Hacker News Reader

Comment by ghort | Hacker News Reader

> The initiatives that the author links to in this section are presumably aimed at reducing the false negative rate for women, and thereby lowering the bar that women need to clear to get hired.

You (he) can't say that you are lowering the bar since a false negative is positive like any other positive. Widening the bar is a more apt analogy.

View on HN

Reducing the false negative makes getting a job "easier" (a higher percentage of applicants succeed) and therefore is lowering the bar. I understand your point about how this metaphor may not be exactly the right language to use, but that seems like a trivial difference in semantics and not content.

I don't think it's a trivial difference. Lowering the bar suggest that people are now being accepted on lesser merits. But a false negative is a binary condition. Both positive and false negative results are over the bar. False negatives are positive results we misjudge. If he is suggesting something else he should say that and support it. I would define "a higher percentage of applicants succeed" as "more likely".

Imagine that applicants were scored between 1 and 100 and anyone with a score over 90 is offered a job. Suppose that being nervous in an interview gives you a -5. A nervous applicant has a higher bar - a 95.

Someone says maybe we aren't hiring enough women because women have been socialized not to be engineers and are therefore nervous when applying. Let's work on a program to help women be less nervous, then we'll hire more.

Now the DM author says "let's apply that program to everyone, not just women, otherwise we'll be lowering the bar for women." And Google immediately fires him. Your observation is that he shouldn't have said "lowering the bar" but it's objectively true.

In this example, the bar for women is 90 + the likelihood of being nervous * 5. If you remove the second term for women and not men, you are lowering the bar for women.

Reply on news.ycombinator.com