by RonPurewal Mon Oct 13, 2008 6:55 am
CAVEAT LECTOR: i'm going to assume that "either I or I" is supposed to say "either F or I". proceed accordingly.
--
one way you could do it: number picking. just try different numbers of P's and I's/F's, and see whether the answer is invariant or whether it changes. if it changes at all, then the data are insufficient.
* let's try 100 p's and 100 f/i's
among the 100 p's, there are 60 males and 40 females.
among the 100 f/i's, there are 80 males and 20 females.
so 60 out of 200, or 30%, are female.
* let's try 1000 p's and 100 f/i's
among the 1000 p's, there are 600 males and 400 females.
among the 100 f/i's, there are 80 males and 20 females.
so 420 out of 1100 are female. this is a weird percentage, but it's clearly not 30% (which would be 330 out of 1100), so, insufficient.
--
you could also do it by realizing that this is a WEIGHTED AVERAGE of 40% (the percentage of p's that are female) and 20% (the percentage of f/i's that are female). like any other weighted average, this one depends on the relative quantities of its components (here, p's versus f/i's), quantities that are not given in either of the two statements.
conceptually, if there are huge numbers of p's and negligible numbers of f/i's, then the percentage of females will be very close to 40%. conversely, if there are huge numbers of f/i's and very few p's, then the percentage will be very close to 20%.
in fact, there's nothing prohibiting the consideration of the two most extreme cases possible:
* ALL grades are p's --> 40% female
* ALL grades are f/i's --> 20% female
these situations are both allowed under the given conditions, so, insufficient.