Rank Estimator Question Forum

[Germaine]

This is probably a pretty basic question about how the rank estimator at LSN works, but for some reason I can't get my head around it.

Okay, so suppose I'm at a school where 30% of the 1L grades in each class have to be either A-, A, or A+. And the mean for each class is supposed to be in the 3.0 - 3.3 range. I don't know any other rank percentiles.

If I'm using the rank estimator, would it make sense to put something like a 3.8 for the top 30%? I get that may not be accurate, but theoretically, if the same 30% from each class were all the A-range students, woudln't that mean that something in the 3.8 range was probably top 30%?

Sorry if this is way off base...maybe it's just late but I can't seem to think through the math of this.

[Tiago Splitter]

Yes you are way off base. The estimator wants you to get close to the exact percentile cutoff, and top 30% is way too high. The 50th percentile cutoff is something like 3.15, and I promise you very few students have a 3.8. If I had to guess I'd put 10% at around 3.7.

[Germaine]

Yes you are way off base. The estimator wants you to get close to the exact percentile cutoff, and top 30% is way too high. The 50th percentile cutoff is something like 3.15, and I promise you very few students have a 3.8. If I had to guess I'd put 10% at around 3.7.

So basically even though 30% in any given class have to get in the A range, you say 10% have an A-range GPA. And that's b/c it's just unlikely for the same 30% (or anything close to it) to be in the A range in each class? (Like I said, I know I'm probably asking a really stupid question so I appreciate the patience.)

[Tiago Splitter]

Yes you are way off base. The estimator wants you to get close to the exact percentile cutoff, and top 30% is way too high. The 50th percentile cutoff is something like 3.15, and I promise you very few students have a 3.8. If I had to guess I'd put 10% at around 3.7.

So basically even though 30% in any given class have to get in the A range, you say 10% have an A-range GPA. And that's b/c it's just unlikely for the same 30% (or anything close to it) to be in the A range in each class? (Like I said, I know I'm probably asking a really stupid question so I appreciate the patience.)

Even if the same people get the exact same grades in every class, the 30% cutoff would be the last A-, which is 3.67. The reality, as you well know (from your previous post history I can see you booked three classes and got a B in a fourth) is that there is a mix of grades among the class, so grades should fall in a relatively normal distribution around the median. The median is around 3.1-3.2. I have no idea why you think the 30% cutoff would so far away from that number, especially given that you also admitted in your post history that a professor gave you an A+ despite having not given one in seven years, suggesting the A+ is rarely used.

[Germaine]

Thanks. I guess I was assuming that my situation was sort of an anomaly -- and that most people had grades that were like A, A, A, A-...or B, B, B, B+. So basically I just figured that the estimator was assuming a normal distribution where the distribution wasn't actually normal. But I see what you're saying and it makes sense.