David R. Henderson  

De-Identifying Race and Ethnicity Reduces Bias

PRINT
Germany is not the problem... Getting Around Economic Sancti...

But not in the way you might expect.

What we found is that de-identifying applications at the shortlisting stage of recruitment does not appear to assist in promoting diversity in hiring. In fact, in the trial we found that overall, APS [Australian Public Service] officers generally discriminated in favour of female and minority candidates. This suggests that the APS has been successful to some degree in efforts to promote awareness and support for diversity among senior staff. It also means that introducing de-identification of applications in such a context may have the unintended consequence of decreasing the number of female and minority candidates shortlisted for senior APS positions, setting back efforts to promote more diversity at the senior management levels in the public service.

This is from "Going blind to see more clearly: unconscious bias in Australian Public Service shortlisting processes," June, 2017. The study was written by Behavioural Economics Team of the Australian Government (BETA).

HT2 Scott Alexander.


Comments and Sharing






COMMENTS (7 to date)
Tennyson writes:
David R. Henderson writes:

@Tennyson,
Thanks. I hadn’t seen it. BTW, it was 2,100 people surveyed, not the 21,000 that Young claimed.

Hazel Meade writes:

IIRC, there is social science research indicating that emphasizing differences makes many people more racist, rather than less. I think I got that from Jonathan Haidt.
This is something many conservatives have had the correct intuition on long before liberals caught up, in rejecting "hyphenated" identities, and insisting that "we're all Americans".
Of course there are other problems with the conservative approach, but the basic idea that it's important to emphasize a common identity rather than multiple possibly opposed group identities is probably correct, at least if your goal is to reduce racial animosity.

LK Beland writes:

Interesting study.

How can we re-conciliate it with the numerous studies that show that if two otherwise identical CVs are sent, the one with a minority-sounding name gets less call-backs?

Is this a public vs private employer effect?

Bob Murphy writes:

That whole summary (from page 5 of their study) is remarkable. They already know what the right outcome is, and they are just going to tweak various stages of their employment process in order to get it. I wonder why they don't just recommend straight-up hiring quotas?

Glen writes:

Not surprising. From 1999 to 2007 I developed and maintained one of the largest performance prediction systems in the world (at least at the time as it now would likely be
considered a toy). In my case there was no difference when certain feature were purposely disallowed. We saw as much or more stereotyping on the part of the applicant as opposed to the employer.

Thaomas writes:

Combating racial and other kinds of biases (like most kinds of regulations) is a tricky, high data-intensive business. Procedure that result in too much compensation for bias (as some kinds of "affirmative action" do) are bad just as those which do not compensate enough.

Comments for this entry have been closed
Return to top