USMLE Scores

[automan]

---

Members don't see this ad.

Is your three digit score the number out of 350 that you got correct?


If not, how to they come up with the three digit score?

Also, any idea how many you need to get correct to get an average score of say, 220?

 

[lord_jeebus]

It's a big mystery.

Anyone who responds with a definitive answer either made it up, or heard it from someone who made it up.

But it's not the number out of 350, because some questions are experimental, and because you can't score over 300.

 

[automan]

So let's say, there are 50 experimental questions. Then it's your score out of 300?

I am just guessing, but it seems reasonable. A 270 = 90% It doesn't seem that crazy.

When they used to give out the percentile information, did they ever give out the actual percentages people got correct?

 

[ddmo]

I doesn't work they way you are describing because that would screw over the people who received more difficult questions. They most likely take the given set of questions, compile the list of results, assign the mean a numerical value around 218, then determine the SD. So your score is a representation of your performance against other students who answered the same questions. This means that even a perfect score would not always equal out to the same 3 digit score.

I believe when I took the exam, the avg. was 218, with an 18 point SD. That gives a pretty good idea of the percentile.

 

[automan]

Thanks, that helps a lot.

 

[genX]

see if this short cut works.

http://www2.uic.edu/stud_orgs/hon/aoa/MedStudentsAmnesia.doc

 

[genX]

posted here by accicent but that is link to 50 pages of high yield board words

 

[fakin' the funk]

But it's not the number out of 350, because some questions are experimental, and because you can't score over 300.

I thought that the reason they can't/don't provide percentiles is because there is no "perfect" score, i.e. there is no ceiling that would be the 100th percentile.

 

[fakin' the funk]

So let's say, there are 50 experimental questions. Then it's your score out of 300?

I am just guessing, but it seems reasonable. A 270 = 90% It doesn't seem that crazy.

Good idea, but have you met someone who could probably get in the 260-270 range? They would probably get 98% or higher correct. 90% would be insulting to them! Freakin geniuses!

 

[automan]

No ****, I thought 70% was around the average of 220. We had at least two people at my school >270 last year, and I figured around 90% would get you that. I think I can write anything over 230 off right now. It must be a pretty steep curve

 

[lord_jeebus]

I would guess that 270 is about 90% correct on a test of average difficulty, but it's hard to say...

 

[guitarguy09]

I once heard from someone that it really isn't a matter of direct percentage. Since things are weighted from everyone who took your same test, it works out a little differently. For example, I heard that when going from 215 to 225, it requires more correct questions than when going from 225 to 235 (less questions are required to distinguish between test takers at the higher scores because less people will be in that range). Obviously, it's not a linear curve. So separating the "elite" from the "extremely elite" is probably only a matter of a couple Qs.

Disclaimer: who knows what the "real" truth is... and who really cares, just do your best and good luck!

 

[toothless rufus]

Disclaimer: who knows what the "real" truth is... and who really cares, just do your best and good luck!

So, they don't explain how the test is scored? Why be secretive?

 

[mesoderm]

So, they don't explain how the test is scored? Why be secretive?

Makes you wonder...

 

[Skills of House]

Well, according to Kaplan, it's not at all weighted. And it's not performace based like the GRE where if you get a question right, then the next one is harder. There's a pool of questions pre-selected before you take the exam and 50 questions are randomly selected for each block, so it may appear that the test is getting harder or easier at certain times, but it's still random, not systematic.
He also said that the popular belief about scores being influenced by who takes the exam at the same time as you is a myth. Your score is your score regardless of how well or how poorly anyone in your cohort performs.
....admittedly he also couldn't tell us an actual "maximum" score though....

 

[spinaldoc]

I have a friend who said he scored in the 97%. When a person recieves their scores does it give your % you scored on average with the others who took the exam?

 

[fakin' the funk]

I have a friend who said he scored in the 97%. When a person recieves their scores does it give your % you scored on average with the others who took the exam?

Why don't you ask your friend.

Seriously, is this forum on a 1-week cycle?

 

[spinaldoc]

The friend i spoke with i have lost contact with. You know you don't have to read threads that seem to be repetitive to you.