Thread by @Scientific_Bird

(1/n) I've been trying to review the evidence of gender bias in the STEM pipeline. It seems somewhat inconsistent tbh.

An interesting experiment I've found (which I don't think many know of) comes from the site interviewing.io, a platform for technical interviewing.

(2/n) Normally female interviewees were accepted at lower rates and were given lower technical scores on average.

To test whether this was due to implicit gender bias, they created a voice modulation tool that masked gender.

(3/n) Contrary to what they expected, they found that masking gender had no effect on interview performance with respect to any of the scoring criteria.

So they didn't really find any strong evidence of gender bias (using this method at least).

(4/n) They started investigating deeper why women were accepted less. They found that women were much more likely to drop out of the platform if an interview didn't go well.

If they factored out people who quit after one or two interviews, the gender disparity went away entirely

(5/n) This isn't peer-reviewed research, but even outside of the "ivory towers of academia", such experiments aren't completely unheard of because gender bias is such a salient topic.

Anyway, thanks for reading. Blog post here:
blog.interviewing.io/we-built-voice…