# MSAR data sets?

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

#### sanche60

##### Senior Member
10+ Year Member
5+ Year Member
15+ Year Member
Does the MSAR give breakdowns of mcat and gpa by school or does it only give national statistics? I know it gives average mcat and gpa for each school but I was wondering if it gave numbers like percentages of individuals with less then 30, at 30 , more then 30, etc, for each school specifically.

Members don't see this ad.

sanche60 said:
Does the MSAR give breakdowns of mcat and gpa by school or does it only give national statistics? I know it gives average mcat and gpa for each school but I was wondering if it gave numbers like percentages of individuals with less then 30, at 30 , more then 30, etc, for each school specifically.

No, it doesn't give that kind of breakdown. It gives averages (or means, I don't really care about the difference if there is one in this case) for each MCAT subsection, the range of each subsection, and the GPA.

So it might say, for example:

University of Alabama
Overall GPA median: 3.76
Science GPA median: 3.72
VR: median 10, ranging from 6-14
PS: median 10, ranging from 6-14
BS: median 10, ranging from 8-15
WS: median P, ranging from K-T

Hope that helps.

exactly what i needed to know. thanks moose.

Members don't see this ad :)
Since the MSAR uses the median instead of the mean for their GPA and MCAT, do you think that the statistics for each school are skewed a little? If so, do think they are skewed higher or lower?

A mean will tend to skew data upwards (compared to a median), while a median will tend to skew it downwards (compared to a mean). This does not work for all distributions, though. For example, the mean and the median MCAT score (out of all tests taken) should be ~ equivalent (if you look at the distribution, you'll see why: it's virtually symmetrical). However, most things in life (someone slap me if you disagree) tend towards a more Boltzmann distribution. That is, there are more people at the low end of the scale than at the high end, so your median will usually be lower than your mean.

Sorry, I think I got way too into that.

The Boltzmann distribution is correct in some instances, but there's a reason why statisticians call the Gaussian curve the "normal" curve. Most things in life (such as experimental data) end up being like a normal curve, with an equal number of data points above and below the mean.

On another note, the median tends to be a better indicator of the "middle" of a distribution as compared to the mean, for the median is resistant to outliers. The mean can easily be skewed by a few high or low scores, while the median would still reflect the "true" middle of the distribution.