Thank you! So all *non-experimental* items are weighted equally. (There is sometimes some misinformation that circulates saying that certain questions are worth more than others.) But experimental questions aren’t scored. And there are a lot of them: ~25% of the questions on a typical version of the test. How many USMLE questions are experimental items? ruclips.net/user/shortsFCuB-mxeYDg?feature=share
Good lecture doctor, thank you. As Step 1 switches to P/F, many students who predictably will score more than average tend to take the exam before Jan 26 and will get higher scores, and those who probably would get lower scores take the exam after Jan 26. So, my question is if the score generates by comparing applicants who took the same test in the same period of time, can we suggest that taking the exam in late 2021 will put the applicants in harder contest and more difficulty to score high?
Great question - and no, you don’t need to worry. Remember, the reference group is a group of *previous* first-time USMLE test-takers from U.S. MD schools. I think you are correct that a more accomplished group of international medical graduates will choose to take Step 1 during the final months before pass/fail score reporting - but the performance of these test-takers is not determined versus each other, it’s determined versus the reference group. I discussed this a little bit in a recent Mailbag: thesheriffofsodium.com/2021/08/13/mailbag-second-looks-the-usmle-step-1-curve-and-residency-placement-rates-for-dos/
when you say 'same form of the exam', is it that alot of people have already the exact same examination which was either difficult or easy? in what way apart from difficulty is it the same?
Hi@sheriffofsodium, i am difficulty understanding as to why is mean taken as 200 and SD as 20, i mean is it just an assumption or is it supported by data?
It’s a matter of historical record. Remember, the USMLE score is a scaled score - a way of representing normative performance on a a standard scale. Most standardized tests are reported this way (SAT, ACT, MCAT, GRE, LSAT, etc.). When the USMLE began, the mean was arbitrarily chosen to be 200 and the standard deviation 20. They could have used a different scaled score - for instance, the old NBME exams used a scaled score similar to the SAT, where the mean was set at 500 and the standard deviation was 100. But they chose 200 as the original mean. (Of course, because examinee performance continues to increase, the observed mean is now much higher.)
It may be easier to follow the logic in this article than in the video: thesheriffofsodium.com/2021/05/04/breaking-the-magic-the-usmle-three-digit-score/
THANK YOU SO MUCH FOR THIS! I HOPE PEOPLE REALIZE HOW VALUABLE THIS VIDEO IS
How
this channel deserves more subs
This is insane good quality
Dude, this is amazing
But I have a question if they use the mean how come the average step 2 score is increasing over time? Thanks
This is dope 🔥 Btw as you mentioned some are experimental questions in your early videos but here you mentioned every qs carry the same marks. How?
Thank you! So all *non-experimental* items are weighted equally. (There is sometimes some misinformation that circulates saying that certain questions are worth more than others.)
But experimental questions aren’t scored. And there are a lot of them: ~25% of the questions on a typical version of the test.
How many USMLE questions are experimental items?
ruclips.net/user/shortsFCuB-mxeYDg?feature=share
@@sheriffofsodium thank you
Good lecture doctor, thank you.
As Step 1 switches to P/F, many students who predictably will score more than average tend to take the exam before Jan 26 and will get higher scores, and those who probably would get lower scores take the exam after Jan 26.
So, my question is if the score generates by comparing applicants who took the same test in the same period of time, can we suggest that taking the exam in late 2021 will put the applicants in harder contest and more difficulty to score high?
Great question - and no, you don’t need to worry. Remember, the reference group is a group of *previous* first-time USMLE test-takers from U.S. MD schools. I think you are correct that a more accomplished group of international medical graduates will choose to take Step 1 during the final months before pass/fail score reporting - but the performance of these test-takers is not determined versus each other, it’s determined versus the reference group.
I discussed this a little bit in a recent Mailbag:
thesheriffofsodium.com/2021/08/13/mailbag-second-looks-the-usmle-step-1-curve-and-residency-placement-rates-for-dos/
is percentage correct include experimental? or just 200 questions
when you say 'same form of the exam', is it that alot of people have already the exact same examination which was either difficult or easy? in what way apart from difficulty is it the same?
I just did an nbme and got a 58%. How would that translate into the three score, estimation?
Hi@sheriffofsodium, i am difficulty understanding as to why is mean taken as 200 and SD as 20, i mean is it just an assumption or is it supported by data?
It’s a matter of historical record. Remember, the USMLE score is a scaled score - a way of representing normative performance on a a standard scale. Most standardized tests are reported this way (SAT, ACT, MCAT, GRE, LSAT, etc.). When the USMLE began, the mean was arbitrarily chosen to be 200 and the standard deviation 20. They could have used a different scaled score - for instance, the old NBME exams used a scaled score similar to the SAT, where the mean was set at 500 and the standard deviation was 100. But they chose 200 as the original mean. (Of course, because examinee performance continues to increase, the observed mean is now much higher.)
It may be easier to follow the logic in this article than in the video:
thesheriffofsodium.com/2021/05/04/breaking-the-magic-the-usmle-three-digit-score/
Damn.. that’s crazy to actually know how this works..
If we give step1 exam in may, are we at advantage ? Our results will only appear in July for data collection purpose