Hi guys, I just noticed that most people in this forum who are using CFA calculator are not aware of ability to download Statistics Data from the CFASUCCCESS.COM site. The download link is http://www.cfasuccess.com/cs/calculator/ This calculator compiles the scores from those who used 40/60/80 rules, and statistics data is created in an excel file for each exam term. You can download these excel files by clicking “Download Statistics Data” button. Make sure you choose correct exam term and level when download them. These excel file was re-generated almost daily depends how many new records are added to the database. We know larger size of sample has better representation, so I encourage all of you to use CFA calculator to get larger size of sample data. Good luck to everyone.
Bump. Thanks, Richard! I know some of the more quantitatively minded among us have been curious about the extensive data you’ve collected. Seriously, it’s great and appropriate that you’ve given us access to the data.
We need some more people who only just passed to get their data in there, and who narrowly failed (so all you band 10ers!) So far (from Richard’s data) We have the highest mark (on a minimum basis) of someone who failed at 225/360 (62.5%) and the lowest mark (on a maximum basis) of someone who failed at 252/360 (70%). That’s not hugely helpful so far!
Richard’s data needs quite a lot of cleaning - there are lots of results clearly not from this year’s level 2 (wrong weights), quite a few incomplete ones (total weight <360), level 3 results (!). I had a crack with manuhack’s data and this seems pretty good (and he does the min/max basis calcs for you). After cleaning the obviously nonsense ones, there is a pass with a maximum possible score of 65% (but only one), and one each @ 67 and 68, then a couple at 69. On the fail side, there are lots that failed with a minimum possible score of 64%. So I think that puts the MPS at around 65. We still don’t know if these marks are already put on a curve though, and that the raw marks are different.
chrismaths Wrote: ------------------------------------------------------- > So I think that puts the MPS at around 65. We > still don’t know if these marks are already put on > a curve though, and that the raw marks are > different. Why would you suggest that the marks are put on a curve? I’ve never heard anything from the CFAI in this regard before. I think the 65 is probably about right (keep in mind that there may always be some deviations based on Ethics scores).
Wow. I knew I went through by the skin of my teeth. Scored a 65% using the 40/60/80, it was ethics (>70%) that pulled me though I am sure. I was sure I had failed too. Better put in a better effort for LIII!
plyon Wrote: ------------------------------------------------------- > chrismaths Wrote: > -------------------------------------------------- > ----- > > So I think that puts the MPS at around 65. We > > still don’t know if these marks are already put > on > > a curve though, and that the raw marks are > > different. > > > Why would you suggest that the marks are put on a > curve? I’ve never heard anything from the CFAI in > this regard before. > > I think the 65 is probably about right (keep in > mind that there may always be some deviations > based on Ethics scores). Because from previous “discussions” about the MPS, the MPS is lower than that on a raw basis. I don’t think the top 1% got 93%! Maybe the top 1…
Based on the slamming anyone who touts their accomplishments gets around here, I doubt you’d hear about how well people really think they did. But on Level I, I’d be surprised if I didn’t get a 93 or better and Level II a 90 or better. The top 1% could be 93%. Anyway, that “rule” is only a guideline used to be sure the psychometric methods they now employ do not produce results inconsistent with previous methods. They could easily tolerate a fairly wide band of variance from the 1% rule. The previous “discussions” tended to be more speculation than raw data. Now we’ve got some data so I’d bank on that more than the speculation.
Interesting … now if you check the statistics on the manuhack.com/cfamps, they also include the MPS estimate, right now it shows between 65.0% and 66.7% for level 2.
plyon Wrote: ------------------------------------------------------- > The previous “discussions” tended to be more > speculation than raw data. Now we’ve got some > data so I’d bank on that more than the > speculation. I’m afraid that’s nonsense. They were based on the minutes of the 2004 MPS meetings that someone “found”.
chrismaths Wrote: ------------------------------------------------------- > plyon Wrote: > -------------------------------------------------- > ----- > > > The previous “discussions” tended to be more > > speculation than raw data. Now we’ve got some > > data so I’d bank on that more than the > > speculation. > > I’m afraid that’s nonsense. They were based on the > minutes of the 2004 MPS meetings that someone > “found”. That’s the speculative part – the MPS in 2004 was less than 65, so anything that tells us about the MPS four years later is speculation, right? The original thread was deleted months ago so I can’t find it now. Point is that the most current info from the CFAI says they uses standards setting (based on a modified Angoff method) with the"70% of top 1%" rule as a check.