You don’t always have the luxury of choosing from among a sample that includes above-median performers.
In that case comparing average members of one sex with the above average members of the other is still not the right comparison to make.
I thought it was clear that the first clause was not controlling for anything, while the second was making a specific measurement of expected performance.
Even this statement is ambiguous. Does the specific measure of expected performance actually screen of gender?
In that case comparing average members of one sex with the above average members of the other is still not the right comparison to make.
Even this statement is ambiguous. Does the specific measure of expected performance actually screen of gender?
You never need to compare the average, because you only ever need to compare a small number of individuals.
Performance in the production environment correlates with the measured expectation equally well for males and females.