Highest score on step 1?

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

stlblues

New Member
10+ Year Member
15+ Year Member
Joined
Jan 20, 2005
Messages
7
Reaction score
0
So I just got my scores back today, and I was wondering what the highest score is on Step 1 - I haven't been able to find it anywhere.

Good luck to all of you!
 
stlblues said:
So I just got my scores back today, and I was wondering what the highest score is on Step 1 - I haven't been able to find it anywhere.

Good luck to all of you!

1 point less than yours (hopefully).
 
stlblues said:
So I just got my scores back today, and I was wondering what the highest score is on Step 1 - I haven't been able to find it anywhere.

Good luck to all of you!
If this was the first question that came to your mind when you saw your scores, I'd say you're safe for any specialty you choose. Congratulations.

In regards to your question, that information isn't published by the NBME. I can't imagine a score > 325, however. 😉
 
stlblues said:
So I just got my scores back today, and I was wondering what the highest score is on Step 1 - I haven't been able to find it anywhere.

Good luck to all of you!

Shouldn't bigfrank know?
 
I thought there was a range given like 144-288 or something like that, but I guess even that says "most" scores will fall in this range.
 
How many questions do we have to get right to get a 230??? Does that mean that we answer 230 questions correctly or is it some sort of percentage?
 
No one knows this. It is all tightly kept secret.

There are tons of theories.

My personal one is 350 ?. 50 ? thrown out.


Raw score is score. I think this makes sense according to most people's anecdotes.



I felt like i missed a ton and yet still did fairly well. Granted there is recall bias. Howvever if they throw out 50 and you miss 50, you got a 250...





When i took it, the range reported was 144-288.

Let me know if you find out the secret...
 
Does anyone know which types of questions they throw out and why?? I took it today, so it won't do me any good but I swear there were some drug questions on there that were a little specific even for pharm.
 
Doc Ivy said:
Does anyone know which types of questions they throw out and why?? I took it today, so it won't do me any good but I swear there were some drug questions on there that were a little specific even for pharm.

Specific how?
 
Wrigleyville said:
Specific how?

Well one question asked about gentamicin showed a picture of the nephron and asked where it caused toxicity-I guessed the proximal convoluted tubule which it turns out is right, but I just don't recall ever reading about it, but maybe it was in my brain somewhere

Hey good luck G 😉
 
How many questions do we have to get right to get a 230??? Does that mean that we answer 230 questions correctly or is it some sort of percentage?

no.

each question has it's own performance profile, and the test is normalized to those profiles for each test. therefore, if you get a 'harder' test (objectively, accord to the aggregate performance profile of the test) you can get more wrong than someone with an objectively 'easier' test. the variation between the extremes of the aggregate performance profile isn't known to outsiders. the %correct/score correlation is almost certainly non-linear-especially at the extremes-though in some parts of the curve it may look linear.

there are not exactly 50 experimental questions. there is variance in the number you can get, with 50 probably being the average number of 'experimental' questions that are sampled out of the large original pool of questions (probably ~5000-10000 questions, but just a wide-ranged guess, probably not more than that). your whole test is sampled randomly from the original pool, so on average you should get ~50.

the 'experimental' questions are most likely not simply 'wierd tester questions', but questions they want to accrue more 'performance data' on before 'releasing' them as a regular question. if 'experimental' questions have a dismal performance profile, being either too easy or hard according to whatever variance cut-off they have, they will probably be screened out reasonably quickly.

they only use 1 years worth of performance data to calculate your score, so it doesn't matter when you take the test, and you are normalized to what is approximately your cohort.
 
Top