whats the highest step1 score possible?

[surething1]

I hope I don't get flamed for this but what is the highest step 1 score possible? three digit score obviously. thanks. also whats the highest you have heard of someone getting?

---

Members don't see this ad.

 

[Bones2008]

I don't think anyone actually knows.

 

[Idiopathic]

I hope I don't get flamed for this but what is the highest step 1 score possible? three digit score obviously. thanks. also whats the highest you have heard of someone getting?

Lets say 285. I heard of a 272. Im sure someone else here has seen a higher one.

 

[bigfrank]

Lets say 285. I heard of a 272. Im sure someone else here has seen a higher one.

Yea, if you get over a 270, you can say whatever you want!

 

[chillypalmer]

For me its 206!!!
If only I had kept my brain in the refrigerator, and not on the window sill, where chugga the neighbourhood crow was sitting, I would have had a brain and scored like 272.Oh, I hate brain eating crows and any score over 206!!!

ps...Chugga is a research scholar at Stanford today.He is the first crow from the east coast to do that.

 

[p53]

The Step 1 range from Jan-Dec 2005 was 142-280. Thus, it is likely the highest possible score is a 280. With such a large sample size I believe the person that scored a 280 maxed out.

 

[vtucci]

Someone at my school got a 281.

 

[ice_23]

Someone at my school got a 281.

Apparently 299 is the highest score as relayed by an NBME representative at an AAMC meeting I once attended.

-Ice

 

[(nicedream)]

LOL, can you imagine getting a 299? Wtf. I wonder what kind of % that would be.

 

[BlondeCookie]

LOL, can you imagine getting a 299? Wtf. I wonder what kind of % that would be.

Maybe 100%. Then again, 299 sounds like such a peculiar number. Why wouldn't it just be an even 300?

 

[DrPak]

Maybe 100%. Then again, 299 sounds like such a peculiar number. Why wouldn't it just be an even 300?

Two ninety nine sounds more impressive. It takes longer to say...

 

[Idiopathic]

Keeping in mind its likely a theoretical score, that 299.

 

[Tristero]

So much bull**** goes on about STEP 1 scores. According to NRMP Data from 2005 the amount of applicants to Derm, Plastics, Ortho, Rad Onc, and Rads with greater than 270?

ZERO. Unless all of the 270+ kids are applying early match to neurosurgery or optho, and considering there were only 14 kids in country with scores from 260-270 that applied to Derm, and only 4 to plastics--it seems to me that 270+ is extremely unlikely, and surely not at the frequency that I see thrown around all over this board.

281? i'm not saying its impossible, but that kid would have to be the smartest/best test taker kid out of all the medical schools in the country. That includes HMS, Johns Hopkins, UCSF, Baylor, WashU, etc etc. Thats ALOT of smart kids he's crushing.

Just my take. I could be wrong, but I was wondering what you guys think about that.

 

[Iwy Em Hotep]

So much bull**** goes on about STEP 1 scores. According to NRMP Data from 2005 the amount of applicants to Derm, Plastics, Ortho, Rad Onc, and Rads with greater than 270?

ZERO. Unless all of the 270+ kids are applying early match to neurosurgery or optho, and considering there were only 14 kids in country with scores from 260-270 that applied to Derm, and only 4 to plastics--it seems to me that 270+ is extremely unlikely, and surely not at the frequency that I see thrown around all over this board.

281? i'm not saying its impossible, but that kid would have to be the smartest/best test taker kid out of all the medical schools in the country. That includes HMS, Johns Hopkins, UCSF, Baylor, WashU, etc etc. Thats ALOT of smart kids he's crushing.

Just my take. I could be wrong, but I was wondering what you guys think about that.

It looks like 1 person with 270+ applied to Ortho, but your point still stands.

 

[Tristero]

It looks like 1 person with 270+ applied to Ortho, but your point still stands.

Duly noted. It looks like 1 person applying to ortho and 1 person applying to rads had >270. Like you said though, my argument still makes sense. >270 is a really, really rare event, and I doubt it occurs with the frequency that is claimed on this board. Its nice to finally see some hard data on this subject instead of being uncertain and relying on anecdotal evidence.

But I could be wrong.

 

[lord_jeebus]

Duly noted. It looks like 1 person applying to ortho and 1 person applying to rads had >270. Like you said though, my argument still makes sense. >270 is a really, really rare event, and I doubt it occurs with the frequency that is claimed on this board. Its nice to finally see some hard data on this subject instead of being uncertain and relying on anecdotal evidence.

But I could be wrong.

How many claims of 270+ have you seen around here? I can think of one. I may have seen 1 or 2 people claiming to know someone else who got 270+, but when you open things up to everyone known by someone on SDN that's a good percentage of the med school population.

 

[Time]

It looks like 1 person with 270+ applied to Ortho, but your point still stands.

Just checked the NRMP Data from 2005. The version I HAVE - unless there is more than 1 version circulating - says the highest score for diagnostic radiology applicants was "<=270" (for 1 applicant), which I take to be LESS OR EQUAL to 270. While several other (but not all, and not ortho) specialties had this score distribution, NONE of the specialties had ">=270" - which I understand to be 'greater than or equal to' 270. There is space allocated for >=270, so it may very well exist, but with anything above 265 being in the top 2% nationwide, I doubt that there are too many of 270+s or 280+s around. Step 2 forum seems to have more of 270+s, though....

I downloaded the PDF file that I have from a link posted some time ago on this forum.

 

[Tristero]

Just checked the NRMP Data from 2005. The version I HAVE - unless there is more than 1 version circulating - says the highest score for diagnostic radiology applicants was "<=270" (for 1 applicant), which I take to be LESS OR EQUAL to 270. While several other (but not all, and not ortho) specialties had this score distribution, NONE of the specialties had ">=270" - which I understand to be 'greater than or equal to' 270. There is space allocated for >=270, so it may very well exist, but with anything above 265 being in the top 2% nationwide, I doubt that there are too many of 270+s or 280+s around. Step 2 forum seems to have more of 270+s, though....

I downloaded the PDF file that I have from a link posted some time ago on this forum.

Time, where did you get that 2% number for greater than 265? just curious. it seems to me that 2% would give much greater numbers than is shown on the NRMP data for 260-270 range in all specialties...

I think I have the same PDF that you do from the NRMP website, I first thought there was no one with
>270 in rads or ortho, but if u look closely, it shows a 1, not a 0. But maybe I am misreading it?

 

[Tristero]

When I did my calc for 258, it worked out to be around the 97th percentile (i.e. 3% of scores were equal to or higher), which means that likely at least 2.5% of scores were 260 or greater, which accounts for about 500 people (assuming 20K test takers). Thats quite a bit actually, and leaves room for people to outlie in that 270+ area, but I think 280 is pusing it for sure

\If thats based on standard deviations that cant be correct bc it is doubtful that the scores fall on a normal bellcurve...but i dunno its all good i think we have beaten this topic into the ground lol

 

[Idiopathic]

\If thats based on standard deviations that cant be correct bc it is doubtful that the scores fall on a normal bellcurve...but i dunno its all good i think we have beaten this topic into the ground lol

Its not doubtful. in fact, its calculated based on the standard curve. admittedly they dont always fall that way, but its close enough to estimate, and Ill argue that the 3 digit/2 digit scores are based completely on the standard distribution.

 

[Tristero]

Its not doubtful. in fact, its calculated based on the standard curve. admittedly they dont always fall that way, but its close enough to estimate, and Ill argue that the 3 digit/2 digit scores are based completely on the standard distribution.

thats not right, check earlier threads. the score distribution is far from standard. if you look at the nrmp data there are prolly round 150/160 >260 scores, just eyeballin it. Unless ~350 people w/ >260 are applying to ENT, ophtho, and neurosurg which is highly unlikely, and if your calculations are correct based on standard deviation, the distribution is far from standard.

 

[inshanesworld]

Not to repeat someone's post but the highest score possible is indeed 299-300. This test is graded on the number correct so if you do not miss the actual graded questions (you can miss all 50 of the experimental cause they don't count) you win. This was the original question and while we can chat about what this years highest score was but it really is an effort in futility.

You may say to yourself that while 299-300 is theoretically possible, short of hell freezing over no one will every get that. However, I am aware of a nice gentleman who is the only person in history to get a perfect score on the pathology board exams (the ones that make you a board certified pathologist). No one before him or after has accomplished this feat. Remember every once in a long while a savant is born. The better question about related to this thread is what benefit is gained by knowing this number?

 

[lord_jeebus]

This test is graded on the number correct

Is this your speculation or do you have a source? (I also think it is likely scored this way but have found no evidence to support my theory.)

 

[inshanesworld]

I have had this relayed to me by sources I trust. So you can consider it partially ancedotal. However, the USLME website hints at this as well in the document explaining the discontinuing of percentile information.

http://www.usmle.org/Scores/percentiles.htm

There are several key phrases that lead to this as a very possible conclusion

In other words, a score of 200 on one administration of a Step indicates the equivalent level of examinee performance as a score of 200 on any other administration of the same Step. This equivalence holds even if the pass-fail standard is changed, which permits comparing performance across time.

As a result, identical three-digit scores are associated with significantly different percentiles if those percentiles are based upon the examinees from the different periods.

Which student had better performance on Step 1? Both students scored 210, so their performance was identical.

So in what way could you ensure that a 210 score in 1980 = a 210 in 2006? Seems crazy as we have to know so much more about so many things due to natural advancement of medicine. You do this by weighing the difficulty of the individual questions as opposed to weighing the entire test itself. The statistics on this are beyond this discussion. But you create a test that is normalized before the test takers even take the exam. This allows you to use the absolute score as a normal predictor of ability.

The second quote about the acknowledgement that the same three digit score could have a different percentile rank from year to year is very telling. In essence the NBME is admitting that there is nothing precluding ever test taker from scoring a 260. A 260 is still a 260 but now the average has shifted and it is now the 50th percentile. So if every test taker can get a 260 then every test taker can get a perfect score. 260, 230, 186 or whatever represents the same level of achievement as every other person who has ever taken the exam throughout history. This is a serious implication that your score represents actual number of questions correct.

This reply is too long so I will stop and let you all read the link and make your own implications. (Though I do reserve the right to retain the idea the the test is graded by an old man with a magic eight-ball and Dion Warwick.)

 

[Reaganite]

I stopped reading SDN for info on Step 1 scores. Reading the scores here, even if you scored >260, you feel like you haven't accomplished anything out of the ordinary. The NRMP data were the best thing to hit SDN because it became very clear that either: 1. people are lying on SDN or 2. that SDN is not a representative sample of the general applicant pool. Regarding high scores, the realeased data show that only a handful of people score above 270. The year I took the exam, the highest score was listed as 288. Given that a very, very small number of people even score above 270, it is likely that this person was an extreme outlier.

 

[t33sg1rl]

It seems to me that a gifted statistician should be able to figure this out. If you remember back to the MCAT, to move a section score from 10 to 11 required getting an extra 10 or so questions right, but moving your score from 13 to 14 only required getting another two or three questions right-the thing was weighted towards the middle, so that a LOT of people scored 28-32. If Step 1 is NOT weighted like that, then it seems to me that could be teased out somehow.

I'd like to offer a 2nd conspiracy theory-the reason that they won't admit that your score corresponds to the number you got right is that they would then have to admit how many "experimental" questions are on each test, plus then everyone will realize how silly it is to prefer a residency applicant with 225 over one with a 220. Five Q's on a daylong test??? C'mon! I feel in my heart that there are probably 50 experimental questions and 300 real ones. A score of 215 would be 72% correct, which seems like a reasonable average to me, and 300 would be the theoretical upper limit.

 

[Time]

Hate to restart this, but I had to actually recheck the PDF that I have. I stick with numbers I posted in my first post. I do see 'a 1, not a 0' in some cases, but not rads, or ortho... (yes, 2005 data!!!!)

Why 98%ile ....2x std. dev.+mean --> assuming normal distribution.

I do agree that there is always someone who is better than others in almost anything, but saying that there must be someone hitting a 299 based on the fact that he is an exception is, maybe, like saying that an ueber-exceptional human can fly given that the olympic champs can jump so high. if you were to analyze extreme performances, you would be sure to find that the 'enigmatic leader' is always (very closely) followed by a huge number of competitors ready to win the moment he blinks and loses focus. whether in sports (most obvious), or in anything else....

given the limited time we are exposed to pretty much the same information, i doubt that it would be possible for anyone to be SO MUCH better than 99% of the population. I would guess that most guys scoring 3std dev above the mean do so by just a point. so far we have seen tangible evidence of 'only' 268 here, with most notable 267 by Bigfrank (thanks for all your salient advice here, bro!). someone did post 272, but that person was not a regular here, and as far as i am concerned may have been off by a 100 points. also, someone getting a perfect score on pathology boards is great, but if the next highest score for generations had been the score just a point or so lower, the new record would hardly matter. could have been luck in this case, (less luck for all others ) - or just the fact that with the advent of computer technology, better books, review material, etc.... it is easier to access more information than in the past.

and of course, i am guessing as much as y'all!

Time, where did you get that 2% number for greater than 265? just curious. it seems to me that 2% would give much greater numbers than is shown on the NRMP data for 260-270 range in all specialties...

I think I have the same PDF that you do from the NRMP website, I first thought there was no one with
>270 in rads or ortho, but if u look closely, it shows a 1, not a 0. But maybe I am misreading it?

 

[velo]

Keeping in mind its likely a theoretical score, that 299.

exactly. To get that you'd probably need both 100% correct and a low mean %correct from the other test takers

 

[exmike]

It seems to me that a gifted statistician should be able to figure this out. If you remember back to the MCAT, to move a section score from 10 to 11 required getting an extra 10 or so questions right, but moving your score from 13 to 14 only required getting another two or three questions right-the thing was weighted towards the middle, so that a LOT of people scored 28-32. If Step 1 is NOT weighted like that, then it seems to me that could be teased out somehow.

I'd like to offer a 2nd conspiracy theory-the reason that they won't admit that your score corresponds to the number you got right is that they would then have to admit how many "experimental" questions are on each test, plus then everyone will realize how silly it is to prefer a residency applicant with 225 over one with a 220. Five Q's on a daylong test??? C'mon! I feel in my heart that there are probably 50 experimental questions and 300 real ones. A score of 215 would be 72% correct, which seems like a reasonable average to me, and 300 would be the theoretical upper limit.

In order for the UMSLE to be normalized to score from year to year, it must be a normal distribution that can be adjusted. As such, it is highly unlikely that your score is exactly what you got in terms of # correct. More likely than not it follows a similar pattern w/ the MCAT where around the mean you have to get many more correct to raise your score (thus making a 230 much better than a 220) whereas the diff between a 270 and 260 is pretty meaningless.

 

[inshanesworld]

I think it is important here to reiterate. The USMLE is NOT normalized. We should not be throwing out statistical terms without understanding what they mean. The usmle is not a normalized test and for good reason.

Anything normal in statistics means that it is fit to a normal gaussian distribution or is washed to fit it. Thus there can only be a certain percentage of people that get each score. You have had classes like this where there are 4 A's, 12 B's 40 C's etc. The percentage need to pass a class like that varies from year to year based on the number of students in the class and their respective spread of scores. So if 10 people got above 95 and you have a 94 you would get a B and be pissed about it. The MCAT is normalized. The number of correct answers needed for each score level varies based on the quality of the students in the sample pool.

The USLME, however, IS NOT A WEEDING TEST, it is a licensing test, or a pass/fail. It is designed to be sure you have achieved a certian level of knowledge to be a doctor. If it were normalized to a curve it would mean that in one year the number of Q's you need right to pass would change from year to year. Could you imagine this?!? Imagine going to the DMV and thinking you could miss up to 6 and still be licensed to find out at the window that because of a better quality of testtakers that year you could not miss more than 4. Absolutely not. Why would anyone make a licensing test were the doctors of 2003 only had to get half of them right to pass but the ones in 2006 had to get 80% right because there were more smart people and it had to fit a gaussian curve. That is STUPID.

It is for this reason that there are NO percentiles given for the test. Because it is not, nor designed to be, normal. If it were there would not be an issue.

It is a very poor understanding of statistics and the purpose of the USMLE that people try to believe there is some normalization that occurs. So how do they update the test to account for the increased success of students through the years. Very simple if you have any experience in writing exams. They test/rank the individual questions, mix them in such a way to maintain standards, and take a chance that their mix of easy and hard questions will do the trick.

More on this in my next post

 

[inshanesworld]

Lets remember that when you compare the USMLE of 1996 and the one of 2006 there will be considerable progression of student success. However, the difference between the success of students in 2005 compared to 2006 will be incremental. How do you measure this change (+ or -) from year to year. One way is to populate your test with questions that are non-graded (for research only) and use statistics to see what the results were. I suspect that this is why the experimental questions are on there.

So the NBME puts into the test experimental or ungraded questions. What they can do then is find questions that were easy, normal, slightly challenging, challenging, and impossible. You can tell which ones are which based on statistics. For example given a large enough sample size I can tell you which questions people were guessing on because it fits a particular statistical pattern. These experimental questions allow them to get an idea where the general direction of intellectual advancement is pointing. (either people are getting smarter or the are not)


Once I have seperated out my questions I populate the next years exam with a predetermined amount of each type of question. I include experiemental questions that were just above the average level of the students of the previous administration to correct for the increase in ability by the students in the future.


By doing this the NBME takes a risk, like the people who make the flu vaccine, that they got it right, because they may have miss calculated the success level of the future class and the distribution will either spread out or get narrow or be skewed (like this year). But they have the HUGE sample size of their stats on their side. Like is said in Vegas the house wins always.

These are my theories based upon my conversations with people at the NBME and what makes sense logically.

Remember the USMLE is like a driving test. It was meant to be pass/fail from the beginning and has unfortunately morphed into this pseudo-comparison test. The NBME has tried to fix this by eliminating the percentile rank publications. Have you ever hear anyone mention their Step 3 score. NO because at that point the ridiculous competition is over and they only care if you passed or failed.

 

[GoBlueMI]

I have a 278 on step 1. Not sure what the highest is but I have to believe there are people much higher than that.

 

[TexasTriathlete]

What is the lowest possible score? Could I get a 0 if I missed every question?

 

[HTxFrog]

I think 300 is the max score if you somehow got every question correct. I was kind of confused when I took one of my UWorld self assesments so I emailed UWorld and they said the self assessment was based on a theoretical maximum score of 300, so I assume they are basing that on what the actual usmle goes up to.

 

[RamblinRose]

This has probably been asked before, but what does it mean when a score is listed with the 3 digit score / 2 digit score. Ex 240/99

I heard it was old nomenclature......But what does 99 mean, and what 3 digit score is equal to 99??