In Schweser’s notes they recommend thinking or the word Horn for Type i errors. The word Horn is an acronym for H0 Reject Null. If you can remember that and you understand that the null hypothesis is that the manager adds no value, then you can remember that a Type i error incorrectly rejects the null hypothesis that the manager adds no value. i.e. Type i, keep managers that do not add value Type ii fires managers that add value. For this question if you want to reduce turnover you’d reduce type ii errors because that is where people are getting fired.
Another point - Sum of Type I and Type II errors is a constant. So reducing I means increasing II, and vice versa. We cannot reduce both I and II at the same time.
ruckmani Wrote: ------------------------------------------------------- > Another point - Sum of Type I and Type II errors > is a constant. > > So reducing I means increasing II, and vice > versa. > > We cannot reduce both I and II at the same time. This is true at a given significance level. Both can be reduced by using a larger confidence interval (higher significance level).
Good point mwvt9! But in this question I am assuming that the null hypothesis has been set, along with the significance level.
I agree with you. I just wrote that for those who possibilty didn’t pick on the correct assumption you made.
kinda helpful to remember Fs: Type II = Fail to reject null, so Fire good managers null of course is manager adds no value
First let’s think what is a REASONABLE null hypothesis? If we assume that it is rare to get a good manager and most are mediocre, then the null is that “this manager is mediocre”. So Type I error would be “fail to reject the null when it is true” and in plain english it means “do not identify (fire) a mediocre manager” and this chance will go up in this question. Type II here would mean " fire a good manager". this chance would go down here. so the answer is B.
I judge REASONABLE by what the cfa says (why swim upstream?). As mwvt9 put it: A, i think. I believe the null is the manager adds no value. So you are more likely to keep a bad manager than you are to fire a good manager? I’m no statistician but how would you judge whether a manager was “mediocre” or not? I believe the null, no value added, is much more easily rejected. Your type two is bang on tho.
i actually typed wrong. my explanation points to A, not b.