Top Law Schools

I've put together a ranking specifically for people who are unsure of where they want to practice after graduation. I've used Simpson's Biologic Diversity Index (1-(D= (Σn(n-1))/N(N-1))), where n (in this instance) equals the % of graduates employed in a specific region and N equals the total % whose location is known. The regions evaluated for diversity were the largest markets - Pacific (Cali), Middle Atlantic (NY), South Atlantic (DC), East North Central (IL), and Other. The schools with the highest index have graduates the most spread out, whereas a lower index means the graduates are concentrated heavily into one or two regions.

It's important to note this doesn't say anything about a degrees actual portability power - a great deal of it is self selection. Columbia students chose CLS because they want NYC, and stay there after graduation. A T6 grad will have much more national reach than a Notre Dame graduate. However, this just looks at which schools do have the most spread out students after graduation.

Rank School (1-D)
1 Michigan 0.778
2 Notre Dame 0.772
3 Chicago 0.769
4 WUSTL 0.758
5 Yale 0.755
6 NU 0.736
7 Duke 0.727
8 Vanderbilt 0.718
9 Harvard 0.702
10 Stanford 0.702
11 UVA 0.690
12 GULC 0.663
13 Iowa 0.650
14 BU 0.596
15 Tulane 0.593
16 BYU 0.590
17 Penn 0.588
18 Indiana U 0.581
19 Cornell 0.580
20 W&M 0.570
21 GWU 0.559
22 Case Western 0.556
23 BC 0.555
24 Penn State 0.551
25 Berkeley 0.551
26 Emory 0.549
27 W & L 0.548
28 NYU 0.534
29 Columbia 0.512
30 UIUC 0.509
31 Minnesota 0.507
32 Ohio State 0.501
33 American U 0.487
34 Wisconsin 0.480
35 U of Cincinnati 0.464
36 Wake Forest 0.462
37 UT 0.430
38 UConn 0.385
39 Pepperdine 0.378
40 U Arizona 0.361
41 Alabama 0.340
42 Northeastern 0.330
43 UNC 0.305
44 U Washington 0.297
45 Maryland 0.289
46 U San Diego 0.285
47 Temple 0.281
48 UCLA 0.277
49 U Tennessee 0.276
50 Cardozo 0.256
51 USC 0.256
52 Arizona State 0.255
53 Loyola - Chi 0.254
54 George Mason 0.249
55 Lewis & Clark 0.237
56 Chicago-Kent 0.234
57 U Utah 0.223
58 U Georgia 0.219
59 Florida 0.216
60 Florida State 0.212
61 Kentucky 0.208
62 UC-Davis 0.200
63 Fordham 0.195
64 Baylor 0.183
65 UC-Hastings 0.171
66 U Houston 0.170
67 Oklahoma 0.157
68 Seton Hall 0.145
69 Richmond 0.137
70 Brooklyn 0.131
71 Colorado 0.116
72 Georgia State 0.115
73 UNLV 0.087
74 SMU 0.080
75 Loyola - LA 0.040



By Region:
Midwest
Rank School (1-D)
1 Michigan 0.778
2 Notre Dame 0.772
3 Chicago 0.769
4 WUSTL 0.758
5 NU 0.736
6 Iowa 0.650
7 Indiana U 0.581
8 Case Western 0.556
9 UIUC 0.509
10 Minnesota 0.507
11 Ohio State 0.501
12 Wisconsin 0.480
13 U of Cincinnati 0.464
14 Loyola - Chi 0.254
15 Chicago-Kent 0.234
16 Kentucky 0.208
17 Oklahoma 0.157


New York
1 Cornell 0.580
2 NYU 0.534
3 Columbia 0.512
4 Cardozo 0.256
5 Fordham 0.195
6 Seton Hall 0.145
7 Brooklyn 0.131


Northeast
1 Yale 0.755
2 Harvard 0.702
3 BU 0.596
4 Penn 0.588
5 BC 0.555
6 Penn State 0.551
7 UConn 0.385
8 Northeastern 0.330
9 Temple 0.281



Pacific
1 Stanford 0.702
2 Berkeley 0.551
3 Pepperdine 0.378
4 U Washington 0.297
5 U San Diego 0.285
6 UCLA 0.277
7 USC 0.256
8 Lewis & Clark 0.237
9 UC-Davis 0.200
10 UC-Hastings 0.171
11 Loyola - LA 0.040



South Atlantic
1 Duke 0.727
2 Vanderbilt 0.718
3 UVA 0.690
4 GULC 0.663
5 Tulane 0.593
6 W&M 0.570
7 GWU 0.559
8 Emory 0.549
9 W & L 0.548
10 American U 0.487
11 Wake Forest 0.462
12 Alabama 0.340
13 UNC 0.305
14 Maryland 0.289
15 U Tennessee 0.276
16 George Mason 0.249
17 U Georgia 0.219
18 Florida 0.216
19 Florida State 0.212
20 Richmond 0.137
21 Georgia State 0.115



Southwest
1 BYU 0.590
2 UT 0.430
3 U Arizona 0.361
4 Arizona State 0.255
5 U Utah 0.223
6 Baylor 0.183
7 U Houston 0.170
8 Colorado 0.116
9 UNLV 0.087
10 SMU 0.080




Here's the data:
https://spreadsheets.google.com/pub?hl= ... utput=html

you sir are an awesome poster. this isn't incredibly valuable, but definitely interesting.

I've been wondering about this for a while. Thank you for posting it in terms of understandable (and concrete) data. Your point about self-selection was also very good.

Really, really interesting. Thanks!

Good data, not incredibly informative, but still interesting. The results do make a lot of sense though. The schools that are higher are ones that don't have a strong "home" market. NY and DC schools are lower, Cali/West Coast schools are really low, and midwest schools are high on the list.

Cool stuff.

I just wanted to note that Tulane is not included on this list.

pinkzeppelin wrote:I just wanted to note that Tulane is not included on this list.

It just goes to the T30. Everything after that is already known to be very strongly regional.

If you're curious for personal reasons, Tulane's index is 0.593, which actually puts it somewhat high on the list (higher than I expected). They do put 18.5% into the Middle Atlantic and 13% into South Atlantic (not New Orleans).

the problem with using D here is you are lacking a control variable. With the simpson bio div index, one compares diversity within a single region. Here, for D to be meaningful, you want to take all midwest schools and compare their D (or 1-D, if you prefer); then all eastern schools and compare their D; etc., rather than just haphazardly compare every school's D to each other/doing cross-regional comparisons.

You essentially did all the work, just one final grouping step is needed for the data to be more informative. When you do that, the 1-D index becomes positively coordinated with usnews rank (with a few outliers like Notre Dame -- which is where this index becomes useful for decisionmaking), and people will stop wondering why yale is number 5 on that list and/or whether a michigan degree is more "portable" than one from yale.

solidsnake wrote:the problem with using D here is you are lacking a control variable. With the simpson bio div index, one compares diversity within a single region. Here, for D to be meaningful, you want to take all midwest schools and compare their D (or 1-D, if you prefer); then all eastern schools and compare their D; etc., rather than just haphazardly compare every school's D to each other/doing cross-regional comparisons.

You essentially did all the work, just one final grouping step is needed for the data to be more informative. When you do that, the 1-D index becomes positively coordinated with usnews rank (with a few outliers like Notre Dame -- which is where this index becomes useful for decisionmaking), and people will stop wondering why yale is number 5 on that list and/or whether a michigan degree is more "portable" than one from yale.

I see what you're saying. I'll mess with it in a little bit, I'm working on something else atm. Group by the primary market they feed to? It'll just look a little weird listing Harvard as a Pacific school.

3 Chicago 0.769

This one was fairly surprising to me, I thought the prevailing wisdom was that UChi grads stayed in Chi?

Zabini wrote:

3 Chicago 0.769

This one was fairly surprising to me, I thought the prevailing wisdom was that UChi grads stayed in Chi?

From the data, only about 36% do, a sizable chunk (40%) goes to California and NYC

Nightrunner wrote:On behalf of a board too often overrun with people arguing over whose secondhand anecdotes are more common: thanks for actual statistics.

I'm stickying this thread for a few days.

yes, and you should sticky his/her other thread on '09 employment data.

aliarrow wrote:

solidsnake wrote:the problem with using D here is you are lacking a control variable. With the simpson bio div index, one compares diversity within a single region. Here, for D to be meaningful, you want to take all midwest schools and compare their D (or 1-D, if you prefer); then all eastern schools and compare their D; etc., rather than just haphazardly compare every school's D to each other/doing cross-regional comparisons.

You essentially did all the work, just one final grouping step is needed for the data to be more informative. When you do that, the 1-D index becomes positively coordinated with usnews rank (with a few outliers like Notre Dame -- which is where this index becomes useful for decisionmaking), and people will stop wondering why yale is number 5 on that list and/or whether a michigan degree is more "portable" than one from yale.

I see what you're saying. I'll mess with it in a little bit, I'm working on something else atm. Group by the primary market they feed to? It'll just look a little weird listing Harvard as a Pacific school.

no, just group it by market/region in which the school is physically located, since you're essentially translating diversity = dispersion from the school's location (into other markets/regions).

Statistics and facts are so damn refreshing to see. Keep it up OP.

Would like to see regional groupings. Would also like to see more schools listed.

Something along the lines of below. The problem here is the regions are too big and too few. I'm not sure if you aggregated your data in the spreadsheet to get them or thats how they came, but I would suggest slicing them down a bit further. Combining DC/VA schools with northeast schools just felt wrong. Moreover, as OP mentioned, self-selection bias taints pretty much all inferential statistics so take any and all inputs with a grain a salt. Again, the usefulness comes from identifying schools that have higher indexes relative to their regional peers than usnews ranking would suggest. And finally, the integrity of the employment data, incl the type of employment is always questionable when culled from self-reporting schools.

Rank School (1-D)

East North Central
1 Michigan 0.778
2 Notre Dame 0.772
3 Chicago 0.769
4 WUSTL 0.758
6 NU 0.736
7 Iowa 0.650
8 Indiana U 0.581
9 UIUC 0.509
10 Minnesota 0.507


Northeast/Middle Atlantic
1 Yale 0.755
2 Harvard 0.702
3 UVA 0.690
4 GULC 0.663
5 BU 0.596
6 Penn 0.588
7 Cornell 0.580
8 W&M 0.570
9 GWU 0.559
10 BC 0.555
11 W & L 0.548
12 NYU 0.534
13 Columbia 0.512
14 Fordham 0.195

South Atlantic/Texas
1 Duke 0.727
2 Vanderbilt 0.718
3 Emory 0.549
4 UT 0.430
5 UNC 0.305


Pacific
1 Stanford 0.702
2 Berkeley 0.551
3 U Washington 0.297
4 UCLA 0.277
5 USC 0.256
6 UC-Davis 0.200

solidsnake wrote:Something along the lines of below. The problem here is the regions are too big and too few. I'm not sure if you aggregated your data in the spreadsheet to get them or thats how they came, but I would suggest slicing them down a bit further. Combining DC/VA schools with northeast schools just felt wrong. Moreover, as OP mentioned, self-selection bias taints pretty much all inferential statistics so take any and all inputs with a grain a salt.

Unfortunately, the groupings are how they are given from US News, the only thing that can be further separated is the Other category (which is mostly just empty regions, and Texas). I still think this list serves a purpose though, which is why I posted it. It's value is limited by certain factors which you've pointed out.

aliarrow wrote:

solidsnake wrote:Something along the lines of below. The problem here is the regions are too big and too few. I'm not sure if you aggregated your data in the spreadsheet to get them or thats how they came, but I would suggest slicing them down a bit further. Combining DC/VA schools with northeast schools just felt wrong. Moreover, as OP mentioned, self-selection bias taints pretty much all inferential statistics so take any and all inputs with a grain a salt.

Unfortunately, the groupings are how they are given from US News, the only thing that can be further separated is the Other category (which is mostly just empty regions, and Texas). I still think this list serves a purpose though, which is why I posted it. It's value is limited by certain factors which you've pointed out.

Oh i agree it has value, which is why i helped organize the stats.

Another way of communicating this data is that the index may show the degree to which grads leave their school's home market (as defined by physical location) compared to other schools in that same region. Of course, whether that migration or lack thereof is by choice isn't clear.

Filed under people who should be working at McKinsey or BCG.

Nice job solidsnake, ty! I agree with your concerns and would make these tweaks.

I'd keep DC schools with the Northeast/Midatlantic and move the Virginia schools into the South. I'd also make the NYC schools their own region: the size of the legal market there warrants it and that helps for the strong self-select to NYC effect that those schools will have.


When you do that:

East North Central
1 Michigan 0.778
2 Notre Dame 0.772
3 Chicago 0.769
4 WUSTL 0.758
5 NU 0.736
6 Iowa 0.650
7 Indiana U 0.581
8 UIUC 0.509
9 Minnesota 0.507


Northeast/Middle Atlantic
1 Yale 0.755
2 Harvard 0.702
3 GULC 0.663
4 BU 0.596
5 Penn 0.588
6 Cornell 0.580
7 GWU 0.559
8 BC 0.555


NYC
1 NYU 0.534
2 Columbia 0.512
3 Fordham 0.195


South Atlantic/Texas
1 Duke 0.727
2 Vanderbilt 0.718
3 UVA 0.690
4 W&M 0.570
5 Emory 0.549
6 W & L 0.548
7 UT 0.430
8 UNC 0.305


Pacific
1 Stanford 0.702
2 Berkeley 0.551
3 U Washington 0.297
4 UCLA 0.277
5 USC 0.256
6 UC-Davis 0.200

Zabini wrote:Nice job solidsnake, ty! I agree with your concerns and would make these tweaks.

I'd keep DC schools with the Northeast/Midatlantic and move the Virginia schools into the South. I'd also make the NYC schools their own region: the size of the legal market there warrants it and that helps for the strong self-select to NYC effect that those schools will have.

I really think VA schools should be kept in South Atlantic. Many of them (W&L, W&M, GMU) are all feeders into the DC market. I'm continuing to update, I'm about up to UC-Hastings in my data.

Cool. Tag.

solidsnake wrote:the problem with using D here is you are lacking a control variable. With the simpson bio div index, one compares diversity within a single region. Here, for D to be meaningful, you want to take all midwest schools and compare their D (or 1-D, if you prefer); then all eastern schools and compare their D; etc., rather than just haphazardly compare every school's D to each other/doing cross-regional comparisons.

You essentially did all the work, just one final grouping step is needed for the data to be more informative. When you do that, the 1-D index becomes positively coordinated with usnews rank (with a few outliers like Notre Dame -- which is where this index becomes useful for decisionmaking), and people will stop wondering why yale is number 5 on that list and/or whether a michigan degree is more "portable" than one from yale.

Thanks for your original idea btw... This does make things much clearer.

aliarrow wrote:

solidsnake wrote:the problem with using D here is you are lacking a control variable. With the simpson bio div index, one compares diversity within a single region. Here, for D to be meaningful, you want to take all midwest schools and compare their D (or 1-D, if you prefer); then all eastern schools and compare their D; etc., rather than just haphazardly compare every school's D to each other/doing cross-regional comparisons.

You essentially did all the work, just one final grouping step is needed for the data to be more informative. When you do that, the 1-D index becomes positively coordinated with usnews rank (with a few outliers like Notre Dame -- which is where this index becomes useful for decisionmaking), and people will stop wondering why yale is number 5 on that list and/or whether a michigan degree is more "portable" than one from yale.

Thanks for your original idea btw... This does make things much clearer.

no prob. thanks for using the 1-(D= (Σn(n-1))/N(N-1)) formula. This is great to test the degree of diversity among various shareholder groups in target companies. good stuff. We need more quants on these boards.

thanks for doing this. Very interesting.