From a statistics point of view, this statement is dispositive: "...decrease the false negative rate, not increase the false positive rate."
The interaction between type I and type II errors virtually always have an inverse relationship. You cannot decrease one without increasing the other. It is, for example, one of the difficulties in making good medical diagnostic tests.
P.S. snarky commentary such as 'read the memo yourself', implying I'm lying about having read it, doesn't contribute to civil discussion.
> The interaction between type I and type II errors virtually always have an inverse relationship.
If you're adjusting a dial, yes. But hiring is complex, and having more time to evaluate candidates means you can do a better job. Spending more effort is a simple way to improve false negatives and false positives. It's entirely plausible that a company might have extra time budgeted for double-checking resumes from certain groups.
The interaction between type I and type II errors virtually always have an inverse relationship. You cannot decrease one without increasing the other. It is, for example, one of the difficulties in making good medical diagnostic tests.
P.S. snarky commentary such as 'read the memo yourself', implying I'm lying about having read it, doesn't contribute to civil discussion.