How do these scores make sense? You can get a 50 on AM, a 65 on PM and still pass?
These are “as posted” on AF
I even have the username associated with each score … and you can always go back to the original thread and check for yourself
Also there would be some difference between what 40/60/80 is indicating and what the actual scores were
could you give us the link to the thread?
and what’s 40/60/80 rule btw?
thx
the bottom line here i think that is its gona be a coin flip… 50pc of us will pass
From my understanding, in the result slip, subjects marks below 50 are assumed to score “40” and marks between 50-70 are assumed to score “60” while marks above 70 are assumed to score 80%. A weighted average method is subsequently used. The estimate indirectly puts a 40% floor and 80% ceiling but as a method of approximation, it can be pretty decent (i.e.: use it as an estimate but don’t rely on it).
These scores are from this thread http://www.analystforum.com/forums/cfa-forums/cfa-level-iii-forum/91324255
The 40/60/80 method is explained well in the post above.
From my experience in Level 1&2 the 40/60/80 method gives a fairly good approximation for a Multiple Choice exam.
However for AM as I stated earlier … It tends to overestimate the scores marked as “less than 50”
Some more details … for the ones interested
Number of scores = 91
Pass = 80
Fail = 11
Overall Median AM % = 54.8%
Overall Median PM % = 72%
Overall Median Total % = 62.5%
Median scores of only those who passed
AM % = 55.7%
PM % = 72%
Total % = 63.1%
Thanks for the data. You guys must have worked very hard for the figures.
My inputs in the 40/60/80:
- For me, for the level 3’s PM, we can even make it more accurate:
For example, a set of 6 question: there is only one scenario that is between 50-70, which is 4/6=67%
Similarly with >70%: only have 5/6 (83.3%) or 6/6 (100%)
So why 60/80? Why not just 67/90 for instance?
I would agree, it’s not appropriate to model morning scores in the same way as a multiple choice paper. From my experience of mocks, I found the morning paper could be quite binary, i.e. i would get all the points or none, where as in the multiple choice your worst case is a 1 in 3 throw the dice scenario.
Another thing to thnk about; at 300Hours they do a results analysis and for those of you who’ve not seen it it suggests the difference between bands is about 1 or 2% per band (i.e between band 9 and Band 10) and that the MPS (in 2012) was between 49% and 68%.
All this goes to show is that performances between 1 candiadate and the next are very tight and that you cannot look back and say I crushed the morning or PM was awful. Frankly after at least 2 exams already we should know that post exam sentiment deos not equal actual performance.
Hope you all did well, though not as well as me, I’d like to pass. In the meantime try to forget about it, it’s only going to make it worse as the actual results day comes. Probably 12th August.
\
Actually, theoretically, there are two scenario for each of “<50”, “50-70” and “>70”.
There are two scenarios where we can get 50-70 (i.e. 3/6 = 50% and 4/6 = 67%). Taking an average of the midpoint, we get the rule of thumb numbers of 25/58/91 instead. That said, as a rule of thumb, nice rounded numbers tend to be prefered and 40/60/80 is pretty good as a rule of thumb assuming that candidates score around the average (the 25/58/91 assumes a more dispersed score).
I am very sure that 3/6 gets counted in the <50 bracket … coz the result email clearly indicates this bracket as =<50.
I agree that 40/60/80 method can be fine tuned … but my goal was not to deduce the actual MPS as accurately as possible … rather i just wanted to get a sense of the combinations of AM/PM that pass/fail.
Also i had done this number crunching sometime in Dec/Jan … as i realised that the ease of scoring in AM and PM is very different and wanted to set targets for each … i had concluded that 55-60% in AM and 70-75% in PM should see you through
That is interesting. I never noticed that myself (probably too happy celebrating?). Incidentally, do the AM / PM combination matter? I was under the impression that you will likely pass / fail if your aggregated percentages exceeds / falls short of the minimum passing score (with a small Ethics adjustment there)?
You are right … AM & PM are weighed equally and the MPS is based on the aggregate score.
What i realised after looking at some of the score matrices of last year’s Level 3 results (and other years too) was that there is a noticable difference between AM & PM scores of most candidates … with PM generally being higher than AM … this is evident from the difference of ~17% between median AM & PM scores
I was just trying to get a sense to this difference …
After the exam all people around me were really confident and told me ahh it was easy. Think I did not it. Lost around 15 18 points in the morning session due to skipping some small questons or not answering complete and making mistakes and already know around 5 mistakes (questions) in the pm session.
how can someone predict their score with a decent confidence margin (i.e. 90%+)??? do you guys remember every single question, or most of them??? i find it laughable when i see someone saying i did 50-55% in the AM and 70-75% in the PM. there’s a 95% chance your actual score is outside those margins…
what confidence do you guys have in your “scores?” i ask that honestly, for my curiosity…
based on what i remember (like half of the test), what i did well and what mistakes i made, i am 95% certain i got above 70% on the AM and above 45% on the PM…. however, i am 75% confident I got above an 80% on the AM and above 60% on the PM. i can’t really put a cap on my “estimate” as, who knows, maybe i got an 85% on the AM and half the guesses i made were correct and got a 70% on the PM… highly unlikely.
out of the roughly 40-50 questions/parts in the AM, i only left 3 parts blank (about 11-12 pts total) as i circled to come back to and didnt have time. about 30 questions in the AM i am positive i got 99% of the points available, which leaves about 10 questions or so that i felt decent about my answer, felt i answered what was asked, but through the quick pace maybe i missed something, a trick, or didn’t answer completely what they are looking for - i’m assuming i got 50% of the pts awarded for those.
out of the 60 PM questions, i only remember about half of the questions. from the 30 or so i remember, I know i got 20 questions correct, and i know i got about 10 incorrect. the remaining 30 i simply just don’t recall. if i got half of those correct that leaves me at an overall avg of 58.5% for the PM section…. maybe i got more right, maybe less, who knows.
just curious what you guys are assessing in coming up with your “estimated” scores?? my confidence band is pretty wide. i’m certain i got between a 65-85% on the AM, and a 45-65% on the PM. but that does nothing for me in easing my tension or giving me a better idea of my score or likelihood of passing. if it’s toward the lower end, i’ll fail miserably… towards the higher end, i’ll obviously pass… in the middle, i’m on the bubble.
suggestions?
Suggestion - estimating your score is pointless. Ride out the wait and find out for sure on August 12th.
junior if it makes u feel any better thats pretty much how i feel too… feels like 65 ish which means were at the mercy of the mps
This synopsis perfectly encapsulates my feelings after the test. I derive my confidence level by comparing how I feel after I complete a test question vs how I feel after completing a mock. If I was able to answer a question confidently, completely, and with a high degree of certainty, I’ll probably wind up getting that question right at least 7 out of 10 times. I assume that at best I get half of my educated guesses correct, and I assume that if I had a high degree of uncertainty I got it wrong. For me that equates to on the bubble.
a comforting factor is that on an average people have left 15 marks in AM
also rem 1 q in PM = 3 marks
so even 2-3 correct q can balance a big 9-10 marks AM q