Makes sense, I guess, but when results seem to cluster so close to the 50% boundary it probably makes the answer a bit too random....
Agreed. That's why they give you two different results--the non-curved and the curved results (relative to others who've taken it). For the ideal test, these numbers would match exactly, but for most tests I've seen they're pretty different.
One negative consequence of having the cutoffs all at 50% is that there is a critical point at 50-50-50-50, at which you're right on the cusp of all 16 possible results. But the alternative is to test variables for which you expect to have a larger portion of the population on one side or the other... which makes it seem kind of "biased" in a way, although the tests usually end up this way empirically anyway.
no subject
Makes sense, I guess, but when results seem to cluster so close to the 50% boundary it probably makes the answer a bit too random....
Agreed. That's why they give you two different results--the non-curved and the curved results (relative to others who've taken it). For the ideal test, these numbers would match exactly, but for most tests I've seen they're pretty different.
One negative consequence of having the cutoffs all at 50% is that there is a critical point at 50-50-50-50, at which you're right on the cusp of all 16 possible results. But the alternative is to test variables for which you expect to have a larger portion of the population on one side or the other... which makes it seem kind of "biased" in a way, although the tests usually end up this way empirically anyway.