News for hackers. Stuff that matters.
Re:Brief thoughts on the “Google memo” by mLuby on Friday August 11, 2017 @04:33AM (#14987690)
I'm simply making an ethical claim. Any` skew sounds unacceptable to me.

If we agree that any` skew is unacceptable, I'd suggest we take advantage of leverage and seek the greatest skew correction per resources allocated. That might well turn out to be in construction workers…

`statistically significant

Re:Brief thoughts on the “Google memo” by gfodor on Friday August 11, 2017 @02:52PM (#14993013)
I guess there is also an inductive problem with not accepting any skew whatsoever. For example, the field of medicine has minimal skew, but if you drill into specialties, there is massive skew. At what level of granularity does skew stop mattering? It has to become noisy at some point, but good luck deciding on a concrete and consistent metric for determining where that is.