Where the Curve Came From:
In 1991, Law Services made up a new LSAT curve on a scale of 120 to 180 (instead of the traditional 200-800, or the previous 10-48). They assigned percentiles to each score on the curve based on anticipated results.
In 1992, they assigned to each test-taker the percentile actually achieved in 1991, which wasn't quite the same.
In 1993, they assigned each test-taker the actual percentile achieved by the test-takers in the previous two years, combined.
In 1994, they assigned each test-taker the actual percentile achieved by the test-takers in the previous three years, combined.
They've done the same thing ever since -- based your reported percentile on the three years' test-takers before you.
How the Curve Has Changed:
Over the last dozen years, the assigned percentile for a particular score
has dropped. That means that more people get a higher score than used
to. Here's the data for three different years:
|
174 |
172 |
170 |
168 |
166 |
164 |
162 |
160 |
158 |
156 |
154 |
152 |
150 |
148 |
146 |
144 |
1992 |
99.6 |
99.1 |
98.4 |
97.2 |
95.2 |
92.5 |
88.9 |
83.9 |
78.3 |
70.4 |
63.3 |
54.8 |
45.9 |
37.9 |
29.4 |
23.8 |
2001 |
99.5 |
99.1 |
98.2 |
97.0 |
94.8 |
91.9 |
88.2 |
83.1 |
77.4 |
70.7 |
63.3 |
55.2 |
47.3 |
39.3 |
32.2 |
25.5 |
2004 |
99.5 |
99.0 |
98.1 |
96.7 |
94.6 |
91.4 |
87.3 |
82.2 |
76.5 |
68.7 |
61.5 |
53.2 |
44.9 |
37.0 |
29.6 |
23.3 |
So in 1992, 97.2% of test-takers, or 972 out of a thousand, had a score of 167 or lower, and 28 had a score of 168 or better. But in 2001, 96.7% of test-takers, or 967 out of a thousand, had a score of 167 or lower, and 33 had a score of 168 or better.
That may not seem like a big deal, but out of 150,000 test takers, that's 150 x 33 people, or 4950 instead of 4200. Seven hundred and fifty more people had a score of 168 or better. Top law schools had 750 more people to choose from!
As I explained above, the reported percentile is based on a three-year average. So in a year when the curve moves, there are three times as many people at the new, lower percentile.
Okay, here's an example.
Suppose that for 2001, 2002, and 2003, 100,000 test-takers took the LSAT, and 3% of them got a 168 or better (as was reported for 2001). That's 30 out of 1,000, or 3,000 out of 100,000.
Are you with me so far? Good. Now in 2004, it's reported that for the last three years, 3.3% of test-takers had a 168 or better. That's 3 more test-takers per thousand, or 300 per 100,000. But all the three year change happened this year. So this year, 900 more people got that score!
In fact, this year (2004-2005) there were 145,000 test-takers. So about 1300 more people had an LSAT score of 168 or better than did last year!
This example is greatly simplified. SInce the actual number of test-takers changes from year to year, we need to adjust each year's percentile by the actual number of test-takers for the last three years. But there hasn't been that much difference in this time period -- the number of tests has dropped from 148,00 to 147,000 to 145,000. So maybe there are only 1200 more people with a 168, instead of 1300. Georgetown has only 1200 more people with an LSAT score at last year's median to choose from; does that make you feel better?
The skewed LSAT curve means that over the long haul, more and more people have higher LSAT scores than you. How many more varies with the year and with your score, but rest assured that it's more.
Note: All the data on this page was actually published by Law Services; there are no numerical assumptions, only simplifications.