MCAT score distribution statistics

[mujaman]

---

Members don't see this ad.

I don't think I've ever seen a thread come up specifically on this... if so, apologies.

As I was reviewing MCAT statistics today in an effort to convince myself that my second attempt score was going to be significantly better than my first (27), a few things stuck out to me about the score distribution:

(https://www.aamc.org/students/download/320554/data/combined12.pdf.pdf):

- The total score is a nice curve with a pretty normal distribution
- The section scores do not have nearly as nice a curve (especially the writing section, which I know is gone for 2013, but still... it's all over the place!)

So, statistics people, can anybody explain how the data presents like this? My understanding is that each section is scaled, and that the overall scale would simply be an outcome of this. It seems this can't be the case though...

 

[inycepoo]

The entire exam out of 45 is scaled to achieve a normal distribution across all test takers of the exam in all years. 0.01% 45, 0.1% 42, etc. The section scores are not scaled like that.

 

[mujaman]

The entire exam out of 45 is scaled to achieve a normal distribution across all test takers of the exam in all years. 0.01% 45, 0.1% 42, etc. The section scores are not scaled like that.

Right... that's pretty much what I said. But the point I was making is that I don't understand how the total score can be scaled so much differently than the section scores... i.e. how is there a meaningful correlation between the two 'scalings'?

 

[LostinLift]

Tangentially related but what I read and spoke with AAMC about is that the curve is pre-set, but if this is the case, even barring two weeks for any complaints about scores or re-grades, why are the scores not available sooner. And I would love to see some actual data on all test taker scores to see how well it lines up with their proposed distribution.

 

[inycepoo]

Right... that's pretty much what I said. But the point I was making is that I don't understand how the total score can be scaled so much differently than the section scores... i.e. how is there a meaningful correlation between the two 'scalings'?

It must mean that for every administration of the exam, they scale it according to the total score, but don't necessarily take do it the same way every time. Don't know another way to explain it.

If I knew more stats, I'd give you a better mathematical answer. Alas, I do not 😛

 

[mcloaf]

Tangentially related but what I read and spoke with AAMC about is that the curve is pre-set, but if this is the case, even barring two weeks for any complaints about scores or re-grades, why are the scores not available sooner. And I would love to see some actual data on all test taker scores to see how well it lines up with their proposed distribution.

The first topic is pretty searchable. IIRC it largely comes down to making sure there were no testing irregularities etc.

As for actual data, I'm a little confused. The PDF OP linked to is the actual testing data. Looks like a pretty nice curve to me.