Median vs. Mean in School Stats

[GreenCenter]

My impression is that the median tends to be higher than the mean in both GPA and MCAT for most schools. When figuring out which schools I should aim for, should I be looking at median or mean?

---

Members don't see this ad.

 

[3dee]

it's about the whole package!!

 

[mms2k7]

median. the mean is always misleading and takes into account outliers with either very special circumstances, that you probably dont fall under. The median is probably a much more accurate estimation of the #s at the school. It might be the whole package, but if you have a 3.7bcpm + a 30 MCAT i BET youll get in SOMEWHERE.

 

[ChristianJames1]

Definitely the median. 50% of people fall above the median, and 50% below.

 

[chickenlittle84]

But it is more complicated than that in this case. The median comes from MSAR and that includes all people who were accepted. So if a mid-tier schools accepts a lot of candidates who have a 4.0/40, that drives the median up. The mean comes from U.S. News and that includes only those students that matriculate...so it could be that the median is high because it excludes outliers, or it could be that the median is high because it includes all students that were accepted (at least for mid or lower tier schools). Similarly, the mean may be lower because it includes exceptions or because it includes only students that matriculate.

 

[mms2k7]

yah what chickenlittle said, i apologize, i thought you were referring to the median or mean of matriculants. Its different between acceptances/matriculants.

 

[TheRealMD]

And so we play the numbers game again. Yay!

Here's what I've learned so far:

1) Only make meaningful interpretations from schools that give you this criteria:

Average MCAT, Average scores in each section, Average cum GPA, Average science GPA.

One thing that I want to point out is that unless the school you are looking at has a rather small class, low outliners will not influence averages. People with low MCAT scores are an exception and should not adversely affect the average. Even if they did, then that means that they show a commitment to looking towards the entire applicant and not just numbers.

2) In my view, medians are misleading. They take the medians of each individual section, add them up, and that is the median. However, unless you're talking about a brilliant applicant, how often do people have truly balanced scores? This is why averages are still important.

That's all I can really say.

 

[Maigret Man]

Yeah. Definitely keep an eye on the number of "accepted" vs. "matric." students. Some schools accept four times more students who actually attend.

 

[uhohspaghettio]

if you have a 3.7bcpm + a 30 MCAT i BET youll get in SOMEWHERE.

hmm. are you sure about this? My friend had 3.65 and 34 but got no acceptances this year 😱

 

[uhohspaghettio]

he didnt have much good ECs though

 

[bunface]

just out of curiosity, what ECs did he have, do you know? my ECs aren't too fabulous either...

 

[uhohspaghettio]

just research work and limited clinical. Nothing that stand out like saving dead african babies or working with noble laureate. Plus, he applied late- like toward the end of July and ealry August

 

[leftmid]

So, I'd pay attention to the US News Mean and look at the MSAR Median only when applying to top 10-20 schools. Or just leave the Median up to those schools exclusively for their bragging right.

 

[Dr.D-man]

Definitely the median. 50% of people fall above the median, and 50% below.

Your 50/50 thing applies to the mean, not the median.

Mean = average number within a set of numbers
Median = the middle number within a set of numbers

 

[uhohspaghettio]

Your 50/50 thing applies to the mean, not the median.

Mean = average number within a set of numbers
Median = the middle number within a set of numbers

Well, if a school mean is 30, it usually indicates that the majority of students have scores around 29 - 31. Some might have 27- 28, or 32 -33. Other scores are probably outliers

 

[linuxizer]

Your 50/50 thing applies to the mean, not the median.

Mean = average number within a set of numbers
Median = the middle number within a set of numbers

He's right. Half the people have scores above the median; half have scores below. You got your definitions right, so I assume there was just some switch-up.

Ari

 

[Dr.D-man]

Your 50/50 thing applies to the mean, not the median.

Mean = average number within a set of numbers
Median = the middle number within a set of numbers

I looked it up on wikipedia and I can see where I was wrong.

I was thinking that the mean or average correlates to the 50th percentile, where 50% are better/worse. I now see that this is wrong. That only happens on a perfectly symmetrical distribution.

As for the median, I was thinking that this is calculated by (low number + high number)/2. I guess this is wrong. It appears that is a count of all the numbers in a set and the middle number is taken to be the median.

It has been a while since I took statistics...

 

[DocCallaway]

If you look at the curves in the msar, they are very skewed, therefore, the best fit statistic that represents the population is the median. Go with the median.

 

[GreenCenter]

I think Stats class is coming back to me:

3.1 3.1 3.1 3.7 3.8 3.8 3.8 Median: 3.7 Mean 3.49
3.6 3.6 3.6 3.7 3.8 4.0 4.0 Median: 3.7 Mean 3.76

The median w/o the mean tells you NOTHING!

Not nothing, but you get my point.