Curve Analysis Forum

[Nulli Secundus]

1 Jun 91 -14
2 Oct 91 -10
3 Dec 91 -11
4 Feb 92 -10
5 Jun 92 -15
6 Oct 92 -11
18 Dec 92 -13
7 Feb 93 -15
8 Jun 93 -12
9 Oct 93 -15
10 Feb 94 -11
11 Jun 94 -11
12 Oct 94 -9
13 Dec 94 -14
14 Feb 95 -10
15 Jun 95 -9
16 Sep 95 -14
17 Dec 95 -15
19 Jun 96 -12
20 Oct 96 -12
21 Dec 96 -14
22 Jun 97 -12
23 Oct 97 -16
24 Dec 97 -14
25 Jun 98 -12
26 Sep 98 -14
27 Dec 98 -16
28 Jun 99 -13
29 Oct 99 -12
30 Dec 99 -14
31 Jun 00 -14
32 Oct 00 -14
33 Dec 00 -14
34 Jun 01 -13
35 Oct 01 -12
36 Dec 01 -12
37 Jun 02 -10
38 Oct 02 -12
39 Dec 02 -13
40 Jun 03 -10
41 Oct 03 -12
42 Dec 03 -11
43 Jun 04 -12
44 Oct 04 -10
45 Dec 04 -12
46 Jun 05 -9
47 Oct 05 -10
48 Dec 05 -8
49 Jun 06 -10
50 Sep 06 -10
51 Dec 06 -11
Free June 07 -8
52 Sep 07 -10
53 Dec 07 -11
54 Jun 08 -10
55 Oct 08 -9
56 Dec 08 -11
57 Jun 09 -11
58 Sep 09 -11
59 Dec 09 -14
60 Jun 10 -12

So these are the curves for all PTs, here are some averages:

Cumulative Avg: -11,9016

June Tests Avg: -11,45
Sep/Oct Avg: -11,7368
Dec Avg: -12,6667

In the 18 years with full sets of Jun -> Sep / Oct -> Dec tests,

From June to Oct, curve got better in 8 years, got worse in 7 and did not change in 3 years. (where better means more lenient)
From Oct to Dec, curve got better in 10 years, got worse in 4 and did not change in 4 years.

For the last 8 years, from June to Oct, curve got better 1 year and got worse the next, last year it got worse by a point. (It by no means should be seen as indicative of a better curve, but just pointing it out)

For the last 4 years, from Oct to Dec, curve got better consistently.

TL;DR: Statistics show that it is better to take the test in December, too bad you end up being too late when you do so.

[Adjudicator]

TL;DR: Statistics show that it is better to take the test in December, too bad you end up being too late when you do so.

This argument follows logically if which of the following is assumed?

A) The December test is not typically more difficult than tests at other times of the year.

[whymeohgodno]

October 10 one better be a -15 curve

[Adjudicator]

October 10 one better be a -15 curve

Amen to that!

[LSAT Blog]

I've done something similar, in case anyone's interested in seeing the averages by month specifically for the past 10 years.

[Blindc1rca]

dude this is why rather than aim for a 170 or a 172 or whatever I just aim for a -10 or better.

[MissLucky]

damn the june 07 ppl got NAILED. in every way...new passage structure AND one of the worst curves in modern LSAT history?

[Hedwig]

Anything below -10 just seems stingy to me. But yeah, I focus on a goal of a certain "-#" per section, rather than trying to aim for a certain score, because the certain -# per section puts you in the general category of scores you might like.

[northwood]

i dont understand why people worry about "their score" when doing prep tests. for me I worry about the actual number of questions I answered correctly. I aim to increase my raw score each time i take the test. I use the score to help indciate what I would have earned if the test counted for real, but i pay more attention to the raw score.

If you want to use a curve, why not use the stingest one. That way you will most likely see a nice bump on test day, if you score the same number correct.

[St.Remy]

i dont understand why people worry about "their score" when doing prep tests. for me I worry about the actual number of questions I answered correctly. I aim to increase my raw score each time i take the test. I use the score to help indciate what I would have earned if the test counted for real, but i pay more attention to the raw score.

If you want to use a curve, why not use the stingest one. That way you will most likely see a nice bump on test day, if you score the same number correct.

It seems perfectly logical to worry about your score, considering that at the end of the day the score is the important thing, not the number you missed. Aiming to increase your raw score each time is fine, but stripping the curve from the test obfuscates the fact that some tests are indeed harder than others and were intended as such. Using the stingiest curve would most likely give you an oscillating score that wouldn't be of any help to you in figuring out where you stand for test day purposes.

[LSAT Blog]

If you want to use a curve, why not use the stingest one. That way you will most likely see a nice bump on test day, if you score the same number correct.

Because it would likely be discouraging. You don't want to beat yourself up.

[09042014]

TL;DR: Statistics show that it is better to take the test in December, too bad you end up being too late when you do so.

This argument follows logically if which of the following is assumed?

A) The December test is not typically more difficult than tests at other times of the year.