Law School Discussion

Nine Years of Discussion
;

Author Topic: Does LSAT predict Bar Exam performance?  (Read 7191 times)

Esq

  • Sr. Citizen
  • ****
  • Posts: 109
    • View Profile
Does LSAT predict Bar Exam performance?
« on: April 16, 2005, 02:36:08 PM »
Is an LSAT score a good predictor of someone's ability to pass the Bar Exam on the first attempt? 

One author has recently argued that the average LSAT scores peaked for first-year students at ABA schools in 1991 at 158.5 (representing the 75th percentile). He then noted that the current mean for all ABA schools was 155 (representing the 64th percentile). He wrote that students with a 155 have a "predictive pass rate" on the bar exam of about 72.5%. Nationally, the bar pass rate has fallen to below 75% (year 2000 data). He then argued that there is a nearly perfect correlation between LSAT score and passing the bar on the first attempt. The correlation he cited was 0.91 to 0.94--a perfect correlation is 1.0. See National Bar Examiner, (vol. 73, no 4 p 11-13).

However, the Law School Admission Council ("LSAC") wrote a Letter to the Editor in response to the author's claims. They said it was not statistically appropriate to only use aggregate data from the law schools (in other words, a law school's mean LSAT) to "predict" how an INDIVIDUAL student of that law school will perform on the bar exam. LSAC had conducted a lengthy study that evaluated data from over 23,000 INDIVIDUAL members of the law school class that entered in 1991. They found that LAW SCHOOL GRADES had the highest correlation with bar passage. (.38 - .41). LSAT score came in second (.30).  In other words, the vast majority of the factors that can predict how an individual will perform on the bar exam are unknown or not captured by these statistics. Law school grades, however, are a stronger predictor of bar exam success than the LSAT score.

Quoting from the LSAC

"LSAC, as sponsor of the LSAT, always walks a fine line between defending the utility of its test against the test's critics, and helping law schools understand that its utility as an admission tool is limited.  Should the mistaken notion that a school's LSAT mean has a direct and nearly perfect relationship with bar passage really take hold, there is a great risk that admissions committees will begin to evaluate individual applicants with that group result in mind.  Such a situation will almost certainly lead to an overreliance on the LSAT score in the admissions process, with one potential result being a decrease in racial and ethnic diversity among law schools." The Bar Examiner, February 2005, p 42. 

AllisonAzee

  • Sr. Citizen
  • ****
  • Posts: 157
    • AOL Instant Messenger - AllisonAzee
    • View Profile
Re: Does LSAT predict Bar Exam performance?
« Reply #1 on: April 16, 2005, 02:50:35 PM »
The LSAT is generally a good predictor for how well you will do your first year in law school, i.e. whether you'll flunk out or not.  It's not a very good predictor of much else.  Of course you take a person with a good LSAT, they are more likely to go to a better school that will prepare them for the Bar and having gotten a good LSAT score (most likely through studying) they are more likely to be better test takers and to prepare more for the bar.  But I don't think that makes the LSAT a good predictor for who will pass the Bar, more that the sort of person that does well on the LSAT is likely to do well at other things too (I'm talking generally, don't yell at me if you got a low sore, they're mountains of exceptions to the rule).
Time to commit: American
They like me: American, Loyola, Santa Clara, U of OR, U of Hawaii
They love me ($):Pepperdine, U of SF, Northeastern, Seattle, Lewis and Clark
Aren't ready to commit: U of Wisconsin, U of Maryland
B@stards won't return my calls: U of San Diego
F#ckers: Hastings

melissamw

  • Sr. Citizen
  • ****
  • Posts: 671
    • View Profile
    • Email
Re: Does LSAT predict Bar Exam performance?
« Reply #2 on: April 16, 2005, 03:36:45 PM »
I buy this argument, Allison!  Obviously, someone who has scored well on the LSAT can't go take the Bar without some prep and expect to pass--- you just can't pass the bar without some training in law school.  But, the type of person who studies for the LSAT will also probably do some major prep for the Bar.  The amount of prep work you do is probably the link in the correlation.

BoscoBreaux

  • Sr. Citizen
  • ****
  • Posts: 789
    • View Profile
    • Email
Re: Does LSAT predict Bar Exam performance?
« Reply #3 on: April 16, 2005, 10:17:42 PM »
Yes, LSAT does predict Bar Exam performance, in the same way that IQ tests and Undergraduate GPAs do as well--but only when large numbers of persons are involved. All things being equal, if 1,000 persons who scored a 155 on the LSAT take the Bar, fewer of them will pass than a set of 1,000 persons who scored a 160. That is why, to a large extent, Bar passage rates of schools are so predictable--the schools with the high average LSATs have higher Bar passage rates than schools with lower average LSATs (note there are some exceptions, but they tend to prove the general rule).
BUT, if you take a random person who got a 155, and a random person who got a 160, you cannot predictably tell who will do better on the Bar.

AllisonAzee

  • Sr. Citizen
  • ****
  • Posts: 157
    • AOL Instant Messenger - AllisonAzee
    • View Profile
Re: Does LSAT predict Bar Exam performance?
« Reply #4 on: April 17, 2005, 04:30:22 PM »
And I think it is also to say that because some one did bad on the LSAT they will do badly on the Bar.  I don't think the skills/knowledge/appititude for each test is the same, but in the highly generally sense the kind of person who does well on the LSAT for reasons X, Y and Z is likely to go to a better law school, study more and prepare better for the bar.  But its not like an algebra test predicting how well you'll do in trig. 
Time to commit: American
They like me: American, Loyola, Santa Clara, U of OR, U of Hawaii
They love me ($):Pepperdine, U of SF, Northeastern, Seattle, Lewis and Clark
Aren't ready to commit: U of Wisconsin, U of Maryland
B@stards won't return my calls: U of San Diego
F#ckers: Hastings

AllisonAzee

  • Sr. Citizen
  • ****
  • Posts: 157
    • AOL Instant Messenger - AllisonAzee
    • View Profile
Re: Does LSAT predict Bar Exam performance?
« Reply #5 on: April 17, 2005, 04:32:29 PM »
Oh and on another sort of random point.  I think this is the same sort of reason that higher ranked schools have better bar passage rates.  Part of it is, sure, the schools themselves.  But an even bigger reason is that better students go to better schools.  If a student who got into Yale, well to a TTT they would still almost certainly pass the bar easily and be very good lawyers.  To a large extent students make the school.
Time to commit: American
They like me: American, Loyola, Santa Clara, U of OR, U of Hawaii
They love me ($):Pepperdine, U of SF, Northeastern, Seattle, Lewis and Clark
Aren't ready to commit: U of Wisconsin, U of Maryland
B@stards won't return my calls: U of San Diego
F#ckers: Hastings

Esq

  • Sr. Citizen
  • ****
  • Posts: 109
    • View Profile
Re: Does LSAT predict Bar Exam performance?
« Reply #6 on: April 23, 2005, 10:01:56 AM »
The LSAC is saying that law school G.P.A. is a better predictor of success for an individual on the bar exam than the LSAT is.  Another point they are making is that if you take a school with an average LSAT of 164 and a bar pass rate of 89 percent, you cannot take an individual with an LSAT score of 164 and say that person has an 89 percent chance of passing the bar on the first attempt.

Other studies at the law school level have shown if you know an individual's law school GPA and tried to predict the individual's success on the bar exam, you would be wrong as to the reasons almost 70 percent of the time, because you can only predict with about 30 percent accuracy. If you only know the individual's LSAT score, and you tried to predict, you would be wrong 98.6 percent of the time.

In Rebecca Luczycki's article, Rankings Issue May Change Test Score Reporting, National Jurist, Feb 2003, page 15, LSAC spokesperson, Ed Haggerty, was quoted as saying: "Many law schools are forced, in order to increase their rankings, to overemphasize LSAT scores in their admissions process.  The LSAT score has been skewed, and not for the purpose of assembling the best class, but more for the purpose of manipulating the rankings.  That's not the purpose of the LSAT."

Interestingly, one LSAC "reform" proposal to address the problem is to have LSAC release only the percentile ranking for a student, not the LSAT "score."  In other words, the law school wouldn't even see a 158 or 160 LSAT, it would only see that student X was in the 76th percentile.  The other "reform" proposal the LSAC working group studying the issue is developing is an option to report LSAT scores in the form of a predicted first-year GPA at a specific law school. Id.   

BAFF213

  • Guest
Re: Does LSAT predict Bar Exam performance?
« Reply #7 on: April 23, 2005, 05:19:02 PM »
interesting thread

hilljack

  • Sr. Citizen
  • ****
  • Posts: 1669
    • View Profile
Re: Does LSAT predict Bar Exam performance?
« Reply #8 on: April 23, 2005, 05:47:06 PM »
How could you be wrong 70 or 98% of the time, wouldn't you be right 50% by just geussing, that is if you are talking about bar passage on 1st try.

Esq

  • Sr. Citizen
  • ****
  • Posts: 109
    • View Profile
Re: Does LSAT predict Bar Exam performance?
« Reply #9 on: April 23, 2005, 06:00:56 PM »
The predictive value is what you'd be wrong about. In other words, and according to that single law school study, you wouldn't have LSAT as the strongest predictor variable. Say you had two individuals, both had a 152 LSAT. There are other variables that could make the prediction stronger. I'd like to know their LS GPAs, for example. But this is restricted to only LSAT, a single variable.

LSAT does have a strong relationship to bar performance in the aggregate. With a larger pool, you will see a relationship. Again, your argument is using aggregate statistics against individual statistics. The LSAC wrote that aggregate data show "that for one year, students from a school with an LSAT mean of 155 had an average bar passage rate of 72.5 percent. Given the fact there is a positive, although somewhat modest, relationship between LSAT scores and bar passage (again, 0.30 from the Bar Passage Study), it is no surprise that aggregate LSAT data have a strong relationship to aggregate bar performance. Aggregate data mask the existence and statistical effects of individual outliers--those whose performance on the bar doesn't fit well with their performance on the LSAT." Jim Vaseleck, Law School Admission Council, Letter to the Editor, The Bar Examiner, February 2005, p 42.