What % correct needed to get 250+

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

m3shocker

Junior Member
15+ Year Member
Joined
Jan 26, 2004
Messages
9
Reaction score
0
Hello.

I am a M2 and asking all those who rocked step 1 this question:

What % of the questions on Step 1 do you need to get correct in order to get a 250+? I assume it would be in the 85% to 90% correct raw- all depending on the mean and distribution.

Thanks
 
I think if you get 75% correct = 230 and 85%= 250, anyone is free to correct me , if I am wrong.😀
 
There is absolutely no one that can answer this question for you, only the NBME knows. All we do know is that you need to answer 55-65% right to pass, and that they usually throw out 30-50 questions per test.
 
the above poster is correct.....NOBODY knows the answer to that.......NOBODY.

Our micro professor told us last semester when we took the NBME shelf exam that he had know idea how to interpret the raw score and how to correlate it to the national percentile etc....
he said over and over to our class that it is a "super secret" thing with NBME and no one knows how many "test" questions there actually are etc....


theories are always interesting about how people think it is graded.

later
 
Originally posted by 12R34Y
the above poster is correct.....NOBODY knows the answer to that.......NOBODY.

Our micro professor told us last semester when we took the NBME shelf exam that he had know idea how to interpret the raw score and how to correlate it to the national percentile etc....
he said over and over to our class that it is a "super secret" thing with NBME and no one knows how many "test" questions there actually are etc....


theories are always interesting about how people think it is graded.

later

True enough, no one except the NBME, knows how the darn thing is scored, though theories abound. There is a conventional wisdom, however, and it goes something like this. If yours differs from mine, please fill me in.

- passing is usually estimated to be around 55-65%, this is from Kaplan, the designer's of Q-bank

- questions are apparently "randomly" selected within categories from a huge database from which there are many variants

-some of these are experimental and/or total duds (Kaplan estimates around 50), which leaves 300 questions

- current consensus seems to be that adaptive testing is not in place (increasing difficulty if you do well, easy if you do poorly), although you hear the occasional person that suspects otherwise

-whether linear scoring (1 question 1 point) is used is open to debate, and, frankly, with NBME's secrecy is unknown; some people suspect there may be a non-linear algorithm used because some folks who got high scores in most domains but bombed 1 or 2 random categores sunk their scores big time
 
Originally posted by jed2023

- current consensus seems to be that adaptive testing is not in place (increasing difficulty if you do well, easy if you do poorly), although you hear the occasional person that suspects otherwise

True. But, if you get a question like "patient: animal or mineral" you've probably got an adaptive test and you probably aren't doing well.
 
One other thing I might add... from how I understand it, the test is scored exponentially - that is, after you hit the threshhold of 182 (passing), the # of correctly answered questions required to bump your score to the next level increases... e.g. it may take 5 correct answers to get to 187, but perhaps 8 correct answers to get to 192...

just my $.02...

-t
 
how can i know it for sure? i can't, as the NBME is locked up tighter than a tetanic jaw... i only know what i've heard from multiple administrators from multiple schools who have been dealing with the dumb*ss test for years upon years... that's why I said "from how I understand it"... not "from how I know it."

I may be wrong... but I doubt I'm all wrong.

-t

ps - possibly doesn't have an 'e'
 
The dude was just offering his thoughts. I don't think everybody has to start off their statements with long qualifying statements like," Well, it is within the realm of possibility that. . ."
 
could you be any more sensitive?

all I was pointing out that all of these "ways NBME grades" are theories. nobody knows and people keep chiming in about this means that and such.....nobody knows.

oh and thanks for correcting my spelling. i usually spend all of 2 seconds responding to things and there are occasional errors.
 
The USMLE is NOT adaptive-- I e-mailed the NBME a while ago, and they seemed polite. I don't know if these guys are as super-secretive as we think. If someone tried e-mailing them a couple of carefully-phrased questions about scoring methodology, they may actually offer some answers.


Date: Thu, 28 Aug 2003 16:11:22 -0400
From: Webmail <[email protected]>
To: <deleted>
Subject: RE: computer adaptive testing for Step I

Currently, we are not actively discussing or striving towards an adaptive
model of any kind. We are seeking improvement in other areas. However, we
continue to be interested in the concepts of CAT. We may start thinking
again about it in the future. But for now, there are no plans to go to an
adaptive model.

Thanks.
 
Top