2019-2020 LSAT Administrations
This Interpretive Guide for the 2019-2020 LSAT was developed to help admission officers, deans, faculty, prelaw advisors, and others who use LSAT scores, to facilitate the law school admission process. This guide does not cover all the technical psychometric information available regarding the LSAT, but it does provide basic information in nontechnical language for those who need to use and interpret these scores.
Interpreting LSAT Scores
Scores on the Law School Admission Test (LSAT) are reported on a scale from 120 to 180 and can be directly compared across testing administrations and testing years. Why are scores on the 120-180 scale comparable to each other? These scores have the same meaning from one administration to the next and from one year to the next as a result of a statistical process called equating. When scores are equated, a given scaled score represents comparable ability regardless of the administration in which it is obtained.
The average ability level of test takers (group performance) is higher at some administrations than at other administrations. Nevertheless, for any individual test taker, a given scaled score represents the same degree of ability regardless of when the score is earned. Information about group performance may be useful to some score users. However, score users should never inflate or discount an individual’s score to take into account the administration at which it was earned, since the scores from different test forms have been made comparable through the equating process. An applicant’s LSAT score provides the same information about the applicant’s ability regardless of the ability of others who tested at the same time. For example, an applicant with an LSAT score of 160 might have a higher relative standing among February test takers than among June test takers, but the ability level represented by a score of 160 is the same regardless of when that score is earned.
Distribution of 2019-2020 Test Takers
To guide and monitor their admission processes, many LSAT score users use information about the percentage of test takers who earn each scaled score. This is frequently referred to as distribution data because it provides information about how test takers are distributed across the score scale. Distribution data based on information available from the 2018-2019 and 2019-2020 testing years are presented in Table 1.
Table 1 shows percentile ranks for 2018-2019 test takers compared to 2019-2020 test takers. The same score would have a different percentile rank within different groups. LSAT score users are likely to be most interested in percentile ranks for three different reference groups: all test takers, all law school applicants, and all applicants to a particular law school. This Interpretive Guide only shows percentile ranks for test takers to date. The monthly applicant reports sent to each law school show the percentile ranks for applicants to all ABA-approved law schools to date and for applicants to that law school. These percentile ranks for the reference group of applicants are different from the percentile ranks for all test takers.
Some Things to Note
- The test-taker volume for the 2019-2020 testing year was approximately 10.3 percent lower than reported for the 2018-2019 testing year. Note: Scores from the test takers who took the Spanish LSAT in Puerto Rico are included in the test-taker volume for both testing years.
- The mean LSAT score for the 2019-2020 testing year was 0.89 points higher than the mean LSAT score for the 2018-2019 testing year.
- Slight differences were noted between the percentile ranks for the two testing periods.
Reliability, Measurement Error, Score Bands, and Score Differences
To assess the reliability or consistency of LSAT scores, a reliability coefficient is computed for each LSAT form. Reliability coefficients indicate how reproducible a test taker’s performance would be over repeated administrations of the same test form. Reliability coefficients are measured on a scale from 0 to 1. The larger the value of the reliability coefficient, the more reproducible a test taker’s performance should be. Values of at least 0.9 indicate a very reliable test form.
Table 2 shows the reliability coefficient for the June 2019 (Form 0LSA131), July 2019 (Form 0LSA937 and 0LSA137), September 2019 (Form 9LSA135), October 2019 (Form 0LSA139), November 2019 (Form 0LSA141), January 2020 (Form 0LSA140), and February 2020 (Form 9LSA136) test forms. These and previous LSAT reliability coefficient values have typically been over 0.9, indicating that the LSAT is a very reliable test.
LSAT scores contain a certain amount of measurement error that is assessed with the standard error of measurement for individual scores (SEMI). The SEMI is more useful than the reliability coefficient for interpreting the precision of individual test scores. The SEMI indicates how close a test taker’s observed score is likely to be to his or her true score. (A test taker’s true score is the score that he or she would obtain on a perfectly reliable test — a test with a reliability coefficient of 1.) The LSAT SEMI is very stable and tends to be about 2.6 scaled score points. The actual SEMI value for Form 9LSA136 is 2.69 points (see Table 2). Smaller SEMI values indicate more precise scores.
Score bands, or ranges of scores that contain a test taker’s true score a certain percentage of the time, can be derived using the SEMI. Score bands are constructed by adding and subtracting a multiple of the SEMI to or from a scaled score. By adding and subtracting one times the SEMI to or from a score, the score band will contain an individual’s true score approximately 68 percent of the time. By adding and subtracting two times the SEMI to or from a score, the score band will contain an individual’s true score approximately 95 percent of the time.
For reporting purposes, LSAC constructs score bands by subtracting one times the rounded SEMI from the LSAT score to obtain a lower bound value, and adding one times the rounded SEMI to the LSAT score to obtain an upper bound value. LSAC adjusts the score bands for LSAT scores lying in the upper and lower regions of the LSAT score scale (i.e., scores close to 120 or 180), which makes them asymmetrical.
Table 1: Score Distributions for the 2018-2019 and 2019-2020 LSAT Administrations
|Score||2019-2020 Testing Year
|2018-2019 Testing Year
Note. Since Spanish LSAT scores are not reported on the 120-180 scale, score distributions from the test takers who took the Spanish LSAT in Puerto Rico are excluded from these results.
Given that the SEMI for Form 9LSA136 is 2.69 points (which we round to 3 to create score bands), the score band for most LSAT scores will be 7 score points. For example, the score band for an LSAT score of 150 will be 147 to 153.
LSAT score users are sometimes interested in comparing score differences among test takers. When this is done, users must keep in mind that the SEM for score differences (SEMD) is larger than the SEM associated with individual scores (SEMI). In fact, it is approximately 1.4 times larger. The interpretation of the SEMD is similar to the interpretation of the SEMI: the difference between scores from two test takers is within one SEMD on either side of the true score difference, approximately 68 percent of the time.
Table 2 shows the SEM associated with score differences for Form 9LSA136. For this form, the SEMD is 3.80 points, which we round up to 4 points to compare scores. If two test takers have scores of 150 and 154, for example, their true score difference will lie in the range of 0 to 8 points (4-point difference, plus or minus the rounded 4-point SEMD), approximately 68 percent of the time.
Note: This example illustrates that small score differences between two test takers may be due to measurement error and may not represent real differences in the abilities of test takers. This underscores the LSAC cautionary policy against putting undue weight on small score differences among test takers. The LSAT is just one source of information that should be considered when evaluating an applicant.
Table 2: Reliability and Standard Error of Measurement to Date: 2019-2020 Testing Year
|Standard Error of
|Score Differences (SEMD)|
|0LSA131 (June 2019 – Paper Mode)||0.94||2.62||3.70|
|0LSA937 (July 2019 – Paper Mode)||0.94||2.57||3.64|
|0LSA137 (July 2019 – Digital Mode)||0.94||2.57||3.63|
|9LSA135 (September 2019 – Digital Mode)||0.93||2.51||3.55|
|0LSA139 (October 2019 – Digital Mode)||0.93||2.50||3.54|
|0LSA141 (November 2019 – Digital Mode)||0.93||2.52||3.57|
|0LSA140 (January 2020 – Digital Mode)||0.92||2.57||3.63|
|9LSA136 (February 2020 – Digital Mode)||0.92||2.69||3.80|