you can find me at www.top-law-schools.com! Forum

[BillsFan9907]

Woot congrats sorcer!

Thanks! Pretty excited for the interview

http://www.top-law-schools.com/forums/v ... 2&t=200887

You might find this useful.

[ballcaps]

thanks for this. the thing i still don't really get is the acceptance rate after interviews.

50% would tell me the interview is a key assessment; 90% would tell me it's a rather large hoop to jump through, and 10% of applicants can't interview to save their lives.

but 70-80%? i just can't wrap my head around that...

Come again. Are you saying that because those who are interviewed get admitted at such a high rate, the the interview is really important?

i'm saying that 75% is a very strange acceptance rate after interviews, because it's unclear what that one person out of 4 (not 2, not 10, but 4) lacks.

make sense?

[BillsFan9907]

I think the most reasonable assumption is that close to 25% are auto waitlist and true true true borderline candidates.

In short those 25% didn't have

1) a real chance of being admitted outright to begin with

2) enough holistically to push them over the edge following the interview (borderline)

3) social grace as reflected in interview

[daedalus2309]

thanks for this. the thing i still don't really get is the acceptance rate after interviews.

50% would tell me the interview is a key assessment; 90% would tell me it's a rather large hoop to jump through, and 10% of applicants can't interview to save their lives.

but 70-80%? i just can't wrap my head around that...

Come again. Are you saying that because those who are interviewed get admitted at such a high rate, the the interview is really important?

i'm saying that 75% is a very strange acceptance rate after interviews, because it's unclear what that one person out of 4 (not 2, not 10, but 4) lacks.

make sense?

That 25% rejection rate also captures those who were rejected via faculty review (which takes place post interview), so it's hard to scientifically quantify interviews.

[BillsFan9907]

Yes but its more reasonable than not that what Is an autowaitlist to adcoms will carry over to faculty; they will generally be on the same page even if independent from each other. We can predict with near certainty who will be admitted and who won't based on numbers and perhaps work experience.

[GordonFreeman]

We've reached a consensus: nobody but the HLS adcom understands the significance of the interview

[RSN]

We can predict with near certainty who will be admitted and who won't based on numbers and perhaps work experience.

Really? Care to elaborate?

[BillsFan9907]

We can predict with near certainty who will be admitted and who won't based on numbers and perhaps work experience.

Really? Care to elaborate?

Look at the graph on law school numbers. Where we see a yellow surrounded by green, almost always we will see something like k-jd in their profile if they disclose that stuff.

I'm not a statistician , but obviously you can see a clear box where people are auto admits, a periphery of auto waitlists and then an outer shell of rejections. Outliers in the rejection part (admitted in a sea of red) are generally URMs.

I wonder if there is a way to bleed the colors together so you get a clear coherent picture. In other words, where there is a lot of red and green the gradient would be orange and not dots. I guess like a color wheel.

http://harvard.lawschoolnumbers.com/stats/1415

[ElizabethKB]

I'm curious if there's any correlation between time between complete and JS1 and likelihood of acceptance. I'd imagine a quicker JS1 turnaround is a better sign. Anybody know if there's anything concrete on this?

Okay so I was curious about this, and did a very basic, probably flawed calculation to come up with some numbers (data from the applicant spreadsheet last year).

Keep in mind there are literally only 5 data points of people who were Denied last year after receiving interviews on the applicant spreadsheet after filtering out URMs, those with Decision Pending, and those with no interview, etc. There is no ways the following should be taken as statistical evidence, it's more just for curiosity sake.

The average amount of days it took from Complete -> JS1 for Denied applicants: 20.6
The average amount of days it took from Complete -> JS1 for Accepted applicants: 14.0

Again I want to reiterate that these numbers are meaningless. The cycle is different this year, and there are so few data points that the variance makes these numbers useless. However, it seems like there could be a correlation between the time of complete -> JS1 and rate of acceptance.

ETA: Waitlist data as per Kobaine's suggestion:
The average amount of days it took from Complete -> JS1 for Denied/WL applicants: 16.5

Thanks a ton for this Pylon! Also to the question about purpose of interviews, I talked to one HLS 2L who theorizes that beyond borderline people, the interview is largely to make sure HLS students will be competent in job interviews down the line (to protect employment stats). Just a speculation but makes sense.

[LSCHI]

Now when we are talking "borderline" what are we saying? Barely at both medians or what?

[xylocarp]

We can predict with near certainty who will be admitted and who won't based on numbers and perhaps work experience.

Really? Care to elaborate?

Look at the graph on law school numbers. Where we see a yellow surrounded by green, almost always we will see something like k-jd in their profile if they disclose that stuff.

I'm not a statistician , but obviously you can see a clear box where people are auto admits, a periphery of auto waitlists and then an outer shell of rejections. Outliers in the rejection part (admitted in a sea of red) are generally URMs.

I wonder if there is a way to bleed the colors together so you get a clear coherent picture. In other words, where there is a lot of red and green the gradient would be orange and not dots. I guess like a color wheel.

http://harvard.lawschoolnumbers.com/stats/1415

myLSN has a feature like that: http://mylsn.info/r/pre-law/admissions/graph/

[BillsFan9907]

Now when we are talking "borderline" what are we saying? Barely at both medians or what?

Or maybe barely below. Check out the graph. The top right corner is almost all green. As you expand you get more yellow and finally red.

[fra]

Now when we are talking "borderline" what are we saying? Barely at both medians or what?

I believe that borderline would be any point on the LSAT/GPA graph where there is a lot of yellow (waitlists). So some splitters would be borderline or below borderline (in a sea of red) even though they are above one 75%.

[BillsFan9907]

Xylocarp - u rock!!!!

[BillsFan9907]

Now when we are talking "borderline" what are we saying? Barely at both medians or what?

I believe that borderline would be any point on the LSAT/GPA graph where there is a lot of yellow (waitlists). So some splitters would be borderline or below borderline (in a sea of red) even though they are above one 75%.

Yup and that's exactly what we see!

[psychmusic]

.

[LSCHI]

Now when we are talking "borderline" what are we saying? Barely at both medians or what?

I believe that borderline would be any point on the LSAT/GPA graph where there is a lot of yellow (waitlists). So some splitters would be borderline or below borderline (in a sea of red) even though they are above one 75%.

Yup and that's exactly what we see!

Suddenly feel a lot less confident at both medians...

[hillz]

I just read through a few articles on predicting binary outcomes in logistic regressions and I think I burst a blood vessel in my eye.

I would be willing to bet that for 95% of applicants, whether or not a JS1 is received can be predicted by LSAT and GPA. Those 5% are the special snowflakes whose statement/resume was compelling enough to overcome their stats. I imagine there are almost no very high stats (close to 180/4.0) who DON'T get a JS1.

Post JS1, I would arbitrarily guess that acceptances are based 40% on LSAT, 30% on GPA, 12% on resume, 10% on PS, 6% on interview, and 2% on faculty review. In my reading of the faculty review blog post, it's really more to approve/disapprove of a decision that the adcom has already made. This is just my absolutely unscientific/random assignment of numbers for the typical applicant. This wouldn't account for the random person who founded an important charity/olympic athlete who had a 3.0.

I agree with your take on faculty review - it seems more like an "approve/disapprove of what we've pretty much already decided" kind of thing and faculty have nowhere near as much say as they do at Y.

[pylon]

Post JS1, I would arbitrarily guess that acceptances are based 40% on LSAT, 30% on GPA, 12% on resume, 10% on PS, 6% on interview, and 2% on faculty review.

Lol. I'd say 9% on PS and 7% on interview, but other than that I agree.

[pylon]

Suddenly feel a lot less confident at both medians...

I could be reading the graph wrong but at both medians is still above the yellow and into the green territory (173, 3.87).

[steakout319]

Suddenly feel a lot less confident at both medians...

I could be reading the graph wrong but at both medians is still above the yellow and into the green territory (173, 3.87).

Not to get anybody's hopes up needlessly but there have been several posts talking about how LSN numbers are slightly skewed...for whatever that is worth...

[RSN]

Suddenly feel a lot less confident at both medians...

I could be reading the graph wrong but at both medians is still above the yellow and into the green territory (173, 3.87).

Not to get anybody's hopes up needlessly but there have been several posts talking about how LSN numbers are slightly skewed...for whatever that is worth...

In the senses that the pool of people who submit to LSN (and post on TLS) are more likely to have higher stats, and the fact that people are less likely to post waitlists/rejections than they are acceptances? Or in some other way?

[antiworldly]

Suddenly feel a lot less confident at both medians...

I could be reading the graph wrong but at both medians is still above the yellow and into the green territory (173, 3.87).

Not to get anybody's hopes up needlessly but there have been several posts talking about how LSN numbers are slightly skewed...for whatever that is worth...

In the senses that the pool of people who submit to LSN (and post on TLS) are more likely to have higher stats, and the fact that people are less likely to post waitlists/rejections than they are acceptances? Or in some other way?

Both, there's a ton of selection pressure on LSN. Whereas it's supremely useful as a general predictor, it's hit an miss in the grey (or in this case yellow) areas.

[BillsFan9907]

Doesn't LSN post their medians? At any rate, under our current examination we are really looking st a gradient. So once we start to see a ton of yellow in a certain place and a tom of red, we can reasonably conclude that it will continue to be that Color until hitting another color. Selection bias should not affect this

[pylon]

Not to get anybody's hopes up needlessly but there have been several posts talking about how LSN numbers are slightly skewed...for whatever that is worth...

In the senses that the pool of people who submit to LSN (and post on TLS) are more likely to have higher stats, and the fact that people are less likely to post waitlists/rejections than they are acceptances? Or in some other way?

Both, there's a ton of selection pressure on LSN. Whereas it's supremely useful as a general predictor, it's hit an miss in the grey (or in this case yellow) areas.

I don't think anyone is saying it's perfect, but for a process that is as closed-door as law school admissions, it's an extremely useful tool and has value as a guideline.