3 ms·
The problem with using diversity as a selection criteria bias means that statistically on average the diverse speakers will be worse. While attending such a con
by abz10 10y ago
The problem with using diversity as a selection criteria bias means that statistically on average the diverse speakers will be worse. While attending such a conference, if one were to optimize for the quality of speakers a simple rule to achieve this at such conferences would be to avoid the diverse speakers and pick a different track.
I avoid 'diverse' speakers, not because I'm racist, sexist, transphobic etc but because I don't like having my time wasted.
- mcphage 10y ago> that statistically on average the diverse speakers will be worse I mean, since you're being technical, then it actually means that the speakers will be slightly worse at writing talk proposals. But ability to write a talk proposal isn't the same as ability to give a talk. And even if it were, it's meaningless—when doing a conference, maximizing the individual speaker "quality" isn't helpful either. 10 speakers speaking about the same topic amazingly is worse for an attendee than 10 speakers speaking about 10 different things merely very well.
- abz10 10y agoI was not proposing alternative metrics for measuring quality but I believe it should be safe to say that talk proposals are a better indicator of talk quality than race and gender. Nor was I recommending maximizing individual speaker quality over maximizing conference quality. To reuse your two metrics, I would expect 10 different talk topics given by people of the same race and gender to yield more value to attendees than the 10 same talk topics given by people of all different race and genders.