CFA pass rate variance

Ok so here is a question relating to the CFA published pass rates… http://www.cfainstitute.org/cfaprog/pdf/candidate_results.pdf Why, year on year, could the L2 (or any level for that matter) pass rate differ by 20% (e.g. 2004-2005). I mean, we are talking about samples of 20,000+ people in recent years and the CFA is trying to tell us that 4000 more people deserved to pass one year than in the last? Perhaps these stats imply, assuming a bunch of things including the material they test being the same and the range of people they test being from the same pool, that the CFAI’s questions or more possibility their grading of them is not consistently testing for the same thing year on year. I have some answers to it myself but i won’t anchor others in their thinking about (with my answers)…

Must be lower than 70%… reasons: - I doubt someone had 100% - in the past when cfai published it was top 70th percentile - the exam was really tough hehe - if you look at last year’s passing thread there are a few that had to get max in each category to get 70% - overall AF impression - I need to pass I agree that 70 is safe territory but you cab get away With less than that I rest my case

Bump… Any comments on this

i also would like to know. someone please say low 60s

Opinions from retakers??

let me clarify this: Passing score is determined by Angoff method which is 70% of top 1% score. For an example: Out of 1000 candidates, CFAI will take top 10 people’s score and average it (like if top 10 people scored 95 on average), passing score would be 67.5 (70% of 95). Whoever scored 67.5 or higher will pass. It means all 1000 candidates pass if all of them score atleast 67.5. I hope this will clarify

@ cfaboston That is not an Angoff method. Angoff assembles a team of Subject Matter Experts. Each one independently rates each question based on the percentage of “minimally qualified candidates” who would answer it correctly. Then they sum each expert’s percentages for the entire test, and average the sums. That average is the MPS. CFAI admits that they do certain adjustments to this average, but in a nutshell, that’s Angoff.

dlpicket Wrote: ------------------------------------------------------- > @ cfaboston > > That is not an Angoff method. > > Angoff assembles a team of Subject Matter Experts. > Each one independently rates each question based > on the percentage of “minimally qualified > candidates” who would answer it correctly. Then > they sum each expert’s percentages for the entire > test, and average the sums. That average is the > MPS. CFAI admits that they do certain adjustments > to this average, but in a nutshell, that’s Angoff. Thanks. But the MPS they determine from what I’ve mentioned.

http://www.cfainstitute.org/cfaprog/overview/pdf/IntoOur5thDecade.pdf search for the heading “standard setting” I think its page 11.

cfaboston28 Wrote: ------------------------------------------------------- > let me clarify this: > > Passing score is determined by Angoff method which > is 70% of top 1% score. For an example: Out of > 1000 candidates, CFAI will take top 10 people’s > score and average it (like if top 10 people scored > 95 on average), passing score would be 67.5 (70% > of 95). > > Whoever scored 67.5 or higher will pass. It means > all 1000 candidates pass if all of them score > atleast 67.5. > > I hope this will clarify ^ yeah if they do it that way, the sole reason that the scores differ should be something like: 2004 , 32% Pass Rate - High Top 1% Scores and Low Mean Scores 2005, 56% Pass Rate - Low Top 1% Scores and High Mean Scores So in 2004 they must have created a test that was pretty hard with not a whole lot of level 9.5 - 10 difficulty questions, but a lot of 8’s, letting the smart people raise the top scores but leaving the average people behind.

the document I linked said they started using angoff in 2005 that could be the difference.

speaking of which, how did a bunch of L2 candidates forget to check if the sample was covariance stationary. There might be different policy regimes ! DOH !

Do you think the top 1% would get more than 112/120? If not, then MPS would be somewhere around 65% (79/120). I think this would be the ultimate borderline case and could depend on how you did on Ethics maybe.

I am afraid that given the high number of candidates this year, there are higher chances of people getting really high grades and therfore raising thw MPS. I think top 1% averaged 116/120. Dont even know if that’s human.

75%, do you think that would be a pass?

Everything above 70% is a pass

Individual sections or overall?

philip.platt Wrote: ------------------------------------------------------- … > letting the smart people raise the top scores but leaving > the average people behind. So true.

Overall… Thealiman, if you got 75… Stop wasting your time in this forum and enjoy the summer… You’ve earned it Af now belongs to the avg joe’s who are hoping for a less than 70 mps in the exam…and are trying to recreate the exam… Sad but I am in that group