My boss is going to bat for me with our comp committee to get me a larger raise and/or bonus for passing L3. I passed all three levels on the first attempt (and as such I’m aware of ethics prohibiting me from using that as a bargaining chip, but my boss is not a charterholder and wants to use that in his argument). He has requested from me an idea of the percentage of people who pass 3/3. Obviously CFAI does not release this. Is there any better way to estimate this other than the admittedly flawed (pass rate LI)*(pass rate L2)*(pass rate L3)? Thanks.
As a charterholder you should tell him not to imply that passing on your first try makes you any better of a person than someone who took each test twice. You are both still allowed to put CFA after your name.
Just out of curiosity, where’s the flaw in: No of candidates passing level III -------------------------------------------- No of candidates registered for Level I Given the assumption that the total number of Level I or Level III candidates does not fluctuate too much in between years.
MehdiOchre Wrote: ------------------------------------------------------- > Just out of curiosity, where’s the flaw in: > > No of candidates passing level III > -------------------------------------------- > No of candidates registered for Level I Over 45 years, that’s 108,234 --------------- 1,025,407 Wow, did you know that 92,000 people took Level I in June 2008?
MehdiOchre Wrote: ------------------------------------------------------- > Just out of curiosity, where’s the flaw in: > > No of candidates passing level III > -------------------------------------------- > No of candidates registered for Level I > > Given the assumption that the total number of > Level I or Level III candidates does not fluctuate > too much in between years. Terrible assumption. Number of L1 candidates appears to be increasing every year.
frisian Wrote: ------------------------------------------------------- > Wow, did you know that 92,000 people took Level I > in June 2008? So roughly as many L1 candidates in 1 year as total charterholders since the beginning of the program. Brutal.
I’d say there’s no real way to come up with a number that isn’t complete BS on some level. There are no statistics on anything remotley related, first of all if you use: > No of candidates passing level III > -------------------------------------------- > No of candidates registered for Level I there could be level III candidates that passed that have been taking the test for 5 or 6 years. There’s no available numbers to be able to acurately come up with a figure. Also this is such a hazzard ethically (Reference to the CFA, Conduct of members/candidates, and Duties to Employers issues all could be raised) that I would say not to do it.
> Over 45 years, that’s > > 108,234 > --------------- > 1,025,407 > CFAI says that the average # of attempts is 4. That means that the top 50% of people who pass all 3 do it in 3 or 4 attempts. I don’t think you can get any certainty about where to draw that line for the 3/4 split, but I’d say that since you’re averaging with 5,6,7,8,9 (please tell me no one wasted a decade on this thing) its going to take mostly 3’s in that top 50% to pull the average to a 4. Worst case, 50% of CFAs did it in 3 attempts. So back to this formula, multiply the result by 0.5 and you have your worst-case-percentage. This is ok internally I guess, but if you advertise a number derived this way, you’re going to meet the nice PCP folks.
Take a moving say 3 yr or 5 yr avg of the pass rates for each level and multiply them by each other. Generally speaking I think 10% is a rule of thumb…
The 4 year average gets even more complicated when you remember that L1 is given 2x per year. Any model based on available data has a ton of assumptions, and CFAI is sitting tight on the right answer. My guess is that the number is between 10% and 25% of charterholders did it 3/3. Since a lot of people give up, the number of candidates who make it through to get a charter 3/3 is probably very small. Most people who failed at least one exam probably failed L1.
bchadwick Wrote: ------------------------------------------------------- > Most people who failed at least one exam probably > failed L1. This is purely anectodal, but most of the people I know who took 4+ attempts failed L2.
frisian Wrote: ------------------------------------------------------- > > Over 45 years, that’s > > 108,234 > --------------- > 1,025,407 > > > Wow, did you know that 92,000 people took Level I > in June 2008? I was more thinking about, not 108234/1025407 = 11%, but rather 12314/54549 = 23% or 12314/43777 = 28% (for one given year some years ago). The process initially suggested 0.40 * 0.48 * 0.76 = 15% seems too small. So maybe the total pass rate is about 20% instead of 10-15%. Still - four out of five don’t make it.
dlpicket Wrote: ------------------------------------------------------- > bchadwick Wrote: > -------------------------------------------------- > ----- > > Most people who failed at least one exam > probably > > failed L1. > > This is purely anectodal, but most of the people I > know who took 4+ attempts failed L2. Could be for those who end up with the charter, but if you include the people who drop out along the way, I think you’ll find L1 is a killer.
dlpicket Wrote: ------------------------------------------------------- > bchadwick Wrote: > -------------------------------------------------- > ----- > > Most people who failed at least one exam > probably > > failed L1. > > This is purely anectodal, but most of the people I > know who took 4+ attempts failed L2. Yeah same here. In my office historically people have gotten stuck on L2 for multiple years if they get stuck at all. Some have taken it 4-6 times I personally doubt they will ever complete the program. In fact I question why they keep going through the torture. Now that I have cleared L2, my co-workers seem to think I already “have my CFA” and it is just a matter of sitting for the *easy* level 3 exam. Unfortunately, word is that L3 is the new L2 in terms of difficulty/differentiating candidates.
MehdiOchre Wrote: ------------------------------------------------------- > frisian Wrote: > -------------------------------------------------- > ----- > > > > Over 45 years, that’s > > > > 108,234 > > --------------- > > 1,025,407 > > > > > > Wow, did you know that 92,000 people took Level > I > > in June 2008? > > > I was more thinking about, not 108234/1025407 = > 11%, but rather 12314/54549 = 23% or 12314/43777 = > 28% (for one given year some years ago). The > process initially suggested 0.40 * 0.48 * 0.76 = > 15% seems too small. > > So maybe the total pass rate is about 20% instead > of 10-15%. Still - four out of five don’t make it. If the current conservative pass rates are L1 40%, L2 40%, L3 50%, the pass rate for 3/3 can’t be higher than 8%. So 92 / 100 don’t make it 3/3.
Multiplying the pass rates is really the only thing you can do that makes any sense, although it’s obviously not 100% accurate it’s a fair proxy.
I gotta’ think that not discouraging your boss from using this approach or assisting him with putting this analysis together would still constitute a violation. I would tread very carefully on this issue… jrumph Wrote: ------------------------------------------------------- > My boss is going to bat for me with our comp > committee to get me a larger raise and/or bonus > for passing L3. > > I passed all three levels on the first attempt > (and as such I’m aware of ethics prohibiting me > from using that as a bargaining chip, but my boss > is not a charterholder and wants to use that in > his argument). He has requested from me an idea > of the percentage of people who pass 3/3. > > Obviously CFAI does not release this. Is there > any better way to estimate this other than the > admittedly flawed (pass rate LI)*(pass rate > L2)*(pass rate L3)? > > Thanks.
I think that approximately 50% of first time L1 passers pass L2 and L3 consecutively. I would put the 3/3 number somewhere in the 20-25% range of all L1 candidates. Of course this is just an educated guess based on a large number of candidates I know.
If you present someone with the facts (passing rate of each level) i think they can draw the conclusion themselves (i.e. Correlation between intelligence and achieving 3/3)… I think 10% would be close to the mark.