Estimating your USMLE Score's percentile?

[wamcp]

Anyone know how to figure out your true national percentile rank based on your 3 digit score?

Based on the following graphs I've found:

Table estimate of USMLE score

Non-normal bell curve of USMLE scores

It seems like it is definitely NOT a normal bell curve.
_________________________________________
Here is a normal bell curve:

USMLE normal bell curve
_________________________________________

Let's assume you score 250, with a test mean of 224 and standard deviation 22.

If you use the normal curve, you calculate that you are 88.1 percentile.
If you use the non-normal curve, you get 94.52 percentile. Which seems like a pretty big difference to me.

Any thoughts? How accurate is all this?

---

Members don't see this ad.

 

[Phloston]

My score report, from December, said that "for recent administrations, the mean and standard deviation for first-time examinees from American and Canadian medical schools are approximately 224 and 22, respectively, with most scores falling between 140 and 260." Apparently, for recent test-takers, the mean and SD are now 227 and 22.

Americans/Canadians score higher on the exam, so the 97.5% for them is currently 271. Sounds ludicrously high, but they only compose ~1/2 of the applicants to the match. You also have to consider that most of the best test-takers are at US MD schools, rather than overseas, and that MD schools' curricula are best geared toward students excelling on the exam.

When taking into account / pooling the numbers from US MD schools, combined with second-time testers, DOs, IMGs, current physicians, etc., who take the exam, 97.5% is still around 260 for all applicants to the match.

Therefore, in terms of what a residency PD actually sees, a 260 is still perceived as +2SD.

My guess is that the "140 to 260" range listed on the score report refers to the bottom 97.5%, rather than +/-2SD (which would be 95% of test-takers).

The net curve is positively skewed (left-shifted).