Page 1 of 1

USNWR why do they use median instead of mean?

Posted: Thu Apr 21, 2016 3:39 pm
by Broncos15
Curious to know- obviously both methods have inherent flaws.

Re: USNWR why do they use median instead of mean?

Posted: Fri Apr 22, 2016 6:24 pm
by theconsigliere
Median is a more robust statistic, i.e. it is much less responsive to outliers. Imagine how much it would suck to be a splitter if schools reported their means instead of medians.

Re: USNWR why do they use median instead of mean?

Posted: Sat Apr 23, 2016 4:18 pm
by devilsadvocatetroll
theconsigliere wrote:Median is a more robust statistic, i.e. it is much less responsive to outliers. Imagine how much it would suck to be a splitter if schools reported their means instead of medians.
Er, depends on the splitter.
Assuming the school weights LSAT and GPA equally:
If his LSAT raises the average LSAT relatively higher than his GPA lowers the average GPA for the school in question...then average would help him.

Re: USNWR why do they use median instead of mean?

Posted: Sat Apr 23, 2016 11:12 pm
by banjo
Using the mean would punish schools for admitting members of groups that have traditionally scored lower on the LSAT.

Re: USNWR why do they use median instead of mean?

Posted: Sun Apr 24, 2016 12:00 am
by philepistemer
theconsigliere wrote:Median is a more robust statistic, i.e. it is much less responsive to outliers. Imagine how much it would suck to be a splitter if schools reported their means instead of medians.
People that score minimally below one stat and way above on another are hurt by using median instead of mean. Who would you expect to be a better law student: someone who was 0.05 above the GPA median and 1 point above the lsat median or someone who was 0.05 below the gpa median and 8 points above the lsat median?

Re: USNWR why do they use median instead of mean?

Posted: Sun Apr 24, 2016 10:55 am
by theconsigliere
philepistemer wrote:
theconsigliere wrote:Median is a more robust statistic, i.e. it is much less responsive to outliers. Imagine how much it would suck to be a splitter if schools reported their means instead of medians.
People that score minimally below one stat and way above on another are hurt by using median instead of mean. Who would you expect to be a better law student: someone who was 0.05 above the GPA median and 1 point above the lsat median or someone who was 0.05 below the gpa median and 8 points above the lsat median?
I guess my comment was for "super" splitters, like someone with below a 3.3 GPA or a low 160s LSAT for a T14. I think those people would be in a lot of trouble if mean was used because unless their other stat is a 180 or a 4+ then there's not much of a chance they're going to be able to pick up one median as much as they drop the other.

Post removed...

Posted: Sun Apr 24, 2016 11:43 am
by Barack O'Drama
Post removed...