Step 1 Percentiles?

[Smokey]

---

Members do not see this ad.

Any source of percentiles for Step 1? (i.e. 220=xth percentile) It obviously changes for each administration, but I am just trying to get a feel.

Thanks

 

[Faebinder]

(I am sure the oldies are tired of hearing this but it has to be said) It's not percentile.. it's a two digit score equivilant. Hasn't been a percentile since the 90s.

 

[gutonc]

Any source of percentiles for Step 1? (i.e. 220=xth percentile) It obviously changes for each administration, but I am just trying to get a feel.

Thanks

Given the statistics they provide with your score report, you can actually calculate a percentile. This will not be anywhere near your 2-digit score though, unless you scored a 99.

More importantly however, it really doesn't matter.

BE (now PE)

 

[Nerdoscience]

Can you really? I thought they gave you the mean score, but to calculate the percentile, you'd need the median. I guess you could assume a perfecl bell without skew. Maybe I just don't remember.

 

[DCMDtobe]

From my USMLE Step 1 score report, I've got:

Mean = 217
Standard Deviation = 23

The basic stats are as follows:

Z-score = (your score minus mean) / standard deviation.

For example, if you got a 230, your Z-score is (230 - 217) / 23 = 13 / 23 = 0.565.

Go to a Z-score table to figure out your percentile.
http://www.math.unb.ca/~knight/utility/NormTble.htm

For example, a Z-score of 0.565 puts you at about the 71st percentile.

 

[Faebinder]

From my USMLE Step 1 score report, I've got:

Mean = 217
Standard Deviation = 23

The basic stats are as follows:

Z-score = (your score minus mean) / standard deviation.

For example, if you got a 230, your Z-score is (230 - 217) / 23 = 13 / 23 = 0.565.

Go to a Z-score table to figure out your percentile.
http://www.math.unb.ca/~knight/utility/NormTble.htm

For example, a Z-score of 0.565 puts you at about the 71st percentile.

Hmm, not a statistician here but the table you linked wont work for the calculation I believe. The population distribution has only 1 tail end not 2. (remember there are a lot of people in the 180s, 190s and very few in the 260s and 250s.) It's not a bell curve for sure since the cut off is a straight line at 182 (Unless 217 is the mean of EVERYONE that took the exam, which I don't think it is.)

 

[WatchingWaiting]

Hmm, not a statistician here but the table you linked wont work for the calculation I believe. The population distribution has only 1 tail end not 2. (remember there are a lot of people in the 180s, 190s and very few in the 260s and 250s.) It's not a bell curve for sure since the cut off is a straight line at 182 (Unless 217 is the mean of EVERYONE that took the exam, which I don't think it is.)

Actually, 217 is the mean of first-time US M.D. takers, so it does include 7-8% below 182.

 

[Mumpu]

Unless it's a true bell curve, the mean and SD statistics are meaningless. The whole idea of mean and SD is that you can plug them into an equation and plot the bell curve of the dataset. If the distribution is not normal, these numbers are misleading and non-parametric values should be reported instead.

 

[Faebinder]

Unless it's a true bell curve, the mean and SD statistics are meaningless. The whole idea of mean and SD is that you can plug them into an equation and plot the bell curve of the dataset. If the distribution is not normal, these numbers are misleading and non-parametric values should be reported instead.

Exactly. 👍 We need the median.

 

[dagr8vik]

the bottom line is that you can assume a normal distribution and get an approximate percentile (to give you the 'feel' it sounds like you were requesting).

But you cannot know the actual percentile because they don't want you to know and do not report the required information to figure it out (which, by the way, is more than just the median).

 

[Dunce]

Do program directors or people at the med schools get a better sense of the range of scores or do they just get the same mean (217) that was given to test-takers?

 

[getinthehole]

I think the PDs may or may not be aware of the mean, but probably don't care. I think they download applications to their program from ERAS, and maybe then sort their applicants by board score. At this point, they have an idea of how competitive the applicants are to their program, if the average is 240 or 200 it is sort of irrelevant to them because they pick from their applicants. Then, I think they go down the list, read their stuff, and decide who to give interviews to. Interviews aren't just given on stats though. If you have 260, AOA, and apply to 40 places, you probably won't get an interview at places you are WAY too competitive for. Reasoning... you won't go there anyways, and they need to interview people who have a decent liklihood of matching in their program.