TIL one decade ago, average MCATs at top medical schools were ~32-34

[efle]

---

Members don't see this ad.

Found this during a discussion buried in another thread and through it warranted it's own space to be shared.

I found this from a SUNY bio department webpage showing stats from 2006.

Apparently, these schools were the only places with a 30+ MCAT in the early/mid 2000s. The best of the best were about 32-34:

Pretty shocking to think that in about ten years, these schools moved their score distribution so high that their 10th percentiles are now at their old medians.

Some of the biggest, like at Chicago, jumped half a dozen points or more.

The MCAT distribution of scores was very similar back in the day, by the way - here is the 2006 distribution (also put it below), with a 30 and 33 the same as today, at 79th and 91st percentiles.

 

[Lawpy]

the score creep is real.... and suffocating

 

[EEtoPre-Med]

So if the percentiles were the same, does that just mean that way more people are applying now?

 

[Nommonnom]

So if the percentiles were the same, does that just mean that way more people are applying now?

I think there are twice as many applicants.

 

[Pusheen]

TIL I had higher stats than the average Harvard matriculant 10 years ago

 

[efle]

So if the percentiles were the same, does that just mean that way more people are applying now?

The percent admitted this last cycle was 40%, and a decade ago it looks like it was about 43% (first number from current AAMC tables and second from one of their old figures, below).

The only way to explain a jump of this magnitude this fast (often 3-4 points, sometimes like 4-5+) is that the medical schools flipped out over class scores/grades and very heavily emphasized them in admissions, much more than they had been doing previously.

This was also right around the time that it became the norm to use the internet to find school stats easily, and lots of rankings and comparison tools started coming out. I know by the time I was applying to colleges in 2011 the system for at least a few years had been to look up what schools had scores like yours and go from there. I imagine something similar happened with med admissions a few years prior and suddenly having a 30 or 31 or 32 median was unacceptable for places like Columbia and Penn.

Also remember this is happening in the far right tail of a distribution. Bumping your median from low 30s to higher 30s means you stop building a class centered in the top 15-20% and start building a class centered in the top 1-3%. It was an order of magnitude jump in less than ten years.

 

[efle]

TIL I had higher stats than the average Harvard matriculant 10 years ago

Pretty insane to think an 11/11/11 breakdown put someone above average at Chicago and Penn huh.

 

[Nommonnom]

The percent admitted this last cycle was 40%, and a decade ago it looks like it was about 43% (first number from current AAMC tables and second from one of their old figures, below).

The only way to explain a jump of this magnitude this fast (often 3-4 points, sometimes like 4-5+) is that the medical schools flipped out over class scores/grades and very heavily emphasized them in admissions, much more than they had been doing previously.

This was also right around the time that it became the norm to use the internet to find school stats easily, and lots of rankings and comparison tools started coming out. I know by the time I was applying to colleges in 2011 the system for at least a few years had been to look up what schools had scores like yours and go from there. I imagine something similar happened with med admissions a few years prior and suddenly having a 30 or 31 or 32 median was unacceptable for places like Columbia and Penn.

Also remember this is happening in the far right tail of a distribution. Bumping your median from low 30s to higher 30s means you stop building a class centered in the top 15-20% and start building a class centered in the top 1-3%. It was an order of magnitude jump in less than ten years.

This is just wrong, your chart shows 30k applicants in early 2000's. There are about 60k now. There are simply more people with higher numbers.

 

[efle]

This is just wrong, your chart shows 30k applicants in early 2000's. There are about 60k now. There are simply more people with higher numbers.

There are far more seats as well too, holding overall admit rate more steady than you'd expect. In the 2006 distribution I linked there were 71,000 scores assigned, while in the year I tested recently there were about 95,000 scores assigned. Schools jumping from an 83rd percentile median to 97th+ isn't readily explained this way either.

Edit: Also in 2006-2007 is when it crossed over the 40k threshold. I drew it out when I guesstimated.

 

[LutGholein]

I wonder why there was a spike in the mid 1990s in terms of applicants-- I thought the economy was generally pretty solid with the whole dot com boom?

 

[efle]

@LizzyM your badge says 10+ years and you're at a top program. Have you seen a change in the admissions mindset? Or has a 37+ always been valued the same, they're just a little less rare now?

 

[Nommonnom]

I wonder why there was a spike in the mid 1990s in terms of applicants-- I thought the economy was generally pretty solid with the whole dot com boom?

Heavy Reagan recession until 93-95. 87 (89?) stock market collapse.

 

[Nommonnom]

There are far more seats as well too, holding overall admit rate more steady than you'd expect. In the 2006 distribution I linked there were 71,000 scores assigned, while in the year I tested recently there were about 95,000 scores assigned. Schools jumping from an 83rd percentile median to 97th+ isn't readily explained this way either.

Edit: Also in 2006-2007 is when it crossed over the 40k threshold. I drew it out when I guesstimated.

Yes but at the high end the seats at top schools have been steady while there are more people with higher numbers, leading to higher numbers at the top schools.

 

[ace_inhibitor111]

There are far more seats as well too, holding overall admit rate more steady than you'd expect. In the 2006 distribution I linked there were 71,000 scores assigned, while in the year I tested recently there were about 95,000 scores assigned. Schools jumping from an 83rd percentile median to 97th+ isn't readily explained this way either.

Edit: Also in 2006-2007 is when it crossed over the 40k threshold. I drew it out when I guesstimated.

Even if all schools had increased the number of seats in their class, the number of new seats in the Top 20 will still be much smaller than the number of new seats in the 20-114 range. An extra 24,000 new scores would mean a good 1,000 or so in the 96%+ percentile. Given this I think it wouldn't be too surprising for the scores to shift considerably in the smallish sample size of ~150 students in each of the Top 20 schools. Nevertheless, I do think there is some influence of a desire for higher scores, but that it does not explain all the data.

 

[efle]

Yes but at the high end the seats at top schools have been steady while there are more people with higher numbers, leading to higher numbers at the top schools.

Well look at it this way. WashU in 2006 had a 34 median, when there were ~6500 people per year with a 34+. WashU just now had a 38 median, when there are only 1000-2000 people per year with a 38+. It is even more egregious for cases like Penn or Chicago that went from half the class drawn among 9000-12000 to being drawn from the top 1000-2000. The population shifts just don't even come close for this!

 

[efle]

Even if all schools had increased the number of seats in their class, the number of new seats in the Top 20 will still be much smaller than the number of new seats in the 20-114 range. An extra 24,000 new scores would mean a good 1,000 or so in the 96%+ percentile. Given this I think it wouldn't be too surprising for the scores to shift considerably in the smallish sample size of ~150 students in each of the Top 20 schools. Nevertheless, I do think there is some influence of a desire for higher scores, but that it does not explain all the data.

See my above post. The sudden score favoritism even looking at raw numbers is marked.

 

[Nommonnom]

Well look at it this way. WashU in 2006 had a 34 median, when there were ~6500 people per year with a 34+. WashU just now had a 38 median, when there are only 1000-2000 people per year with a 38+. It is even more egregious for cases like Penn or Chicago that went from half the class drawn among 9000-12000 to being drawn from the top 1000-2000. The population shifts just don't even come close for this!

Schools select from the applicant pool not the test taker pool.

https://www.aamc.org/newsroom/newsreleases/446400/applicant-and-enrollment-data.html

52k applicants in 2015. 35k in 2004 per your chart.

 

[efle]

Schools select from the applicant pool not the test taker pool.

https://www.aamc.org/newsroom/newsreleases/446400/applicant-and-enrollment-data.html

52k applicants in 2015. 35k in 2004 per your chart.

Try thinking about it slowly for a while. In 2006 there were about 9000 people that would be top half of Penn. In 2016 there were only a couple thousand. Nothing about overall pool growth explains that. The numbers of 37+ scorers did not increase so much that the same degree of MCAT emphasis would move schools to a 37-38 median. Not even close.

 

[Nommonnom]

Try thinking about it slowly for a while. In 2006 there were about 9000 people that would be top half of Penn. In 2016 there were only a couple thousand. Nothing about overall pool growth explains that. The numbers of 37+ scorers did not increase so much that the same degree of MCAT emphasis would move schools to a 37-38 median. Not even close.

I don't know where you're getting your numbers from so I can't check them, but off the cuff 9k out of 35k top half at Penn seems completely wrong. Sourced numbers would be helpful.

 

[Nommonnom]

There were 37,000 applicants in 2000 with about 14,000 first year seats for a ratio 2.6:1
Compare that with 52,000 applicants in 2016-17 for roughly 22,000 seats for a ratio 2.3:1

The applicant to seat ratio has gone down

But seats at top schools, which is what the original post cited, have stayed roughly constant leading to number inflation.

 

[efle]

I don't know where you're getting your numbers from so I can't check them, but off the cuff 9000 out of 35k top half at Penn seems completely wrong. Sourced numbers would be helpful.

71000 people take the MCAT
About half apply (still the case)
9000 scored a 32+
Penn median was 32

Decade later in recent years:

95000 take the MCAT
About half apply
A thousand or two scored 37-38
Penn median 37-38

 

[Nommonnom]

71000 people take the MCAT
About half apply (still the case)
9000 scored a 34+
Penn median was 34

I believe you're assuming that the highest scorers apply, but you haven't given all your math/assumptions. I don't believe this is true.

 

[efle]

I believe you're assuming that the highest scorers apply, but you haven't given all your math/assumptions. I don't believe this is true.

Check the AAMC tables. It is indeed true that app rates from higher scorers are very high and from lower scorers are very low. E.g. from the old aggregate table of the last 3 years, there were 33000 people that applied with a 33+ out of about 39000 awarded those scores in the same time period. On the flip side, out of the ~115,000 scoring a 23 and below, only about 20k applied.

Regardless, even if we assumed a dropoff rate that was significant at the high end (and it isn't), my argument would be the same because both values would get reduced.

E.g. 9000 becomes 7500 for the old admissions game, but then the 2000 now would drop to 1660 as well.

 

[Nommonnom]

Check the AAMC tables. It is indeed true that app rates from higher scorers are very high and from lower scorers are very low. E.g. from the old aggregate table of the last 3 years, there were 33000 people that applied with a 33+ out of about 39000 awarded those scores in the same time period. On the flip side, out of the ~115,000 scoring a 23 and below, only about 20k applied.

I believe another point is that electronic applications were new then I believe, so I'm sure people didn't just spam applications to every school like they do now, so the applicant pool for every school would have much fewer high numbers.

 

[efle]

I believe another point is that electronic applications were new then I believe, so I'm sure people didn't just spam applications to every school like they do now, so the applicant pool for every school would have much fewer high numbers.

You're that committed to your version huh? That it becomes more plausible the high achievers couldn't be bothered to mail something, than that admissions put more emphasis on stats?

 

[Nommonnom]

You're that committed to your version huh? That it becomes more plausible the high achievers couldn't be bothered to mail something, than that admissions put more emphasis on stats?

I'm pretty sure the average apps were around 7 or lower with mail.

Also your numbers argument doesn't contradict what I was saying that there are more people with higher numbers to choose from.

You're simply saying they are also choosing more people with higher numbers as well, which may be true, but we are speculating because of the number of assumptions in play to reach your point, which I am illustrating by discussing the number of apps etc.

 

[efle]

I'm pretty sure the average apps were around 7 or lower with mail.

Also your numbers argument doesn't contradict what I was saying that there are more people with higher numbers to choose from.

You're simply saying they are also choosing more people with higher numbers as well, which may be true, but we are speculating because of the number of assumptions in play to reach your point, which I am illustrating by discussing the number of apps etc.

Not really tho.

"For the 2006-07 entering class, there were 483,148 applications from 39,108 applicants—an average of 12 applications by each aspiring medical student"

The number was 15 in 2016.

Seriously I'm not making a bunch of crazy assumptions. The schools have started filling half their class from a much smaller population that is way up at the top end of the curve. The 10th percentiles for the current distributions are above the 2006 medians in a lot of cases. It's a massive shift not explained at all by the numbers of 37s being handed out or applying.

 

[ace_inhibitor111]

I wonder how they were able to sift through all those applications back then without weighing MCAT as heavily? Extracurriculars? But even then I doubt applicants had as many or as impressive ECs as you see today. Even doing EMT or getting an MPH was considered unique back then.

 

[Nommonnom]

Not really tho.

"For the 2006-07 entering class, there were 483,148 applications from 39,108 applicants—an average of 12 applications by each aspiring medical student"

The number was 15 in 2016.

Seriously I'm not making a bunch of crazy assumptions. The schools have started filling half their class from a much smaller population that is way up at the top end of the curve. The 10th percentiles for the current distributions are above the 2006 medians in a lot of cases. It's a massive shift not explained at all by the numbers of 37s being handed out or applying.

Top 5%, about a 34, was about 1500 in 2002. Top 2% maybe a 38 or so is about 1500 people now.

 

[efle]

Top 5%, about a 34, was about 1500 in 2002. Top 2% maybe a 38 or so is about 1500 people now.

Bruh, stop. I have exact numbers on this already.

In 2006 a 34 started at 90.6th percentile, and there were 71000 test takers, so about 6670 with that score. We know from the AAMC tables that the vast majority apply (like 85+%) but lets say 75% go ahead and apply. Almost exactly 5000.
In 2016 a 37+ was about the top 2000, yeah. Assume again only 75% apply and we get the 1500. Very different number.

And that's not even taking one of the more obvious cases. WashU has always liked the crazy high scores. But a place like Penn? Chicago? They cut the pool they draw the top half from down to a couple thousand instead of around ten thousand.

These schools went from medians at ~32-33 to 10th percentiles at 32-33 and medians in the top percent or two. In less than a decade. Population change doesn't cover it dude, not even close!

 

[LoveBeingHuman:)]

Check the AAMC tables. It is indeed true that app rates from higher scorers are very high and from lower scorers are very low. E.g. from the old aggregate table of the last 3 years, there were 33000 people that applied with a 33+ out of about 39000 awarded those scores in the same time period. On the flip side, out of the ~115,000 scoring a 23 and below, only about 20k applied.

Regardless, even if we assumed a dropoff rate that was significant at the high end (and it isn't), my argument would be the same because both values would get reduced.

E.g. 9000 becomes 7500 for the old admissions game, but then the 2000 now would drop to 1660 as well.

Most of those low scorers apply DO

 

[Dr.Sticks]

Hmm
So I wonder what the EC expectations were a decade ago as well?

 

[Lawpy]

Hmm
So I wonder what the EC expectations were a decade ago as well?

probably not as insane as expectations today

 

[Nommonnom]

Bruh, stop. I have exact numbers on this already.

In 2006 a 34 started at 90.6th percentile, and there were 71000 test takers, so about 6670 with that score. We know from the AAMC tables that the vast majority apply (like 85+%) but lets say 75% go ahead and apply. Almost exactly 5000.
In 2016 a 37+ was about the top 2000, yeah. Assume again only 75% apply and we get the 1500. Very different number.

And that's not even taking one of the more obvious cases. WashU has always liked the crazy high scores. But a place like Penn? Chicago? They cut the pool they draw the top half from down to a couple thousand instead of around ten thousand.

These schools went from medians at ~32-33 to 10th percentiles at 32-33 and medians in the top percent or two. In less than a decade. Population change doesn't cover it dude, not even close!

You're misreading your chart. When you say a 34 starts at 90.6 percentile and ends at 93 %ile, that means it's 93 %ile. This lowers your number to about 3000. Then with a 20% increase in number of apps per person the 3000 vs 2000 isn't a huge difference.

 

[Flirt]

Interesting. Is there a step 1 equivalent of this chart? I wonder if step 1 average scores have gone up now as well

 

[Nommonnom]

Interesting. Is there a step 1 equivalent of this chart? I wonder if step 1 average scores have gone up now as well

Ya step 1 scores go up every year.

 

[efle]

You're misreading your chart. When you say a 34 starts at 90.6 percentile and ends at 93 %ile, that means it's 93 %ile. This lowers your number to about 3000. Then with a 20% increase in number of apps per person the 3000 vs 2000 isn't a huge difference.

Nope, the values describe the bin it is assigned to. As in 6.9% scored above a 34 and 2.5% were assigned a 34. So the total number of 34 or above is 9.4% x total.

The MCAT has since changed to only listing the ceiling, because they wanted to unambiguously refer to how many people beat you rather than how many tied+beat you. The bins and the math behind it is the same though.

And again this is the easiest case to try and argue. Look at schools that went 31-32 up to 37-38 and try your (already faulty!) math.

Ya step 1 scores go up every year.

Like the average scaled score goes up or the top residency admits are going up?

 

[Flirt]

Ya step 1 scores go up every year.

Any idea what the average was about the time presented on this chart? Assuming 34 then is the equivalent of 34 today, which it should be, what was step 1 average for WashU for example then?

 

[efle]

Any idea what the average was about the time presented on this chart? Assuming 34 then is the equivalent of 34 today, which it should be, what was step 1 average for WashU for example then?

Do you mean the average step1 earned by WashU students ten years ago? There is no such data on that afaik

 

[Flirt]

Do you mean the average step1 earned by WashU students ten years ago? There is no such data on that afaik

Yes that is what I meant. Ah, okay. If you ever do find it, as you are the resident data expert, please share.

 

[Lucca]

Yes that is what I meant. Ah, okay. If you ever do find it, as you are the resident data expert, please share.

http://www.neuronix.org/2012/06/effects-of-score-creep-trends-in.html?m=1

another SDNer did a post on this several years ago on his blog.

 

[Flirt]

Bruh, getting into WashU with a 34 has be laughing hysterically.

http://www.neuronix.org/2012/06/effects-of-score-creep-trends-in.html?m=1

another SDNer did a post on this several years ago on his blog.

LOL from that chart, both derm, ent and ortho had average step 1 scores of less than 240. I wish, man. I wish.

 

[LLLC]

.

 

[Nommonnom]

Nope, the values describe the bin it is assigned to. As in 6.9% scored above a 34 and 2.5% were assigned a 34. So the total number of 34 or above is 9.4% x total.

The MCAT has since changed to only listing the ceiling, because they wanted to unambiguously refer to how many people beat you rather than how many tied+beat you. The bins and the math behind it is the same though.

And again this is the easiest case to try and argue. Look at schools that went 31-32 up to 37-38 and try your (already faulty!) math.

If your whole argument depends on the percentile difference of one point of the MCAT, it's not much of an argument in my opinion haha!

Like the average scaled score goes up or the top residency admits are going up?

The average scaled step 1 score goes up.

 

[bigindian4891]

So whats the endgame in these situations? Will the score creep stop eventually?

 

[Gastrapathy]

The importance of these exams is better understood by applicants so you spend more time and $ preparing. Same thing holds for Step 1. UWorld, etc, didn't exist when I took that test. I flipped through an MCAT prep book halfheartedly for a couple weeks (which was more than my roommate who scored a cool 43 did in between his other activities).

There were still folks who prepared heavily (correctly given the game) but they were in the minority. We were told that preparing "wasn't fair." Increased transparency created an arms race.

 

[efle]

there were more higher scoring people in "lower" tier schools, and the distribution of score was less polarized

This seems backed up by data, e.g. people with very high MCATs got in at the same or even higher rates back then, yet the averages for top schools were so much lower. So what we've seen is a concentration of very higher scorers into a couple dozen places. The driving force is still the big question then - why would schools want to be so much more packed with that type of person? @Lucca mentioned an interesting idea to me that it might be in response to more emphasis on Step score by residencies. I still favor the idea that applicants became a lot more likely to compare and evaluate schools by their standardized metrics, so schools drove the numbers up to make it clear they were elite.

I do think the creep upwards in scores has been offset a little on the left side by new programs, though often the new programs have been public/mission driven so it is only a lucky subset of the below-median crowd that gets the benefit a lot.

 

[efle]

The importance of these exams is better understood by applicants so you spend more time and $ preparing. Same thing holds for Step 1. UWorld, etc, didn't exist when I took that test. I flipped through an MCAT prep book halfheartedly for a couple weeks (which was more than my roommate who scored a cool 43 did in between his other activities).

There were still folks who prepared heavily (correctly given the game) but they were in the minority. We were told that preparing "wasn't fair." Increased transparency created an arms race.

How long ago are we talking that you could score a 43 with no serious prep ???

 

[LLLC]

.

 

[Flirt]

The importance of these exams is better understood by applicants so you spend more time and $ preparing. Same thing holds for Step 1. UWorld, etc, didn't exist when I took that test. I flipped through an MCAT prep book halfheartedly for a couple weeks (which was more than my roommate who scored a cool 43 did in between his other activities).

There were still folks who prepared heavily (correctly given the game) but they were in the minority. We were told that preparing "wasn't fair." Increased transparency created an arms race.

Hahahaha, can't imagine how I would have done without preparing. I do believe the spirit of your post though; most physicians I have spoken to admit they basically free-styled the mcat, and can't really seem to get why some of us study for it as we do.