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

[efle]

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!

I'll just answer this with the post that prompted my response, where you made an argument based on the percentile difference of one point of the MCAT. All I did was correct you that my numbers were actually right.

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

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

I think the top schools have just about hit their max. To get higher than top 1-2% MCAT + 3.9 GPA you'd have to start giving up too much in the ECs arena. The overall creep of like a fraction of a point every few years I think will continue a little more, med schools aren't going to collectively run out of applicants with low 30s and everything else in order.

---

Members don't see this ad.

 

[Gastrapathy]

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

That dude could probably do it again tomorrow. I'm not sure he was sober and I know he was up all night the night before. You could meet him someday since he's moderately well known now.

But to answer your question, it was mid-90s. Now I need my geritol.

 

[LLLC]

.

 

[Gastrapathy]

The MCAT is scored on a percentile basis. The increase in preparation is probably reflected in the increase in the difficulty of the exams. As previously stated, the distribution of MCAT scores has remained stable over time.

You're missing the point. Percentiles stay the same but raw scores can change. Even if raw scores stay the same, the tests require a greater investment because they've scaled the difficulty to the current competition.

 

[Gastrapathy]

This is a bit interesting, and begs the question, what inspired the residencies to place more emphasis on Step scores? (similar to the question, what inspired the med schools to place more emphasis on MCAT scores?)
There could be an easy answer, I'm not well-versed enough to comment. I'm going to check if there was a tightening of spots due to larger IMG populations. Or perhaps even the expanding US med students

Charting outcomes

 

[efle]

Charting outcomes

I've recently learned there's a step 1 score creep of about 1 point per year. What reason is there for residency programs to care about this though? In med school, it makes sense to want students who crushed the MCAT because they will crush the step and get great residencies. But in residency? What exam is there to crush then that helps you get a good job? Why would they care so much about having cohorts that did great on step 1?

 

[Nommonnom]

I've recently learned there's a step 1 score creep of about 1 point per year. What reason is there for residency programs to care about this though? In med school, it makes sense to want students who crushed the MCAT because they will crush the step and get great residencies. But in residency? What exam is there to crush then that helps you get a good job? Why would they care so much about having cohorts that did great on step 1?

I would guess it's more that they don't care whether you were the president of 10 clubs or 20 clubs.

 

[Gastrapathy]

I've recently learned there's a step 1 score creep of about 1 point per year. What reason is there for residency programs to care about this though? In med school, it makes sense to want students who crushed the MCAT because they will crush the step and get great residencies. But in residency? What exam is there to crush then that helps you get a good job? Why would they care so much about having cohorts that did great on step 1?

It's the easiest way to compare your program to a standard. Everyone wants to be above average. I don't think it's conscious but we all know that our own programs are special 😉.

 

[mimelim]

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

I know of two students in the last two cycles who scored 524+ with about a week of prep. The arms race has shifted the average, but the overall population hasn't changed.

I've recently learned there's a step 1 score creep of about 1 point per year. What reason is there for residency programs to care about this though? In med school, it makes sense to want students who crushed the MCAT because they will crush the step and get great residencies. But in residency? What exam is there to crush then that helps you get a good job? Why would they care so much about having cohorts that did great on step 1?

Written boards. Most of residency/being a physician has little to do with being able to take tests or even knowledge. But, written boards demand focus and good studying. I don't have data to show that step 1 correlates with board pass rate, but that is a common reason given by PDs.

 

[MDProspect]

That dude could probably do it again tomorrow. I'm not sure he was sober and I know he was up all night the night before. You could meet him someday since he's moderately well known now.

But to answer your question, it was mid-90s. Now I need my geritol.

Are you talking about Scott Roberts?

 

[Gastrapathy]

I know of two students in the last two cycles who scored 524+ with about a week of prep. The arms race has shifted the average, but the overall population hasn't changed.



Written boards. Most of residency/being a physician has little to do with being able to take tests or even knowledge. But, written boards demand focus and good studying. I don't have data to show that step 1 correlates with board pass rate, but that is a common reason given by PDs.

https://www.ncbi.nlm.nih.gov/m/pubmed/17368416/

https://www.ncbi.nlm.nih.gov/m/pubmed/17615848/?i=2&from=/17368416/related

https://www.ncbi.nlm.nih.gov/m/pubmed/21292930/?i=6&from=/17615848/related

 

[sazerac]

I've recently learned there's a step 1 score creep of about 1 point per year. What reason is there for residency programs to care about this though? In med school, it makes sense to want students who crushed the MCAT because they will crush the step and get great residencies. But in residency? What exam is there to crush then that helps you get a good job? Why would they care so much about having cohorts that did great on step 1?

Because applicants apply to many more programs than they did in the past, and the programs being flooded with applications have to filter on something.

When you have the infrastructure to evaluate 1,000 applications for 100 interview slots, and you suddenly find yourself with 4,000 applications, it's pretty easy to set the Step1 filter to 225+ to reduce your workload back to manageable levels like in prior years.

 

[efle]

I know of two students in the last two cycles who scored 524+ with about a week of prep. The arms race has shifted the average, but the overall population hasn't changed.

Written boards. Most of residency/being a physician has little to do with being able to take tests or even knowledge. But, written boards demand focus and good studying. I don't have data to show that step 1 correlates with board pass rate, but that is a common reason given by PDs.

Pass rate? Were people at competitive residencies previously failing?

 

[mimelim]

Pass rate? Were people at competitive residencies previously failing?

Yes, and they still do, just at a lower rate than other programs. ~80% pass rate for general surgery and internal medicine for first time takers. Didn't bother looking up others, but I would assume similar.

 

[AnatomyGrey12]

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've recently learned there's a step 1 score creep of about 1 point per year. What reason is there for residency programs to care about this though? In med school, it makes sense to want students who crushed the MCAT because they will crush the step and get great residencies. But in residency? What exam is there to crush then that helps you get a good job? Why would they care so much about having cohorts that did great on step 1?

I think the idea in the above post is a big factor, everyone likes to think they are the best. With the advent of things like charting outcomes for residencies, med school applicants looking more at numbers, and the like we are seeing these schools and residencies raising their average numbers so they can simply report having the best. I don't know if it is a conscious decision but I think this happens and explains a lot.

Simple applicant population growth doesn't explain the huge bump in MCAT scores at these types of places. It's not just MCAT scores either, go look at the average GPAs on your posted list. Some of these schools have GPAs posted at 3.5, and now their averages are well above 3.75 with many above 3.8 and 3.9. I actually wonder if the ceiling has been hit and averages will slowly start to ease up a bit.

 

[hungrybox]

Haven't read every single post, so this may have already been pointed out, but one factor that could be skewing the data is MCAT retakes.

Correct me if I'm wrong, but my understanding is that back in the day MCAT retakes were not as acceptable as they are now. However, I think that schools only use top scores when averaging their matriculant stats.

 

[Shreman]

Really interesting topic. One point of contention I can see is the use of average MCATs vs. efle's comparison to current median MCATs. Let's take Pittsburgh (a USNWR top 20 school) as a case study.

According to the document, Pittsburgh's average MCAT was 31.2 in 2006.
Looking at their website (http://www.medadmissions.pitt.edu/our-students/), their current average MCAT is ~33.

That is a 1.8 point change over 10 years, or 0.18 point change per year. This seems in line with my intuitions of MCAT score creep.

If you look at UPenn's current average MCAT (http://www.med.upenn.edu/admissions/entering-class-profile.html) however, it is a 38, compared to a 32.1 average 10 years ago. This is a 6.1 point change, or a 0.61 point change/year.

Tl;dr: Changes in scores seem to be school specific. The highly accelerated score creep is probably focused on the ~top 10 USNWR schools. Maybe this indicates how "holistic" each school is in screening and selecting applicants.

 

[efle]

Really interesting topic. One point of contention I can see is the use of average MCATs vs. efle's comparison to current median MCATs. Let's take Pittsburgh (a USNWR top 20 school) as a case study.

According to the document, Pittsburgh's average MCAT was 31.2
Looking at their website (http://www.medadmissions.pitt.edu/our-students/), their current average MCAT is ~33

That is a 1.8 point change over 10 years, or 0.18 point change per year. This seems in line with my intuitions of MCAT score creep.

If you look at UPenn's current average MCAT (http://www.med.upenn.edu/admissions/entering-class-profile.html) however, it is a 38, compared to a 32.1 average 10 years ago. This is a 6.1 point change, or a 0.61 point change/year.

Tl; dr: Changes in scores seem to be school specific. The highly accelerated score creep is probably focused on the ~top 10 USNWR schools. Maybe this indicates how "holistic" each school is in screening and selecting applicants.

I'd point out the 33 is the enrolled/matriculating cohort average for Pitt. MSAR data (which appears to be what the SUNY used) has always been accepted data, matriculant info wasn't added until last year. The fact that pitt was average 31 and is now 10th percentile 31 makes me think they are not absolved! Though it is very surprising that they can be admitting a 10th percentile of 31 and enrolling an average 33.

 

[aprilfools]

However, I think that schools only use top scores when averaging their matriculant stats.

This isn't true. Schools almost universally average MCAT scores so if you are looking at top 20 schools with MCAT averages of 37-38, you pretty much have one shot to do it.

 

[Dr.Sticks]

probably not as insane as expectations today

It seems as competition for seats in medical school has gone up Medical Schools have sought alternative ways to judge applications.
I wonder what will happen 10 years from now?
Think about it: DO schools are close to where M.D schools were 10 years ago..
Close, not exactly there but close..
Will D.O schools be where M.D schools are in 10 years? Hmm

 

[efle]

I'd also look at a couple more schools around Pitt. Vandy online webpage says "mean GPA for the 2016 entering class was 3.87 and the mean MCAT was the 98th percentile" (up to 37 from 33) and Northwestern was entering class 3.87/36 despite being a 29 back then. Not just the top 10!

 

[Dr.Sticks]

I'd also look at a couple more schools around Pitt. Vandy online webpage says "mean GPA for the 2016 entering class was 3.87 and the mean MCAT was the 98th percentile" (up to 37 from 33) and Northwestern was entering class 3.87/36 despite being a 29 back then. Not just the top 10!

I wonder if we will see 40 averages?
Which is 525 right?

 

[efle]

I wonder if we will see 40 averages?
Which is 525 right?

I don't think so, there are only a couple hundred people per year with a 40+ (99.8th percentile)
Once you start getting that high you literally couldn't fill all the top school seats any more

 

[Jalby]

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

This is the most important data. While the number of applicants and the number of spots have increased, the number of spots at the top schools have not increased. There are still 150 at Penn, 200 at Harvard, etc, etc.

So if the top schools can take the top 3000 applicants, in 2002 they would have had and average of 34. If the top schools take the top 3,000 applicants today, they would have an average score of 38.

The fact that the UC Riversides of the world created medical schools which increased the total number of spots does not change the best schools taking the best students.

 

[efle]

This is the most important data. While the number of applicants and the number of spots have increased, the number of spots at the top schools have not increased. There are still 150 at Penn, 200 at Harvard, etc, etc.

So if the top schools can take the top 3000 applicants, in 2002 they would have had and average of 34. If the top schools take the top 3,000 applicants today, they would have an average score of 38.

The fact that the UC Riversides of the world created medical schools which increased the total number of spots does not change the best schools taking the best students.

His math is all completely wrong, detailed in the surrounding posts. There are not 3000 34s back then but 6500, and that's just considering WashU, the biggest scorewhore. Places that went 31/32 -> 37 you cannot even come close to explaining from the small growth in raw numbers of 37+ applicants.

 

[Dr.Sticks]

I don't think so, there are only a couple hundred people per year with a 40+ (99.8th percentile)
Once you start getting that high you literally couldn't fill all the top school seats any more

If anyone will get there WashU will be there first.
I call it already!

 

[Shreman]

I'd point out the 33 is the enrolled/matriculating cohort average for Pitt. MSAR data (which appears to be what the SUNY used) has always been accepted data, matriculant info wasn't added until last year. The fact that pitt was average 31 and is now 10th percentile 31 makes me think they are not absolved! Though it is very surprising that they can be admitting a 10th percentile of 31 and enrolling an average 33.

One key mistake your are making is assuming this info. is from MSAR. MSAR has historically never reported average statistics, only median. As such, I believe this data set is from a non-AAMC source (school websites, US news, etc.). That, combined with the fact that the document reports # of matriculants leads me to believe these are matriculant scores, not applicant scores.

Another counter example is U Michigan (USNWR Top 10). Their 2006 average was 32.1. Their current average is 33 (https://medicine.umich.edu/medschoo...unity/students-faculty/admitted-class-profile). This is a 0.9 change over 10 years. I imagine similar results will be seen for public top 20 schools.

Generalizations are not sufficient for this debate; each school judges differently.

 

[efle]

One key mistake your are making is assuming this info. is from MSAR. MSAR has historically never reported average statistics, only median. As such, I believe this data set is from a non-AAMC source (school websites, US news, etc.). That, combined with the fact that the document reports # of matriculants leads me to believe these are matriculant scores, not applicant scores.

Another counter example is U Michigan (USNWR Top 10). Their 2006 average was 32.1. Their current average is 33 (https://medicine.umich.edu/medschoo...unity/students-faculty/admitted-class-profile). This is a 0.9 change over 10 years. I imagine similar results will be seen for public top 20 schools.

Generalizations are not sufficient for this debate; each school judges differently.

It shows matric data for the full population. Full population matric info = full accept info. The two can only differ on the individual school basis where they lose more of the upper end of their cohort. Anyways the big problem with this is that many of these schools do not post their average stats. Try to go find WashU's average accepted or enrolled GPA and MCAT on their website for example. Using US News wouldn't fix it either, they also report admit info not matric info.

 

[efle]

One key mistake your are making is assuming this info. is from MSAR. MSAR has historically never reported average statistics, only median. As such, I believe this data set is from a non-AAMC source (school websites, US news, etc.). That, combined with the fact that the document reports # of matriculants leads me to believe these are matriculant scores, not applicant scores.

Another counter example is U Michigan (USNWR Top 10). Their 2006 average was 32.1. Their current average is 33 (https://medicine.umich.edu/medschoo...unity/students-faculty/admitted-class-profile). This is a 0.9 change over 10 years. I imagine similar results will be seen for public top 20 schools.

Generalizations are not sufficient for this debate; each school judges differently.

I should add I would love this explanation if there was such data out there. But the only source(s) for average MCAT on all schools have been sources that post accepted. Searching websites for matriculant scores wouldn't get you a list this complete.

 

[efle]

What factor we are leaving out of this mix is the number of applications per student, which has been steadily climbing

I found this value earlier, it was 12 per person in 2006-2007 and 15 per person in 2015-2016.

 

[Shreman]

Just for fun, here's the data I could collect from school websites.

Edit: Check a few posts down for the updated table of schools

 

[aalamruad]

Just for fun, here's the data I could collect from school websites.

View attachment 215553

Where'd you get the data for UCSF? I'm pretty sure their average is somewhere around 34-35 for applicants and 33-34 for matriculants

 

[efle]

Can I get sources? Like I just checked UCSF and it says average MCAT 33/91st percentile (much closer to the 34 median in the MSAR)

 

[Shreman]

It shows matric data for the full population. Full population matric info = full accept info. The two can only differ on the individual school basis where they lose more of the upper end of their cohort. Anyways the big problem with this is that many of these schools do not post their average stats. Try to go find WashU's average accepted or enrolled GPA and MCAT on their website for example. Using US News wouldn't fix it either, they also report admit info not matric info.

I am unclear on your assertion. The key fact is: MSAR has historically never published any averages. Therefore, this document cannot be MSAR official information.

It is likely this information is matriculant info from US News or from school websites. As such, a comparison to USNWR reported averages (I'm assuming, since I don't have access to compass) is the best source for a comparison.

 

[Shreman]

Can I get sources? Like I just checked UCSF and it says average MCAT 33/91st percentile (much closer to the 34 median in the MSAR)

Oops thanks for the catch! I flipped the UCSF numbers. Here's the updated stats. I drew them all from their individual websites.

As for the weird 2016 values, here are some notes:

1. Yale posted a single average score for the "3 subsections" as 12.2 (http://medicine.yale.edu/facts/education/)

2. Emory seemed to be mistaken on the old to new MCAT conversion, as they state: "The average "old" MCAT score is a 34 (approx. the 90th percentile), with at least a 10 in each of the science subtests. The average "new" MCAT score is 514 (approx. the 90th percentile)." (http://www.med.emory.edu/education/admissions/md/faq.html) So using Efle's trusty conversion table, I averaged 34 and 32.5 (90th percentile) for their 2016 average. *this point can be contested

3. Baylor had a 514 average (https://www.bcm.edu/education/schools/medical-school/admissions/faqs) so I converted that to 33.5 on the old scale

 

[efle]

I am unclear on your assertion. The key fact is: MSAR has historically never published any averages. Therefore, this document cannot be MSAR official information.

It is likely this information is matriculant info from US News or from school websites. As such, a comparison to USNWR reported averages (I'm assuming, since I don't have access to compass) is the best source for a comparison.

Has it never? It was medians ten years ago too? My assertion is that either the MSAR once gave averages, or SUNY erroneously put mean when it was actually median. Or that USNWR once gave averages. In no case do I think they used matriculant info though, since this was not available anywhere except for the handful that published it.

US News doesn't give matriculant info either. School websites in like half or more cases don't give it either. I have compass and the USNWR values line up with MSAR from a year prior.

 

[efle]

The current location of 10th percentiles might even be of more interest. I don't have access right now but if you plotted 2006 avg against 2016 10th percentiles I wonder how many have moved up to/past their median from a decade ago.

 

[efle]

Also interesting for UCSF is that another of their pages says the same entering class, 2016, had a 34 average. Maybe they mistakenly put average when they meant median 😉

 

[efle]

2. Emory seemed to be mistaken on the old to new MCAT conversion, as they state: "The average "old" MCAT score is a 34 (approx. the 90th percentile), with at least a 10 in each of the science subtests. The average "new" MCAT score is 514 (approx. the 90th percentile)." (http://www.med.emory.edu/education/admissions/md/faq.html) So using Efle's trusty conversion table, I averaged 34 and 32.5 (90th percentile) for their 2016 average. *this point can be contested

34 and 514 are 94th and 92nd percentiles, that doesn't seem off.

 

[Shreman]

One thing that I think would be cool/valuable would be for someone to store all the median gpa and MCAT data from the current MSAR in Excel and compare it 10 years down the line. Obviously there are some issues with sharing such info, since MSAR is a paid product. It nevertheless would be cool to look at in 10 years.

 

[Shreman]

34 and 514 are 94th and 92nd percentiles, that doesn't seem off.

I guess I'm too neurotic and overplaying the difference between 90 and 94th percentiles

 

[efle]

Here's what I was interested in above for the first 20 names, you can see some schools rocketed their bottom end way up while others still span quite low:

 

[Shreman]

The current location of 10th percentiles might even be of more interest. I don't have access right now but if you plotted 2006 avg against 2016 10th percentiles I wonder how many have moved up to/past their median from a decade ago.

Edit: Efle beat me to it

 

[efle]

Wow I didn't do Case in mine, JFC their bottom 10% mark is now 4-5 points above their typical student just ten years ago.

Also Columbia def wasn't a 27.9 average?

 

[Shreman]

Yup, Columbia is def not that low. I need to be more careful on this. Am going back and checking all my numbers again

 

[libertyyne]

Wow I didn't do Case in mine, JFC their bottom 10% mark is now 4-5 points above their typical student just ten years ago.

Also Columbia def wasn't a 27.9 average?

There is also a consideration of the accetpee MCAT being different then the matriculant mcat averages medians and cuttoffs.

 

[efle]

There is also a consideration of the accetpee MCAT being different then the matriculant mcat averages medians and cuttoffs.

Yeah we were discussing that above. I'm inclined to believe the SUNY numbers are acceptee data from MSAR or US News rather than matriculant data taken from websites, because so many of the schools listed do not share matriculant data online.

 

[Shreman]

Ok so I've run out of patience, but after 39 schools, here are the top 10

 

[Shreman]

At least last year (i'm not sure about this year), the MSAR posted the median data which in a skewed sample is going to be different than the mean.

According to this SDN post, after 2005 MSAR was median data. It seems before 2004 MSAR may have been average data.

 

[efle]

Ok so I've run out of patience, but after 39 schools, here are the top 10 worst offenders

View attachment 215564

Awesome stuff! How many of the 39 had moved their 10th to be equal or higher than their old avg?