Looking At LSN, It Is Obvious That Lines Are Being Drawn & Forum

[nonsequitur]

Second, LSN is not just a self-selecting sample of applicants, but a highly skewed sample. I noticed that you have similar numbers of users on the site in the 150-160 range (the mode should be somewhere in this set) as you do in the 170-180 range (top 2-3 percentile). If you look at just about any school in the middle of the top 100 the averages of matriculated students in no way match the averages on LSN. The only way you can really know what's happening is to check the profile of the 2010 entering class when that information is published, or to have an inside line to the admissions office.

While true, this is almost completely irrelevant when considering LSN's effectiveness as an admissions tool. That person after person seems to miss this fact amazes me.

I wasn't saying that it's not useful as an admissions tool, I was trying to point out the flaw in using that data to extrapolate the school's acceptance/rejection cut-off point. I thought that's what the OP was about.

[D. H2Oman]

Second, LSN is not just a self-selecting sample of applicants, but a highly skewed sample. I noticed that you have similar numbers of users on the site in the 150-160 range (the mode should be somewhere in this set) as you do in the 170-180 range (top 2-3 percentile). If you look at just about any school in the middle of the top 100 the averages of matriculated students in no way match the averages on LSN. The only way you can really know what's happening is to check the profile of the 2010 entering class when that information is published, or to have an inside line to the admissions office.

While true, this is almost completely irrelevant when considering LSN's effectiveness as an admissions tool. That person after person seems to miss this fact amazes me.

I wasn't saying that it's not useful as an admissions tool, I was trying to point out the flaw in using that data to extrapolate the school's acceptance/rejection cut-off point. I thought that's what the OP was about.

The data can absoluetly be used for that.

[qualster]

Second, LSN is not just a self-selecting sample of applicants, but a highly skewed sample. I noticed that you have similar numbers of users on the site in the 150-160 range (the mode should be somewhere in this set) as you do in the 170-180 range (top 2-3 percentile). If you look at just about any school in the middle of the top 100 the averages of matriculated students in no way match the averages on LSN. The only way you can really know what's happening is to check the profile of the 2010 entering class when that information is published, or to have an inside line to the admissions office.

While true, this is almost completely irrelevant when considering LSN's effectiveness as an admissions tool. That person after person seems to miss this fact amazes me.

I wasn't saying that it's not useful as an admissions tool, I was trying to point out the flaw in using that data to extrapolate the school's acceptance/rejection cut-off point. I thought that's what the OP was about.

The data can absoluetly be used for that.

Yes, it can. Again, see Kentucky: Last year, the obvious cut off, according to the LSN graph, was 159. That was their median. This year their line has moved to an ambitions 162. Weird, wacky shtuff.

[09042014]

Second, LSN is not just a self-selecting sample of applicants, but a highly skewed sample. I noticed that you have similar numbers of users on the site in the 150-160 range (the mode should be somewhere in this set) as you do in the 170-180 range (top 2-3 percentile). If you look at just about any school in the middle of the top 100 the averages of matriculated students in no way match the averages on LSN. The only way you can really know what's happening is to check the profile of the 2010 entering class when that information is published, or to have an inside line to the admissions office.

While true, this is almost completely irrelevant when considering LSN's effectiveness as an admissions tool. That person after person seems to miss this fact amazes me.

I wasn't saying that it's not useful as an admissions tool, I was trying to point out the flaw in using that data to extrapolate the school's acceptance/rejection cut-off point. I thought that's what the OP was about.

The data can absoluetly be used for that.

Dude just got pwnd in stats by an education major with a 2.8. He should seriously consider suicide.

[nonsequitur]

Ok, looking at Kentucky's statistics it is fairly representative on the surface, but SSS makes for problems.

On LSN University of Kentucky's 2009 (the school in question) LSAT stats are as follows:

Applied - 1255
Admitted - 384
Matriculated - 138

75 - 162
Median - 159
25 - 155


LSN sample size = 72 total, 43 admitted (21 of whom withdrew)
If all the remaining students attend, that leaves 22.

The LSN sample shows 11 people who were accepted below the median, and 2 people accepted who were below the 25th percentile. 13 below the median is pretty close to representative. It's when you look at the specific numbers that some holes start to appear. There are zero acceptances with a 155, or 154 LSAT, and the two below the 25th percentile are both 153s. There's one 156, which would appear to be the bottom end of the scale if you exclude the 153s as outliers (i.e. special circumstances affected admissions). It would appear from a quick look at the data that only students with a 156 or better are accepted. That just can't be true given the published admission statistics of the school.

Plus when this skewed small sample size is put together with my first point that many of the acceptances given to people with lower LSAT scores are likely later in the cycle it seems just about impossible to use the data that are currently available for this cycle to identify the school's cut-off LSAT. People with perfectly good scores are rejected from schools for other reasons, so the fact that some people with X score have been rejected doesn't necessarily mean anything.