- Joined
- May 15, 2003
- Messages
- 24
- Reaction score
- 0
Do any of you guys have a breakdown of the scaled ITE scores and their related percentiles for the 2006 ITE? For instance, if resident scaled score was 35, what was the percentile?
ITE and Percentiles?
- Thread starter LRingers
- Start date
- Joined
- Jul 6, 2005
- Messages
- 1,062
- Reaction score
- 156
depends on what level of training you are. if you are getting ready to start CA-3, then you still have a lot of work to do. but, if you are just finishing your internship year, then you would have rocked the house.
- Joined
- Sep 11, 2004
- Messages
- 1,834
- Reaction score
- 4
If you got a 35 and you are a:
PGY-1 = 99%
CA-1: 89%
CA-2: 70%
CA-3: 23 %
Double check though
- Joined
- Sep 11, 2004
- Messages
- 1,834
- Reaction score
- 4
It came with my ITE results.
The percentiles for the 2007 written exam are in. Passing this year was a 209 (about 20th percentile), the mean was 250 with a std dev of 50.
Pctle : Score
10%: 186
20%: 208
30%: 224
40%: 237
50%: 250
60%: 263
70%: 276
80%: 292
90%: 314
95%: 332
98%: 350
Pctle : Score
10%: 186
20%: 208
30%: 224
40%: 237
50%: 250
60%: 263
70%: 276
80%: 292
90%: 314
95%: 332
98%: 350
- Joined
- Mar 18, 2004
- Messages
- 131
- Reaction score
- 1
Not sure those percentiles are right.
Honestly, did you get that information from the ABA?
It looks like to me like you have constructed a curve based on the mean and standard deviation which assumes a normal distribution using a program like excel.
The problem is, the curve is probably not normal. On the written exam, to fail 21% of test takers might be a bit high, especially in a group as bright as this one. There are some retakers of course. The groups have been getting brighter each year and will get even brighter for the next couple of years and then things will level off. To increase the fail rate would be a bit strange. Now failing 21% of oral board takers would be consistent with past years which is interesting since all of these people passed their written exam.
Anyway, the distribution of scores usally isn't normal. For a test like the USMLE where you have tens of thousands of test takers - this almost works, although not perfectly. For the ABA written with some number of test takers considerably less than 2000 (you can't count the all the ITE people for several reasons including different ranking group and differet question subset being scored) and a large number of questions of quite variable difficulty the curve is almost certainly not normal. That said between about the 30th and 70th percentile the numbers probably aren't too far off. At the ends the numbers are likely way off.
The ABA occasionally sends out information to residents that shows yearly passing rates and percentage of test-takers passing for written and oral board exams. This information doesn;t seem to show up on their website for whatever reason.
Anyway, if you got this info from the ABA and the curve is just perfectly normal then thanks for sharing but the ABA doesn't just send out the percentiles with exam results like they do for the ITE. That said, the ITE results do have the percentiles for the all classes including the graduating CA-3's who are testing for credit (doesn't include retakers) but the information is presented on the 2 digit scale only and is for the entire part of the exam which is considered the ITE, as opposed to that part which is used for board cert.
cheers
Honestly, did you get that information from the ABA?
It looks like to me like you have constructed a curve based on the mean and standard deviation which assumes a normal distribution using a program like excel.
The problem is, the curve is probably not normal. On the written exam, to fail 21% of test takers might be a bit high, especially in a group as bright as this one. There are some retakers of course. The groups have been getting brighter each year and will get even brighter for the next couple of years and then things will level off. To increase the fail rate would be a bit strange. Now failing 21% of oral board takers would be consistent with past years which is interesting since all of these people passed their written exam.
Anyway, the distribution of scores usally isn't normal. For a test like the USMLE where you have tens of thousands of test takers - this almost works, although not perfectly. For the ABA written with some number of test takers considerably less than 2000 (you can't count the all the ITE people for several reasons including different ranking group and differet question subset being scored) and a large number of questions of quite variable difficulty the curve is almost certainly not normal. That said between about the 30th and 70th percentile the numbers probably aren't too far off. At the ends the numbers are likely way off.
The ABA occasionally sends out information to residents that shows yearly passing rates and percentage of test-takers passing for written and oral board exams. This information doesn;t seem to show up on their website for whatever reason.
Anyway, if you got this info from the ABA and the curve is just perfectly normal then thanks for sharing but the ABA doesn't just send out the percentiles with exam results like they do for the ITE. That said, the ITE results do have the percentiles for the all classes including the graduating CA-3's who are testing for credit (doesn't include retakers) but the information is presented on the 2 digit scale only and is for the entire part of the exam which is considered the ITE, as opposed to that part which is used for board cert.
cheers
MDentropy you are right - those numbers were constructed using assuming a normal bell curve distribution, given a mean of 250 and a standard deviation of 50. About 2150 people took the exam last July, and about 650 of them were repeaters.
The ABA has a graph on it's newsletter showing the pass rate for the written exam (for all test takers, including repeaters) from 1996 to 2006. Estimating from the graph, the pass rate looks as follows:
Year : % Passing
1996 : 64%
1997 : 62%
1998 : 69%
1999 : 53%
2000 : 46%
2001 : 55%
2002 : 63%
2003 : 64%
2004 : 59%
2005 : 72%
2006 : 74%
If you're just talking about the pass rate of first time examinees (i.e. excluding repeaters), the pass rate is significantly higher:
From 2002-2006: 71%, 76%, 73%, 82%, 88%
However for percentile purposes, I would think that one would want to include everyone taking the exam in that particular year.
The ABA has a graph on it's newsletter showing the pass rate for the written exam (for all test takers, including repeaters) from 1996 to 2006. Estimating from the graph, the pass rate looks as follows:
Year : % Passing
1996 : 64%
1997 : 62%
1998 : 69%
1999 : 53%
2000 : 46%
2001 : 55%
2002 : 63%
2003 : 64%
2004 : 59%
2005 : 72%
2006 : 74%
If you're just talking about the pass rate of first time examinees (i.e. excluding repeaters), the pass rate is significantly higher:
From 2002-2006: 71%, 76%, 73%, 82%, 88%
However for percentile purposes, I would think that one would want to include everyone taking the exam in that particular year.
- Joined
- Mar 18, 2004
- Messages
- 131
- Reaction score
- 1
I like your response. I was trying to link my last comments to that newsletter but it wasn't working for me.
You are right, it seems that repeat test takers would be included for percentile purposes. And I implied that this wasn't the case, because that is what I was thinking. That is probably wrong, but either way things are still a little murky and here why I say this:
Interestingly, last year the mean, sd, and passing score were all exactly the same as this year. I am basing that on old post on this website from last year which i went back and re-read earlier today. So does that 209 passing score in 2006 match up to the 26% percentile or the 12th percentile?
The calculated percentiles are a little closer to 26 but not much. It is almost right in between which means either the distribution is either much tighter or much looser than one would calculate.
And certainly if re-takers are included in the percentiles this makes the comparison to the ITE difficult if not impossible.
Most of the repeaters fail the exam, most score less than 209. So one would expect the scoring percentiles to jump from when one takes the exam from starting CA-3 (the ITE) to finishing CA-3 (the real thing) just because you've added a bunch of people to the group who do quite poorly on the exam. This would be true even if one has stayed at the relative same percentile within their peer group. i.e. if you have tended to score around 50th percentile or 75th or 95th on your ITE, and you have actually done sone when compared to the FIRST-TIME test takers, your percentile on the real exam would be significantly higer IF ALL SCORES are in included. Based on my relatively small N this doesn't seem to be the case, which has led me to further believe that the numbers are skewed, but that may have something to do with my sample being at the higher end of the score range.
Anyway, i dont have an answer to all this. I just think it could be more clear. The ABA is a bit cryptic in my opinion.
In 1999 the NBME stopped reporting percentiles for the USMLE. I don't really know why exactly. Seemed like they wanted to make the information a little less clear and people would then focus more on pass and fail. I think the ABA exam is made in conjuction with the NBME. In the defense of the NBME and the USMLE, I think the pass rates are a little more appropriate (95% range). It seems kinda harsh that in 1 in 8 finishing CA-3's failed the exam last year. Hopefully fewer failed this year.
I must say, I never worried about the written exam because the ITE let me know it wouldn't be a problem but quite honestly, the fail rate for the orals is somewhat worrisome especially given the length of the exam (short) and the subjectivity. Why in the world would 25% of people who just passed their written exam need to fail this test? Again, seems pretty harsh.
You are right, it seems that repeat test takers would be included for percentile purposes. And I implied that this wasn't the case, because that is what I was thinking. That is probably wrong, but either way things are still a little murky and here why I say this:
Interestingly, last year the mean, sd, and passing score were all exactly the same as this year. I am basing that on old post on this website from last year which i went back and re-read earlier today. So does that 209 passing score in 2006 match up to the 26% percentile or the 12th percentile?
The calculated percentiles are a little closer to 26 but not much. It is almost right in between which means either the distribution is either much tighter or much looser than one would calculate.
And certainly if re-takers are included in the percentiles this makes the comparison to the ITE difficult if not impossible.
Most of the repeaters fail the exam, most score less than 209. So one would expect the scoring percentiles to jump from when one takes the exam from starting CA-3 (the ITE) to finishing CA-3 (the real thing) just because you've added a bunch of people to the group who do quite poorly on the exam. This would be true even if one has stayed at the relative same percentile within their peer group. i.e. if you have tended to score around 50th percentile or 75th or 95th on your ITE, and you have actually done sone when compared to the FIRST-TIME test takers, your percentile on the real exam would be significantly higer IF ALL SCORES are in included. Based on my relatively small N this doesn't seem to be the case, which has led me to further believe that the numbers are skewed, but that may have something to do with my sample being at the higher end of the score range.
Anyway, i dont have an answer to all this. I just think it could be more clear. The ABA is a bit cryptic in my opinion.
In 1999 the NBME stopped reporting percentiles for the USMLE. I don't really know why exactly. Seemed like they wanted to make the information a little less clear and people would then focus more on pass and fail. I think the ABA exam is made in conjuction with the NBME. In the defense of the NBME and the USMLE, I think the pass rates are a little more appropriate (95% range). It seems kinda harsh that in 1 in 8 finishing CA-3's failed the exam last year. Hopefully fewer failed this year.
I must say, I never worried about the written exam because the ITE let me know it wouldn't be a problem but quite honestly, the fail rate for the orals is somewhat worrisome especially given the length of the exam (short) and the subjectivity. Why in the world would 25% of people who just passed their written exam need to fail this test? Again, seems pretty harsh.
- Joined
- Jun 21, 2006
- Messages
- 2,743
- Reaction score
- 1,768
Call your home program and ask. They should have gotten the report with the scaled score.
As for the original question and the initial response by volatile, I tend to disagree with his response. You don't have "a lot of work to do." If you were taking it for real, you just got a passing score and would be studying for the oral board exam. I think his point is that you would not be in a high percentile. But you would be sitting for your oral exams with the people who may have gotten a 50 on the scaled score. And in the end, it won't matter.
That being said, don't get complacent and not study your butt off. The last thing you want is to go backwards in the year that it really counts. Good luck!
- Joined
- Jul 12, 2002
- Messages
- 446
- Reaction score
- 19
Gern is absolutely right. There is a guy in the class ahead of us that did just that. Everyone was shocked. He passed it as a CA1 (99% all years). Just didn't pass it the year it counted. Sad but true.
Mean CA3 score for 2007 was 37 with a std deviation of 5. If you juxtapose that with a mean certification score of 250 and a std deviation of 50, you get roughly:
2 dig __ 3 dig __ %
27 ____ 150 ___ 3
32 ____ 200 ___ 16
37 ____ 250 ___ 50
42 ____ 300 ___ 84
47 ____ 350 ___ 98
2 dig __ 3 dig __ %
27 ____ 150 ___ 3
32 ____ 200 ___ 16
37 ____ 250 ___ 50
42 ____ 300 ___ 84
47 ____ 350 ___ 98
Thanks, but I'm most interested in the upper edge (the top %iles each year), which is right where the ITE's not-normal distribution gets tricky, and where the math you did here less accurately reflects actual %iles.
Someone out there must have the norm table that can be scanned & posted as a gif or something.
