Efle's Percent Correct vs Percentile Earned Charts for the Old MCAT

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

efle

not an elf
7+ Year Member
Joined
Apr 6, 2014
Messages
14,146
Reaction score
22,775
I was always curious about the relationship between number of correct answers and resulting score percentiles. So I decided to go ahead and plot it out when I saw someone asking about it earlier in a thread.

The charts are built using the scoring scales for AAMC Official Practice Tests 9, 10, and 11 plus the Final Percentile Ranks for the old MCAT.

Method: Since percentiles listed in the AAMC ranks are the ceilings of percentile bins, I paired to the highest number in the bins of correct answers possible to earn a score (averaged across the three tests). I know that's confusing, so...

For example, in PS the numbers correct (out of 52) for a section score of 13 were 47-48, 49-49, and 47-48 across Practice Tests 9, 10 and 11 respectively. I averaged the 48, 49, and 48 and converted into percentage to get 93% correct answers. Paired with a 13 being 97th percentile, this appears on the PS chart as the point at (93, 97).

Points appearing in red mark the boundary beyond which the percentile values become higher than the percent correct (such as the 93, 97 pair). For all points before the red one, the percent correct is a higher value than percentile. For all charts the trendline is polynomial to the 4th power.

Here's my excel file for putting this together. Let me know any questions or mistakes. If the same number correct to section score conversion data becomes available for the MCAT 2015, please PM me! I'd love to chart that out as well to see if you need more or less correct answers for an equivalent percentile between the old and new MCATs.

It's interesting to note that Bio demands more correct answers to earn a similar percentile to PS or V!

Charts:
iWLCuPe.png

Members don't see this ad.
 
  • Like
Reactions: 3 users
I essentially did this a few days ago but didn't make it quite as fancy as you have, and I attempted to correlate it to new exam scores. Thoughts?

http://forums.studentdoctor.net/threads/tentative-aamc-sample-fl-score-conversion-sheet.1149516/
Things do look similar! I pretty much did what you have on the first sheet except only for FL 9-11 since the average across those used to be the gold standard for predicting your score. We still have to wait for a few MCAT 2015 practice exams to be released before we can see if there has been a big change in the % correct needed to achieve a given percentile!
 
  • Like
Reactions: 1 user
Members don't see this ad :)
I was always curious about the relationship between number of correct answers and resulting score percentiles. So I decided to go ahead and plot it out when I saw someone asking about it earlier in a thread.

The charts are built using the scoring scales for AAMC Official Practice Tests 9, 10, and 11 plus the Final Percentile Ranks for the old MCAT.

Method: Since percentiles listed in the AAMC ranks are the ceilings of percentile bins, I paired to the highest number in the bins of correct answers possible to earn a score (averaged across the three tests). I know that's confusing, so...

For example, in PS the numbers correct (out of 52) for a section score of 13 were 47-48, 49-49, and 47-48 across Practice Tests 9, 10 and 11 respectively. I averaged the 48, 49, and 48 and converted into percentage to get 93% correct answers. Paired with a 13 being 97th percentile, this appears on the PS chart as the point at (93, 97).

Points appearing in red mark the boundary beyond which the percentile values become higher than the percent correct (such as the 93, 97 pair). For all points before the red one, the percent correct is a higher value than percentile. For all charts the trendline is polynomial to the 4th power.

Here's my excel file for putting this together. Let me know any questions or mistakes. If the same number correct to section score conversion data becomes available for the MCAT 2015, please PM me! I'd love to chart that out as well to see if you need more or less correct answers for an equivalent percentile between the old and new MCATs.

It's interesting to note that Bio demands more correct answers to earn a similar percentile to PS or V!

Charts:
iWLCuPe.png


Do you mind looking at my raw AAMC practice test scores for the new mcat and translate these to a percentile:
Phys/Chem: 49%
Bio: 53%
CARS: 70%
Psyc/Soc: 49%

Trying to figure out where I stand. I'd like to score around a 500, which is my goal.
 
Do you mind looking at my raw AAMC practice test scores for the new mcat and translate these to a percentile:
Phys/Chem: 49%
Bio: 53%
CARS: 70%
Psyc/Soc: 49%

Trying to figure out where I stand. I'd like to score around a 500, which is my goal.

Based on mcatjelly's scale, you are at a
123/123/126/122 = 493
 
Do you mind looking at my raw AAMC practice test scores for the new mcat and translate these to a percentile:
Phys/Chem: 49%
Bio: 53%
CARS: 70%
Psyc/Soc: 49%

Trying to figure out where I stand. I'd like to score around a 500, which is my goal.
This can't be done yet for the new test, they have not released any score breakdowns that show how many correct you need for certain scores.
 
go further up on this thread, mcatjelly is the poster who provides a conversion.
No, his is essentially identical to mine, and is only valid for taking your raw score on the OLD EXAM practice tests and converting to percentile, which you can then compare to percentile on the new test.

There is currently no way to estimate percentile from NEW EXAM raw scores, which is what he wants.
 
No, his is essentially identical to mine, and is only valid for taking your raw score on the OLD EXAM practice tests and converting to percentile, which you can then compare to percentile on the new test.

There is currently no way to estimate percentile from NEW EXAM raw scores, which is what he wants.

Well we have been using it in most MCAT threads for a "general idea"
 
Well we have been using it in most MCAT threads for a "general idea"

Ya thats basically what I was hoping for. I know it sucks but without any AAMC written practice test to give anyone a score we all have to resort to other methods :)
 
Ya thats basically what I was hoping for. I know it sucks but without any AAMC written practice test to give anyone a score we all have to resort to other methods :)
Just read off the chart then! Plug in your percent correct to the appropriate curve polynomial and you'll get what percentile you could've expected if your percent corrects had been on an old practice exam
 
Just read off the chart then! Plug in your percent correct to the appropriate curve polynomial and you'll get what percentile you could've expected if your percent corrects had been on an old practice exam

Lol yep thats what I ended up doing. Thanks for responding.
 
Top