Medicine Shelf Scoring

[bigfrank]

---

Members do not see this ad.

Can anyone comment on how recent Medicine Shelf Exams are scored? I've heard that your "raw" score is the # of questions answered correctly out of 100; conversely, I've heard that 70 is assigned as the mean with a SD of 7-8.

Can any recent medicine shelf-ers comment on this please?

Thanks

 

[imtiaz]

the high/mean/SD are "for this administration" meaning those that took the exam with you. for us it was 83/73/6.8. i think it varies with each administration.

 

[Hawk22]

"the high/mean/SD are "for this administration" meaning those that took the exam with you. for us it was 83/73/6.8. i think it varies with each administration."

Does "this administration" include everyone at your school? Everyone who took it that month? Who exactly is included in that group?

I got my score from my school today, but I haven't received anything official in the mail yet. Will be interesting to see what it says.....

 

[Stinger86]

As per the nbme website:

"Scores are provided on a scale that has been developed individually for each
test. These scores are scaled to have a mean of 70 and a standard deviation of 8 for a specific group of examinees. It is important to note that although the scores have the "look and feel" of percent-correct scores, they are not. This scale provides a useful tool for comparing the scores of your students with those of a large, nationally representative group taking the subject test as an end-of-course assessment"

So your "raw score" is only as good as your performance compared to your peers

 

[bigfrank]

Thanks guys. It's still confusing, because I need a certain "national percentile" on my test to get honors at my school. It appears that this is a strange way to do it if we're graded against our peers at our school.

 

[Cpt.Hook Hamate]

NBME really gets on my nerve with that scoring crap. Why does every thing have to be all 007 top secret agent stuff? I still would like to know just how many questions I got right on the darn Step 1 test...just out of curiosity. But nooooo...then they would have to kill me if they gave that info. 🙄

 

[imtiaz]

Does "this administration" include everyone at your school? Everyone who took it that month? Who exactly is included in that group?

I got my score from my school today, but I haven't received anything official in the mail yet. Will be interesting to see what it says.....

the 83 was within our group so i'm guessing that it only includes those students at your school that took the exam with you, and not a national group that took it on that same day.

 

[bigfrank]

the 83 was within our group so i'm guessing that it only includes those students at your school that took the exam with you, and not a national group that took it on that same day.

Thanks, that's a pretty small sample size, but it should all work out OK. Best,

 

[Hawk22]

the 83 was within our group so i'm guessing that it only includes those students at your school that took the exam with you, and not a national group that took it on that same day.

I'm not sure that this makes sense to me. I always thought that shelf exams were supposed to be a "standardized" test in which folks at different schools could be compared with a similar test. If the scoring isn't absolute (i.e. an 80 means you got 80 out of 100 questions right), but is instead based on a specified mean for a group at a medical school (i.e. the mean for the group will be given a 70, then go from there), then it would be impossible to compare other students.

Lets say someone gets an 80 at the hypothetical #1 medical school in the country (take your pick). If the 2nd method of scoring is true, then an 80 at the weakest medical school (again, take your pick) wouldn't be anywhere near the same performance because the group as a whole just isn't as good. I know that you can make the same arguement for any medical school test, but again I thought that this was supposed to be a "level the playing field" type test.

I also go to a fairly large school and we probably had around 40 or so people take the test. That's a decent number, but in many other smaller schools I could easily see how less than 20 people would be taking it at any given time. That's a very small "N" on which to base that type of statistical analysis.

In the end, I know it doesn't mean much, but it would be nice to at least understand the method behind the madness.

 

[imtiaz]

i think these tests are already standardized based on testing that was done previously. so they still hold true even in small N's because they are already proven to be standard. it's the same test every time, no different forms or anything.