yankihote wrote: pacifica wrote: WhiteyCakes wrote:
ChampagnePapi wrote:The more people who fill it out, the more it looks like its for borderline applicants. I highlighted in green everyone who is at or above both medians. Everyone else is a URM, splitter, or reverse splitter.
Yeah that made me a little nervous
Naw, don't be nervous. If anything, I think this kinda shows there's no bias in the interviews right now. I don't proclaim to be an expert in this area (especially since I don't know adcom of Chicago's past history), but right now, spread sheet shows 7/21 (33%) above both medians. I mean, statistically speaking, being above both medians should be 1/2*1/2, or 25%, of the admitted pool right? So it's right around there. Most of the applicants got their things in early, but that's about the only trend. Hope that calmed nerves, if not, feel free to continue to freak out, that's always cool too.
There is no guarantee that 25% of applicants would be above both medians. If it works out that way for any given cycle, it is likely the the result of adcomms wanting it that way. Statistically it is only one possibility among many others that would result in any given set of median gpa and lsat scores.
Just to clarify: 1/2 of applicants are above one of the medians. By multiplying this 1/2 by 1/2 again, you are saying that 1/2 of all applicants that are above one median, say the LSAT median, must also be above the GPA median. This is the faulty assumption. The 1/2 of students above LSAT median could all have GPAs below median. This can work in the opposite as well. My point is that there is no guarantee that 25% are above both.
I agree with your clarification. I don't want to get into a detailed statistical argument here (mostly cus I'm very rusty), but what you pointed out is the other extreme of what I was saying. If we assumed LSAT and GPA are completely independent variables (i.e. everyone is a splitter and reverse-splitter), then it's 0.5*0.5 for probability of people being above both medians, or 25% for candidates are above both medians. If we assumed LSAT and GPA are completely dependent (i.e. there are no such thing as splitters), then its basically 50% for candidates above both medians because you're saying if you're above one, you're automatically above the other.
It's probably somewhere in between those two extremes, which means candidates above both medians will make up between 25-50% of the population, and the spread sheet seems to show that, since approximately 1/3 of the interviewees thus far are above both medians.
Which gets back to my original argument, that nobody needs to freak out that somehow this interview is selected for "borderline candidates", splitters, whatever you named, since it seems like the scores so far shows a random sampling.
[Edit: well, not anymore I guess, it's now 8/32? It was closer to 1/3 earlier during the day.]