# Correlation with USMLE World

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

#### coastal

##### Member
7+ Year Member
15+ Year Member
20+ Year Member
Anyone have a rough idea of correlation between passing Step I and score on USMLE World?

I started studying about a week and a half ago (just going through USMLE World questions), have gotten through 800 questions, and am getting 50% of them right. I write Step I in 2 1/2 weeks, and plan to try and go through about 100 questions a night until then (that's really all I have time for).

Will 50% on USMLE World get me a pass? (don't care about score).

cheers.

2.3 * (% correct in UWORLD) + 84 = predicted score.

2.3 * (% correct in UWORLD) + 84 = predicted score.
How accurate is this formula?

Are you sure about this? I used this formula but it didn't seem to match my NBME scores.. From what I have heard, the NBME exams are supposed to be very accurate for most people

Members don't see this ad :)
Are you sure about this? I used this formula but it didn't seem to match my NBME scores.. From what I have heard, the NBME exams are supposed to be very accurate for most people

I just read this equation on these forums (in multiple places). Are your NBME scores higher or lower than the equation would predict?

for reference:

UWorld.......Score estimate
50.00%......199
55.00%......210
60.00%......222
65.00%......233
70.00%......245
75.00%......256
80.00%......268
85.00%......279
90.00%......291

Last edited:
I think this formula is working if you did timed, unused, 350 question per day, like a real simulation of the exam. My wife had 70% correct and 246 on real exam.

Anyone have qbanks score correlation for kaplan?

try the clinical review usmle score calculator (just google it, for some reason SDN isn't letting me link to it)

2.3 * (% correct in UWORLD) + 84 = predicted score.

that can't be right. my USMLEworld score is 88 (80%), with 20% unused. if i do unused q's, my score is 70-80. if i do used questions (all), or incorrect q's, its usually 90-100%. i take the test march 1st, so we'll see how this all pans out.

You'll crush the Step 1!

Seriously? has that formula worked for anyone you know? i'll definitely report back when i get my score (unless its horrible).

I know the third years here who were getting ~75 ended up with 240s. I want a 240.

Good luck.

back to studying for me.

that can't be right. my USMLEworld score is 88 (80%), with 20% unused. if i do unused q's, my score is 70-80. if i do used questions (all), or incorrect q's, its usually 90-100%. i take the test march 1st, so we'll see how this all pans out.

...you are missing the point. Of course the equation doesn't work if you redo your answers, lol.

I know I'm resurrecting an old thread here, but are these estimations still accurate?

I know I'm resurrecting an old thread here, but are these estimations still accurate?

There are so many variables that formula has no way of taking into account..... It is worthless. Stick to NBME & UW exam averages.

There are so many variables that formula has no way of taking into account..... It is worthless. Stick to NBME & UW exam averages.

that guy who averaged an 80% ended up with a 245 (looked at his history)... per the formula, he should have gotten 268! so yeah, the fomula does not work... there's almost no way to tell the future besides studying ur buff off and getting lucky on the exam day!!

1 user
At that end of the bell curve, there is very little that separates a 245 from a 265 in terms of knowledge base. If you take the exam one day earlier you might get a 265; one day later you get a 245.

Even still, I know a couple who studied everything together for Step 1, always got within 3 pts of each other on NBMEs and within 2% on UWorld. They took the test on the same day -- the one scored 23 pts higher on the real thing. It's all about the luck of the draw with questions. Your exam will have 322 Qs, and that's just a drop in the bucket in terms of how many pieces of information you've learned. So it's not entirely representative. Just study your butt off and don't worry about a stinking formula.