MCAT number score versus percentile

[BaronVonZ]

---

Members don't see this ad.

There has been a lot of discussion (particularly in the 6/17 mcat) about some tests being more difficult than others. I'm curious to see just how much variance there is in the scores awarded on a certain day. To answer this, could you post your Total MCAT score and what percentile that score is in? Expectation seems to be that the percentile breakdown will be curved in such a way that each score will be roughly as common on each testing date, but I can't find any solid information to back that.

 

[JSU]

Pretty sure the percentile is to compare you with EVERY other MCAT test taker and are not specific to each test. Here is my break down.

6/17/10 Administration.

Verbal Reasoning (VR):09 Percentile Range: 52.0 - 67.7
Physical Sciences (PS):12Percentile Range: 88.8 - 95.2
Writing Sample (WS):R Percentile Range: 83.8 - 94.3
Biological Sciences (BS):12 Percentile Range: 88.9 - 95.5
Total Score:33R Percentile Range: 87.8 - 91.1

 

[BaronVonZ]

Pretty sure the percentile is to compare you with EVERY other MCAT test taker and are not specific to each test. Here is my break down.

Thats exactly what I'm trying to figure out, all we need is someone with a 33 or a 36 from any other test date to find out. If the percentiles don't match then we can determine if the difficulty of getting a certain score extends beyond luck and personal strengths and into the administration.

Mine:
6/17/10 Administration.

Verbal Reasoning (VR):13 Percentile Range: 98.3 - 99.7
Physical Sciences (PS):12Percentile Range: 88.8 - 95.2
Writing Sample (WS):Q Percentile Range: 62.9 - 83.7
Biological Sciences (BS):11 Percentile Range: 77.1 - 88.8
Total Score:36Q Percentile Range: 95.8 - 97.3

 

[NickNaylor]

COPY EVERYTHING AFTER THIS SENTENCE IN FUTURE POSTS.

Directions:
1) Quote the most recent post
2) Remove the quote tags
3) Look for your test administration and add your data following the same format

5/1/2010 Test
===================

Verbal Reasoning (VR): 15 (99.9 - 99.9%)
Physical Sciences (PS): 13 (95.3 - 97.3%)
Writing Sample (WS): S (94.4 - 99.3%)
Biological Sciences (BS): 13 (95.6 - 97.6%)
Total Score: 41S (99.8 - 99.9%)


6/17/2010 Test
===================

Verbal Reasoning (VR): 9 (52.0 - 67.7%)
Physical Sciences (PS): 12 (88.8 - 95.2%)
Writing Sample (WS): R (83.8 - 94.3%)
Biological Sciences (BS): 12 (88.9 - 95.5%)
Total Score: 33R (87.8 - 91.1%)

Verbal Reasoning (VR): 13 (98.3 - 99.7%)
Physical Sciences (PS): 12 (88.8 - 95.2%)
Writing Sample (WS): Q (62.9 - 83.7%)
Biological Sciences (BS): 11 (77.1 - 88.8%)
Total Score: 36Q (95.8 - 97.3%)

 

[Don Draper]

Correct me if I'm wrong. Isn't a 12 in PS always supposed to be the same percentile? And so on.

I think the argument in the other post was that scoring say a 50/52 would be a 14 on one exam and 15 on another exam. I don't think the argument was a 14 or a 13 is a different percentile on different test days.

The other part of the argument was that scoring a 52/52 would not result in a 15 on some exams, but only a 14. I believe the # by definition is associated with a percentile (in the specific section or cumulative).

 

[moto_za]

Do med schools see or look at the percentiles??

 

[BaronVonZ]

Correct me if I'm wrong. Isn't a 12 in PS always supposed to be the same percentile? I believe the # by definition is associated with a percentile (in the specific section or cumulative).

You may be correct, thats exactly what I'm trying to sort out. A lot of people in the 7/16 thread were complaining that their test scores were unusually low, if we can see that the percentile scores were the same as other administrations then we well be able to discredit that assumption.

 

[Stahs]

I have been wondering this myself and I thought that since I have gotten a lot from this forum that it was about time I made an account and contributed something. I added both my MCAT breakdowns (8/20/2009 and 6/17/2010) and it appears that the percentiles are different. It seems that either my first test from last year was harder (same number mark in a section = higher percentile back then) or that the percentiles are being updated over time as more people take the test and people are doing constantly better (same mark = lower percentile over time). Note: I took the pm 6/17 test and the percentiles are the same as you guys. If one of you took the am it would either mean that the percentiles are based on the whole day (rather than one for each time) or that they are being updated over time as a total test taker pool.

Correct me if my logic is flawed though.

8/20/2009 Test
===================

Verbal Reasoning (VR): 09 (52.8 - 69.0%)
Physical Sciences (PS): 11 (82.0 - 90.8%)
Writing Sample (WS): Q (66.6 - 84.8%)
Biological Sciences (BS): 11 (77.0 - 89.1%)
Total Score: 31Q (80.5 - 85.0%)

5/1/2010 Test
===================

Verbal Reasoning (VR): 15 (99.9 - 99.9%)
Physical Sciences (PS): 13 (95.3 - 97.3%)
Writing Sample (WS): S (94.4 - 99.3%)
Biological Sciences (BS): 13 (95.6 - 97.6%)
Total Score: 41S (99.8 - 99.9%)


6/17/2010 Test
===================

Verbal Reasoning (VR): 9 (52.0 - 67.7%)
Physical Sciences (PS): 12 (88.8 - 95.2%)
Writing Sample (WS): R (83.8 - 94.3%)
Biological Sciences (BS): 12 (88.9 - 95.5%)
Total Score: 33R (87.8 - 91.1%)

Verbal Reasoning (VR): 13 (98.3 - 99.7%)
Physical Sciences (PS): 12 (88.8 - 95.2%)
Writing Sample (WS): Q (62.9 - 83.7%)
Biological Sciences (BS): 11 (77.1 - 88.8%)
Total Score: 36Q (95.8 - 97.3%)

Verbal Reasoning (VR): 11 (83.5 - 95.4%)
Physical Sciences (PS): 12 (88.8 - 95.2%)
Writing Sample (WS): Q (62.9 - 83.7%)
Biological Sciences (BS): 11 (77.1 - 88.8%)
Total Score: 34Q (91.2 - 93.7%)

 

[phltz]

They are not giving you your ranking with respect to your test date. They're not even giving you your ranking based on your year - they don't yet have the info for your whole year when they give you your report. They're giving your percentile based on the previous years' distribution. The guy with the 2009 test has his percentiles given according to this distribution, while the 2010 scores are given according to this distribution.

 

[NickNaylor]

If you compare the two charts from the above side-by-side, they are very similar. In theory the scores should be scaled such that they're somewhat normalized, however I'm sure there is some variation. I'm assuming that's why most schools won't accept scores that are 2-3 years old.

 

[JSU]

8/20/09

Verbal Reasoning (VR): 09 (52.8 - 69.0%)

VS.


6/17/10

Verbal Reasoning (VR): 9 (52.0 - 67.7%)

Odd, I thought the percentages were always the same for a certain number score. I guess it changes from year to year (like above).

 

[moto_za]

8/20/09

Verbal Reasoning (VR): 09 (52.8 - 69.0%)

VS.


6/17/10

Verbal Reasoning (VR): 9 (52.0 - 67.7%)

Odd, I thought the percentages were always the same for a certain number score. I guess it changes from year to year (like above).

2008 Test date

Verbal Reasoning (VR): 08 Percentile Range: 39.5 - 53.3

 

[BaronVonZ]

Can anyone confirm that percentiles are consistent through the year? All we need now is someone from 2010...

 

[KickedOut]

The scaled scores for each section are determined prior to anyone actually taking the test. Example: they determine that missing 5 questions in biology is a 13, so everybody who takes that test and misses 5 questions in the biology section will get a 13.

Percentages are then applied ex post facto. The only reason that there should be any variation in the percentages from one test to the next is that there is now a larger sample size to calculate percentages with. That's why there is a range and not an exact percentage given.

If you purchase "The Official Guide to the MCAT Exam" all this information is provided.

 

[phltz]

Percentages are then applied ex post facto. The only reason that there should be any variation in the percentages from one test to the next is that there is now a larger sample size to calculate percentages with. That's why there is a range and not an exact percentage given.

Wrong and wrong. Percentages are all calculated on the same sample, and thus the same sample size. They are calculated using the data from the previous year. The fact that there's a range is due to the fact that multiple people get the same score. If you got, say, a 30 in 2010, it will be described as a 73.8-79.0 percentile. This is because 5.3 percent of 2009 test takers (that's about 4200 students) got exactly a 30. The range has nothing to do with uncertainty in the calculations. It has to do with the fact that you aren't the only person to get that score.

 

[UnclePhil]

The scaled scores for each section are determined prior to anyone actually taking the test. Example: they determine that missing 5 questions in biology is a 13, so everybody who takes that test and misses 5 questions in the biology section will get a 13.

Percentages are then applied ex post facto. The only reason that there should be any variation in the percentages from one test to the next is that there is now a larger sample size to calculate percentages with. That's why there is a range and not an exact percentage given.

If you purchase "The Official Guide to the MCAT Exam" all this information is provided.

That completely eliminates the standardization of a standardized exam. If that were the case, then we would be comparing our percentages instead of our scaled scores (which would essentially be meaningless).

 

[ModerateMouse]

8/20/09

Verbal Reasoning (VR): 09 (52.8 - 69.0%)

VS.


6/17/10

Verbal Reasoning (VR): 9 (52.0 - 67.7%)

So, essentially if you took both of the above tests and scored better than 68% of the other test takers on the verbal each time you would have got a 10 in 2010 and only a 9 in 2009!

That's a huge difference, given that you have the exact same ranking.

(especially given the psychological difference between a 9 and a 10)

 

[poohbear12345]

I got a 36 on the 2011 Exam and its a 95.8-97.3 percentile as well. Looks like its fairly consistent