikikomorist]

tl;dr: my ranking would be:
2510% Employment Score (similar to ATL)
250% Quality Jobs Score (similar to ATL)BL+FC
20% Federal Clerkship Score
120% Student Quality Score (LSAT + GPA: similar to USNWR)
10% Reputation Score (similar to USNWR)
10% Resources Score (per-student funding)

---

*70% outputs (jobs)
*10% inputs (student quality)
*10% reputation
*10% resources

Most students seem to attend law school targeting BL+FC, so that's what I'd emphasize. Also, most people with clerkships end up going into BL anyway, right? I think inputs should matter more than maybe they do, so I'd be open to changing that number.

[abl]

tl;dr: my ranking would be:
2510% Employment Score (similar to ATL)
250% Quality Jobs Score (similar to ATL)BL+FC
20% Federal Clerkship Score
120% Student Quality Score (LSAT + GPA: similar to USNWR)
10% Reputation Score (similar to USNWR)
10% Resources Score (per-student funding)

---

*70% outputs (jobs)
*10% inputs (student quality)
*10% reputation
*10% resources

Most students seem to attend law school targeting BL+FC, so that's what I'd emphasize. Also, most people with clerkships end up going into BL anyway, right? I think inputs should matter more than maybe they do, so I'd be open to changing that number.

The point of weighting clerkships so heavily is to get a sense for the availability of "better than biglaw" options--and to distinguish among the T14--and not because clerkships themselves are super crucial. Clerkship placement therefore becomes a proxy for "better than biglaw" placement (because clerkship placement is representative of graduate attractiveness). It's probably impossible to track "better than biglaw" outcomes, so if this is something people care about (and they should), it's going to be necessary to use a proxy. Clerkship placement also drops off very quickly and pretty proportionately to access to "better than biglaw" positions. So, although clerkship placement is going to give Yale a pretty big boost (and help account for the fact that a lot of Yale grads end up in "better than biglaw" positions) it's going to make only a small difference between somewhere like UCLA and USC.

So, for the majority of schools that place very few students in clerkships, it's as if that 20% is a non-factor (which makes each of the remaining factors 25% more powerful -- so for comparing Wisconsin to Maryland, for example, you'd really be looking at something more like 31% employment / 31% quality jobs / 12.5% student quality / 12.5% reputation / 12.5% resources (because both schools are going to score in the low single digits for the fed clerkship score).

[OtterLaw]

Akron will no longer be able to claim it's a "top 50 law school"....Or maybe they will just cite the prior rankings

But dat lazy river tho....

[jbagelboy]

So now I'm on a roll. If you were going to design a ranking system, what would it look like?

Mine would probably be a mishmash of ATL's and USNWR's.


I also think that real law jobs matter. I'd probably make that around 50% of the total ranking (probably half and half between real jobs and something similar to ATL's quality jobs).

I think that there are far better outcomes than biglaw. There's no perfect way to account for these, but I think fed clerkship placement isn't bad: these jobs are often or even usually going to be open to federal clerks and will rarely be open to those without federal clerkships. I'd make this maybe 20% of the total. This isn't going to impact schools outside of the T14 much for better or for worse: it'll mostly be a differentiating factor between T14 schools. The only other thing I'd consider is trying to do just fed COA clerkships or some mishmash of uber-prestigious outcomes on a three- or five-year running average: SCOTUS + Bristow + Skadden, or something similar. But I'm pretty skeptical that this captures "great" outcomes better than just using the overall fed clerkship %.

I think resources are worth considering. How many (and how good) speakers a school can bring in, what sort of clinics that school operates, funding to go to conferences, etc, are all things that'll have a material impact on your experience. I'd make this maybe 10% of the total and use funding-per-student or something similar.

I do think that LSAT scores and UG GPA matter. The quality of students in school with you will make a noticeable impact on your experience, and I'm not sure how better to measure this. I'd make them maybe 10% of the total ranking.

Finally, I do think that reputation matters. I'm not a huge fan of USNWR way of doing this, but I'm not sure of a better one either. I'd make this 10% of the total.

I don't think average educational cost or debt matters at all. The only thing that matters is your COA. The most useful rankings are ones that you can use with your COA to make a decision. I would not give any weight to this factor. Ditto alumni satisfaction (also it's too easily manipulated), library size, etc.

---

tl;dr: my ranking would be:
25% Employment Score (similar to ATL)
25% Quality Jobs Score (similar to ATL)
20% Federal Clerkship Score
10% Student Quality Score (LSAT + GPA: similar to USNWR)
10% Reputation Score (similar to USNWR)
10% Resources Score (per-student funding)

---

*70% outputs (jobs)
*10% inputs (student quality)
*10% reputation
*10% resources

The "Resources Score" is a fucking curse that drives up the cost of law school each year. You MUST eliminate that metric since it doesn't measure anything valuable and inflates administrative budgets and tuition cost with no perceivable benefit to students.

Different categories of employment are important, but how do you gauge an accurate clerkship figure? everyone recognizes that the 9-month mark is an atrocious measure of clerkship outcomes, since the majority of students clerking at many top schools do so 1-2 years after graduating. A ranking including the ABA mark would over-inflate schools like Alabama and UGA, that place into regions where less desirable judges are hiring late into 3L for the coming term (11th circuit districts are always last in the country to close), to the detriment of schools where competitive judges are already hiring 2Ls for years out from their graduation date or that require work experience and many semesters of grades. Anyone seeking realistic answers for employment should have a huge problem with this, unless you can collect historical and alumni clerkship data. (The numbers would be striking--for example, over 25% of HLS students and 16% of CLS students will clerk for a federal court, many more than the 9-month figure captures, whereas e.g. Vanderbilt's clerkship numbers are pretty stable).

BL+FC is not sufficient at schools like T6 where nearly all the PI and government hiring is elite (for example, nearly all the fellows at HLS and YLS and pro bono scholars at CLS are in jobs much more competitive than some V100 firm), but including something as vague as "reputation" falls prey to the same stagnation and rankings internal loop that has made US News worse than useless for the past ten years. I'm not sure what to do about this, but having no ranking would probably be the answer.

[jbagelboy]

I also don't think "student quality" as evaluated by college grades and the LSAT exam are meaningful in the least.

[Tempo]

Thank God NYU dropped so far. I was super worried about attending a T6, and now I don't even have the pressure of a T14! Law school will be a breeze now.

[Br3v]

DROVES

[jbagelboy]

tl;dr: my ranking would be:
2510% Employment Score (similar to ATL)
250% Quality Jobs Score (similar to ATL)BL+FC
20% Federal Clerkship Score
120% Student Quality Score (LSAT + GPA: similar to USNWR)
10% Reputation Score (similar to USNWR)
10% Resources Score (per-student funding)

---

*70% outputs (jobs)
*10% inputs (student quality)
*10% reputation
*10% resources

Most students seem to attend law school targeting BL+FC, so that's what I'd emphasize. Also, most people with clerkships end up going into BL anyway, right? I think inputs should matter more than maybe they do, so I'd be open to changing that number.

The point of weighting clerkships so heavily is to get a sense for the availability of "better than biglaw" options--and to distinguish among the T14--and not because clerkships themselves are super crucial. Clerkship placement therefore becomes a proxy for "better than biglaw" placement (because clerkship placement is representative of graduate attractiveness). It's probably impossible to track "better than biglaw" outcomes, so if this is something people care about (and they should), it's going to be necessary to use a proxy. Clerkship placement also drops off very quickly and pretty proportionately to access to "better than biglaw" positions. So, although clerkship placement is going to give Yale a pretty big boost (and help account for the fact that a lot of Yale grads end up in "better than biglaw" positions) it's going to make only a small difference between somewhere like UCLA and USC.

So, for the majority of schools that place very few students in clerkships, it's as if that 20% is a non-factor (which makes each of the remaining factors 25% more powerful -- so for comparing Wisconsin to Maryland, for example, you'd really be looking at something more like 31% employment / 31% quality jobs / 12.5% student quality / 12.5% reputation / 12.5% resources (because both schools are going to score in the low single digits for the fed clerkship score).

As I pointed out above though, clerkships would become a factor. The SEC schools, as just one example, always have more clerks than mid-west schools because of the hiring timelines and structure of the districts near those schools. It doesn't mean UGA is better than Iowa for a student who wants to work at a mid-sized firm in Des Moines or Minneapolis, but its exactly what these god-awful lists encourage.