CFA Discussion Topic: Bias on AnalystNotes Percentile Ranking

To post a new topic or reply to a topic, please log in or register for a free user account.

Author Topic: Bias on AnalystNotes Percentile Ranking
Rguerra
@2005-08-28 10:25:18
I´ve read AnalystNotes (AN) explanation on how an 80% absolute result on a mock exam does not matter and the reason why the relative score is so important. However, when I do the calculations, many times I score over 70% correct and still get an AN relative score of lower than 50%. Isn´t there a downward bias in this result? How does AN account for all people who take the test unprepared? If you think that only people who really care about studying pay the $60 or so and dedicate hours to take mock questions, it seems to me that the relative scores are somewhat loaded with a "survivorship bias". Isn´t that so? How "good" is AN sample of students as compared to the whole bunch who take the exam? Does anyone know? Any comments?
yanpz
@2005-10-07 12:19:32
I have the same feeling... Since it's based on last 50 questions you just did, why cares? I don't. :)
JZino
@2006-06-01 14:09:42
Rguerra I agree. People who pay for this service are likely to be more prepared and better test takers. Your relative ranking should be increased to compensate for all the un-prepared and under-prepared people you will be facing. The sample is biased.

That being said, scoring 50%-tile ranking sure does encourage you to study.
jam99003
@2006-06-02 03:40:05
I think a 50%-tile would on here would put you much higher on actual test day. But that may be wishful thinking as I am consistenly getting in the 50-65%tile range. ONly thing I can hope for is that the top 2% do worse than on saturday than they statistically have in the past.
jcramirez
@2006-06-29 18:29:44
if anything, all the "unprepared" and non-caring people should boost your AN relative score, no? The unprepared should score low, and if there a lot of them, than you should be scoring higher than a lot of people
george2006
@2006-07-27 13:14:54
I think the downward bias is much better for our final actual exam result because it helps keeping us on the toe. You will feel much more confident in the real exam.
gene80
@2006-10-30 03:13:55
Totally agree with George2006
ravincfa
@2006-10-30 08:24:28
Hi All,

I have registered for CFA Level 1 June 2007 and bought CFA institute material. Do i need to refer some other material or CFA institute material is sufficient.

Please help on this and give me some tips regarding preparation.
Nathan
@2006-11-04 22:13:38
I have noticed something unexpected in the comparative statistics. When I look at the "study progress" bar charts, I believe that the page is well described by the text at the bottom - "This chart shows the total number of review questions you have attempted and answered correctly, the average attempted by members of this site and the total questions available for each study session."

As luck would have it there's this statistical anomaly where members of this site on average have attempted about 86% of the questions in each and every group of questions. One would expect some deviation from set to set, but the only deviation can be accounted for by rounding errors.

It just seems odd, but that said ... I would certainly rather have something telling me to work harder than giving me a false sense of security.
sahild05
@2006-11-15 22:12:35
the top 2% on a website should easily be the top 1% worldwide, given that they have liberal internet access and put in this much time. So 50% takes you to 75% (simple regression). Therefore the average person on this website will pass.
kirkgerry
@2008-07-29 11:23:47
The mock exams are harder than the real CFA exam. I got 55-60% in mock exams here but got all 70%+ in the real exam.
capitalpirate
@2010-11-06 09:50:35
the exams on AN are definitely harder than the CFA! keeps you on your toes!