Hacker News Reader

Comment by ryandrake | Hacker News Reader

Why are you (and many others) just assuming the black candidate is less qualified?

That’s literally what this whole article was about. Removing a high correlation performance test, that black candidates didn’t pass as frequently, and replacing it with a very low correlation questionnaire that provided a more diverse applicant pool while weeding out highly qualified individuals.

They still had to pass the performance test. It was just no longer the first step in the process. I want to be clear, that doesn't mean the questionnaire was a good thing. It just means that the questionnaire did not lower the bar.

Instead it reduced the applicant pool in a sudden and unfair manner, which is it's own issue.

No, read the article again. They didn't need to pass the same test to the same degree - the criteria was also changed to have "qualified" and "well qualified".

It's worth nothing that this change happened before the questionnaire was instituted. (The paper referenced in the article was from 2006, I haven't dug enough to find a date for when this change was made, but the narrative in the article also establishes this act as happening in the '00s.) Additionally, from the Conclusions:

"Reweighting was based on data collected from incumbent ATCSs who took AT-SAT on a research basis; some of these employees achieved overall scores less than 70 (that was one of the reasons for the reweighting effort – a belief that incumbent employees should be able to pass the entry-level selection test)."

I don't think this proves that the update to the test was good or bad in overall competency, but I do think it's worth investigating if the test should be updated when existing employees are unable to pass.

Was it replaced, or was the questionnaire an addition?

Exactly. From the article: "As originally scored, the test was intended to pass 60% of applicants, but predictions suggested only 3% of black applicants would pass"

Reply on news.ycombinator.com