What exactly do Step 1 scores mean?

[Dogen]

What do these scores really mean? My school (Case) has a class average of 227. My score is 232, which I'm VERY happy about. When I called to get my scores, the dean asked me what specialty I was interested in, and said I was highly competitive. I know this is a "good" score, but being the number-obsessed medical students we are, anyone know where we can find information on our percentile?

Is the two digit score out of 100?

Anyone also have any idea what type of board scores competitive internal medicine residencies look for?

Thanks

---

Members don't see this ad.

 

[sanfilippo]

fellow buckeye,
i called my dean's office and they gave me the score over the phone: 237. I don't know where I fall with that, especially if the mean and SD's have gone up. they said it was a "competitive" score, but i don't know if that means competitive enough for ob/gyn surgery or nowhere close to start considering ortho and derm.

-s.

 

[Dogen]

hey sanfilippo,
averages were up 1 point this year (so far with most scores reported) at my school. Last year's averages at WashU were 232. I don't know if scores are up across the country.

And if I wanted to do derm I would definitely be out of luck. I don't know any averages for any residencies. From what I've seen from surfing the net, residency programs don't seem to be making that easy to find. I know they would like higher scores because there's a very high correlation between the steps and licensing.

And congrats on your score!

 

[sanfilippo]

i think i oughta wait for my 2-digit percentile score. a lot of residencies like surgery use it for cut-off's for interviews. for example, there's a website for Brown's plastic surgery residency program where it clearly states it won't invite anyone with a score less than 225 for an interview. i think if you get by these cut-off's, even it's just a little with a score like mine, you still have a shot because you still have letters, clinical grades, etc. to boost your potential. now, i just have to work hard on my rotations...

when i got my score, i didn't know if it was good or bad especially since i'm not at a very "prominent" institution.

-s.

thanks, THOUGH, for replying!

 

[southerndoc]

Here's some stuff on percentiles...

The national mean is 216 with a standard deviation of 24.

Here are some percentiles I figured up using Z tables. If you want to find out your exact percentile, then go to http://psych.colorado.edu/~mcclella/java/normal/normz.html (thanks to an earlier poster, whom I cannot recall, for posting that link).

182: 7th
190: 14th
195: 19th
200: 25th
205: 32nd
210: 40th
215: 48th
220: 56th
225: 65th
230: 72nd
235: 79th
240: 84th
245: 89th
250: 92nd
255: 95th
260: 97th
265: 98th
270: 99th

Keep in mind that those percentiles are estimates, but they are generally accurate so long as the USMLE follows a standard Guassian curve (which I've been told it normally does, as with most other tests).

 

[scootad.]

Yep, those stats are incorrect. My sister just found out she got a 252 and is in the 99th percentile!

 

[LoneSEAL]

The 2 digit USMLE score is NOT your percentile. According to the national board of medical examiners (aka people who devise the USMLE), the two digit score "is derived in such a way that a score of 75 always corresponds to the minimum passing score...it is used in score reporting because some medical licensing authorities have requirements that include language describing a 'passing score of 75'.


Personally, I dont' see why the the NBME can't just give percentile scores in addition to the three digit scores although they do give you the mean and SD so you can plausibly figure out your unofficial percentile, assuming a standard distribution of course, which Geek Medic has kindly provided.


What I'm really curious about is why the ceiling for high scores has been raised so much. These days, it's not unusual to hear of people who have score 270+ on their USMLE1. However, just 5 years ago, it was unusual to hear people who even broke 240. It seems to me that the grading scale might have been readjusted ever since the transition to computer testing took place.

 

[scootad.]

OH! I guess my sister misinterpreted her two-digit score as being her percentile. I dont want to be the one to bring her down from her cloud!

 

[southerndoc]

I'm still wondering if the 3-digit score is a raw score. I have a strong feeling that it is.

 

[LoneSEAL]

Geek Medic,

I was thinking the same thing. Step I is a 350 question exam. It is my understanding 50 questions are "experimental" and are not factored into your score. Basically, NBME uses these experimental questions to "test" them out and see how students do on them. If you think back to high school, the SATs had a section-- either verbal or math-- that was experimental too. Statisticians use these trial runs to get a statistical account of how difficult the questions are.

So if 5o questions are experimental, that leaves 300 actual, scorable questions. Conveniently, and maybe not coincidentally enough, 300 is the theoretical max. And also, coincidentally, according to the 2001 First Aid, you need a score of 179 to pass. Well, 179/300 equal exactly 60%, which has been a figure that has been touted as the percentage of questions that one needs to answer correctly to pass.

Of course, this is all a hunch, but I wouldn't be too surprised if the 3 digit score now correlates quite closely with your actual raw score.

 

[southerndoc]

USMLE also eliminates an additional 10-20 questions that had problems of some sort (deemed too difficult, too easy, question was messed up, answer choices messed up, etc.). Hopefully they factor these in. 80% of 290 questions is a different raw score than 80% of 280 questions.

By the way, 182 is the new minimum passing score. I've read that anywhere from 55-65% is considering passing, but if I call correctly, the USMLE bulletin states that 60-70% is considering passing.

 

[Jim Picotte]

One of the reasons they don't post percentages anymore is because it is basically unfair. Let me show you what I mean. I took Step I in June 2000. I'm looking at my score report now and the mean on this test was 215 with a SD of 20. I received a 230 on Step 1. Figuring out my percentile, I was at the 77.3% for that group of students. So, lets say that this is reported and all the residency programs see this number as well. I go on a 2 year hiatus before residency and decide to apply for July 2004 residencies (just like most who just took Step 1). Well, when the PD look at my report, they say wow this guy is in the 77.3% and they look at someone who took Step 1 this year, had the same 230 but their percentile is 72%. Now it looks like I'm a bit ahead of this guy when in fact, we both got the same score on the exam. The NBME is trying to be consistant in saying that if you get a 230 3 years ago, it means you have/had the same amount of knowledge as someone who scores a 230 in 5 years even though scores will always rise as review books/courses get a better idea on what the exam is like and become more effective in teaching toward the exam. I'm betting a 230 in another couple of years will be around 67% and will keep dropping unless the NBME decide to change the style of the exam again and the minimum passing score will continue to rise as well.

 

[southerndoc]

As long as the USMLE provides means and SD's, then residency directors will always have the option of estimating percentiles using Z tables. This could be both good and bad.

 

[hotbovie]

The NBME used to report percentile along with both the two digit and 3 digit score. When I took step 1 last year, they stopped reporting the percentile. Why? To try to prevent residency programs from using the test to screen applicants. It didn't, though. I'm looking at some programs that publish a minimum score. However, I believe I'll be OK at 244. Still, made me kind of mad to know that I can't find out my percentile, but could only estimate it from the z tables (statistcs isn't my strongest thing, so I'm not totally confident with my result)