The New MCAT & The Drop in Admit Rate

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

efle

not an elf
10+ Year Member
Joined
Apr 6, 2014
Messages
14,146
Reaction score
22,781
I've talked about this theory in some threads but finally plotted it out of curiosity and thought I'd share the visual.

Back when the AAMC tables updated with data for the new MCAT (MCAT2015), there was something very odd:

The admit rate for people with the MCAT2015 was only 32.0%, while the overall admit rate for the cycle had been 41.5%.

Crazy, right? What was going on, were adcoms subconsciously biased against the unfamiliar new exam?

Well, it doesn't look like that is true, because in comparable stats bins the admit rates stayed the same - for example on the new grid, 3.9/518+ has an 87.5% admit rate, while on the old grid the same percentile group (3.9/36+) had an 87.9% admit rate. Similarly, all GPAs with 518+ had a 79.2% admit rate, while all GPAs with 36+ had a 79.5% admit rate. This kind of similarity got a little rougher in other bins, since the percentiles on MCAT types were not exactly equivalent, but overall the pattern held up with very similar success rates for similar scores.

So why the drastically lower admit rate with the new MCAT? It looks like it was because people on the lower end of the curve were much more likely to apply now than they used to be under the old exam.

Here are the applicant distributions for the most recent cycle and the composite of the three prior cycles, with proportion admitted also shown:

Edit: Note that the far right are equivalent percentile (36+ = 518+) and the far left are approx. equivalent (486 = 17). See this conversion table. So for both graphics you are looking at the same percentile ranges across the same sized axis, the bins are just different.

CHWJfDm.png


So there you have it - you got a whole lot more pink going on for the bottom distribution's middle and left sections. Mystery solved, in my opinion - either because people didn't understand how their scores compared to MSAR ranges, or because they expected schools to change their behavior with the new exam, many more people with average and below average numbers went ahead and applied than we saw historically.

Also: Important to note GPA distributions did not change much despite the overall left shift in the applicant score distribution. For example among all applicants with a 3.00+ GPA, each bin of 0.2 contained the following percentages:

9TEfxuy.png

Very similar.

Members don't see this ad.
 
Last edited:
  • Like
Reactions: 10 users
tw1xOjL.png


tDNNOJm.png

tjLBtQd.png

fc2fca5cc8833d35c92822a6af386914480a4e7c2411a6329d9742eb6bc89110.jpg
 
  • Like
Reactions: 34 users
Members don't see this ad :)
Also an addendum that it's also not really the applicants' fault entirely for being misled about their chances: the new MCAT repeatedly stressing something along the lines that "the center score of 500 is the ideal/curved grade for med school success" isn't helping matters much either. Since unless you've got some extenuating circumstances a 500 is not gonna happen, but the test promotional materials suggest that it's fine.
 
  • Like
Reactions: 6 users
Also an addendum that it's also not really the applicants' fault entirely for being misled about their chances: the new MCAT repeatedly stressing something along the lines that "500 is the ideal/curved score for med school success" isn't helping matters much either. Since unless you've got some extenuating circumstances a 500 is not gonna happen, but the test promotional materials suggest that it's fine.
Definitely, and it didn't help that they updated the MSAR to show applicant ranges for the new MCAT, whereas it had always been admitted applicant ranges before. Looking at WashU and seeing the median applicant had a 508, IQR 500-515, is going to mislead a ton of people.
 
  • Like
Reactions: 4 users
Who are these people with 518+ not getting accepted? This boggles my mind.
 
Who are these people with 518+ not getting accepted? This boggles my mind.
I imagine very low GPAs, bad LoRs, bad essays, bad interviews, etc.

There are even people in the 3.9/39+ bin from the old tables that get rejected, about 10% failed to get in anywhere in fact.

When people say numbers aren't everything, it's true! You have to play the game in its entirety and check other boxes too
 
I imagine very low GPAs, bad LoRs, bad essays, bad interviews, etc.

There are even people in the 3.9/39+ bin from the old tables that get rejected, about 10% failed to get in anywhere in fact.

When people say numbers aren't everything, it's true! You have to play the game in its entirety and check other boxes too
I guess this is true. This chart makes it look like its a decent proportion of them though. Maybe closer to 20%. Wow. Great visual!
 
Who are these people with 518+ not getting accepted? This boggles my mind.

I have a 517 and received 0 II's out of 20+ applications outside of my state. My stats are well within the range or above a lot of the schools I was outright rejected from

My guess is that schools prefer to interview candidates that have a good chance of going to their school - they're not gonna waste an interview on someone who is most likely going to take the in-state tuition from his home state

So yeah i can see how people with high stats (518+) can end up with 0 acceptances if they don't play their cards right





Sent from my iPhone using SDN mobile
 
  • Like
Reactions: 2 users
In all seriousness, you should submit this data to one of the medical education journals. I think this is publishable!


I've talked about this theory in some threads but finally plotted it out of curiosity and thought I'd share the visual.

Back when the AAMC tables updated with data for the new MCAT (MCAT2015), there was something very odd:

The admit rate for people with the MCAT2015 was only 32.0%, while the overall admit rate for the cycle had been 41.5%.

Crazy, right? What was going on, were adcoms subconsciously biased against the unfamiliar new exam?

Well, it doesn't look like that is true, because in comparable stats bins the admit rates stayed the same - for example on the new grid, 3.9/518+ has an 87.5% admit rate, while on the old grid the same percentile group (3.9/36+) had an 87.9% admit rate. Similarly, all GPAs with 518+ had a 79.2% admit rate, while all GPAs with 36+ had a 79.5% admit rate. This kind of similarity got a little rougher in other bins, since the percentiles on MCAT types were not exactly equivalent, but overall the pattern held up with very similar success rates for similar scores.

So why the drastically lower admit rate with the new MCAT? It looks like it was because people on the lower end of the curve were much more likely to apply now than they used to be under the old exam.

Here are the applicant distributions for the most recent cycle and the composite of the three prior cycles, with proportion admitted also shown:

CHWJfDm.png


So there you have it - you got a whole lot more pink going on for the bottom distribution's middle and left sections. Mystery solved, in my opinion - either because people didn't understand how their scores compared to MSAR ranges, or because they expected schools to change their behavior with the new exam, many more people with average and below average numbers went ahead and applied than we saw historically.

Also: Important to note GPA distributions did not change much despite the overall left shift in the applicant score distribution. For example among all applicants with a 3.00+ GPA, each bin of 0.2 contained the following percentages:

9TEfxuy.png

Very similar.
 
  • Like
Reactions: 1 user
I have a 517 and received 0 II's out of 20+ applications outside of my state. My stats are well within the range or above a lot of the schools I was outright rejected from

My guess is that schools prefer to interview candidates that have a good chance of going to their school - they're not gonna waste an interview on someone who is most likely going to take the in-state tuition from his home state

So yeah i can see how people with high stats (518+) can end up with 0 acceptances if they don't play their cards right





Sent from my iPhone using SDN mobile
This process is so strange. I don't understand why med schools think I need a tie to them to want to go there. They could be my first choice and they'll never know.
 
I have a 517 and received 0 II's out of 20+ applications outside of my state. My stats are well within the range or above a lot of the schools I was outright rejected from

My guess is that schools prefer to interview candidates that have a good chance of going to their school - they're not gonna waste an interview on someone who is most likely going to take the in-state tuition from his home state

So yeah i can see how people with high stats (518+) can end up with 0 acceptances if they don't play their cards right





Sent from my iPhone using SDN mobile
My impression is this kind of thing is an especially common experience for Texan residents
 
This process is so strange. I don't understand why med schools think I need a tie to them to want to go there. They could be my first choice and they'll never know.

They only have a limited amount of interviews - 400 (or so) out of 10,000+ applicants for some schools.

Giving one of those interviews to someone who is most likely going to turn them down is a total waste on their part - they get nothing out of it other than a wasted interview spot.





Sent from my iPhone using SDN mobile
 
Last edited:
  • Like
Reactions: 1 user
Members don't see this ad :)
My impression is this kind of thing is an especially common experience for Texan residents

I'm sure it's true for a lot of applicants from states with several state schools with in-state bias.

It makes sense too - why would an applicant choose out of state med school A if they have relatively comparable in-state med school B right in their backyard for half the price?


Sent from my iPhone using SDN mobile
 
  • Like
Reactions: 1 user
I'm sure it's true for a lot of applicants from states with several state schools with in-state bias.

It makes sense too - why would an applicant choose out of state med school A if they have relatively comparable in-state med school B right in their backyard for half the price?


Sent from my iPhone using SDN mobile
I mean I understand that from their perspective. But from mine, I don't want to be stuck in the same place all my life.
 
Goes to show applicants are generally misinformed and ignorant, while admissions committees are smart and overworked. Not surprising! I'm just glad adcoms aren't swayed by the mere exposure effect of old MCAT scores.
 
  • Like
Reactions: 1 user
This is super interesting. I also wonder if the people who chose to apply with the old mcat score and not retake were more likely to have done well on the old mcat and people who did poorly on the old mcat were more likely to retake and applied with new scores (and perhaps were more likely to have done poorly again?). Or maybe that makes no sense
 
This is super interesting. I also wonder if the people who chose to apply with the old mcat score and not retake were more likely to have done well on the old mcat and people who did poorly on the old mcat were more likely to retake and applied with new scores (and perhaps were more likely to have done poorly again?). Or maybe that makes no sense
It makes sense, I thought about this too. It is true that the first cohort of new scores would include a lot of people retaking, having gotten a poor score on the old version earlier that year.

Thing is I don't think it influences what we're seeing here - they got back the second, still weak score and decided to apply with it, a behavior apparently not as common in prior years. People back in 2012-2014 would often get a bad score and retake, but apparently would then give up rather than applying anyways.
 
Who are these people with 518+ not getting accepted? This boggles my mind.

Along with what others have said, a well crafted school list is crucial. If you apply to only t20 schools, you're competing with very high stat/accomplishment students. Whereas, if you apply to t20 and 10 schools 30-50, the pool of candidates will have worse stats, so your odds of acceptances are going to be much better.
 
Along with what others have said, a well crafted school list is crucial. If you apply to only t20 schools, you're competing with very high stat/accomplishment students. Whereas, if you apply to t20 and 10 schools 30-50, the pool of candidates will have worse stats, so your odds of acceptances are going to be much better.
That's true. I just find it hard to believe that people who are intelligent enough/worked hard enough to get a 518+ on the MCAT wouldn't also do some research into making a good school list. lol. But I guess there will always be those people.
 
It makes sense, I thought about this too. It is true that the first cohort of new scores would include a lot of people retaking, having gotten a poor score on the old version earlier that year.

Thing is I don't think it influences what we're seeing here - they got back the second, still weak score and decided to apply with it, a behavior apparently not as common in prior years. People back in 2012-2014 would often get a bad score and retake, but apparently would then give up rather than applying anyways.

That's a good point, I don't know anything about behavior from previous years. Perhaps a combination of people retaking due to poor scores and confusion about what's a competitive score with the new system?
 
Along with what others have said, a well crafted school list is crucial. If you apply to only t20 schools, you're competing with very high stat/accomplishment students. Whereas, if you apply to t20 and 10 schools 30-50, the pool of candidates will have worse stats, so your odds of acceptances are going to be much better.

I definitely agree that crafting a good school list is important. However, how many schools 30-50 are going to interview you if your stats are greater than their 90th percentile?
 
I definitely agree that crafting a good school list is important. However, how many schools 30-50 are going to interview you if your stats are greater than their 90th percentile?

IDK. Depends on your application, how well your secondary convinces them you are gonna matriculate there if accepted or how much they want a high stat student.
 
I'm glad the numbers support anecdotes I've come across exhibiting the same phenomena. The new MCAT appears to be poorly understood among many people. I've heard many claim to be "shooting for a 500," which is a wholly useless goal unless you're planning on going pod. It's unsurprising that people would then be happy with their "goal" score only to be disappointed after the app cycle.
 
I'm glad the numbers support anecdotes I've come across exhibiting the same phenomena. The new MCAT appears to be poorly understood among many people. I've heard many claim to be "shooting for a 500," which is a wholly useless goal unless you're planning on going pod. It's unsurprising that people would then be happy with their "goal" score only to be disappointed after the app cycle.
I'm sure many premed advisors went off the AAMC recommendations and told their students 500+ = good to go for med school
 
@efle

suggestion for the graphs.

Can you make the X axis the either both Old mcat scores or new mcat scores? It would make comparing the two much easier.
 
@efle

suggestion for the graphs.

Can you make the X axis the either both Old mcat scores or new mcat scores? It would make comparing the two much easier.
I wish! Unfortunately not because the percentiles do not line up correctly for most bins. I'd need access to the AAMC's full data set on numbers per score, rather than per score range.
 
I should add that the 518+ and 36+ are equivalent percentile, and both graph left tails are about lined up because 486 = 17. So what you're looking at above is at least the same percentile range stretched across the same size X-axis, just with the bins divvied up slightly differently.
 
I wish! Unfortunately not because the percentiles do not line up correctly for most bins. I'd need access to the AAMC's full data set on numbers per score, rather than per score range.

ah. You should link your Old MCAT to New MCAT score chart in the original post! That would be handy
 
  • Like
Reactions: 1 user
I'm glad the numbers support anecdotes I've come across exhibiting the same phenomena. The new MCAT appears to be poorly understood among many people. I've heard many claim to be "shooting for a 500," which is a wholly useless goal unless you're planning on going pod. It's unsurprising that people would then be happy with their "goal" score only to be disappointed after the app cycle.

500 is good enough for some decent DO schools out there.
 
Top