"Brian" == Brian E Carpenter
<brian(_dot_)e(_dot_)carpenter(_at_)gmail(_dot_)com> writes:
Brian> The null hypothesis would be that no significant differences
Brian> exist. If that turns out to be true, we know that our
Brian> problem is only lack of diversity among registrants. If it
Brian> turns out to be false, we know that we have an internal
Brian> problem of some kind as well.
Yes.
I'll admit that that particular question--which is far more involved
than the numbers I've seen thrown around to date--is somewhat
interesting.
Although while it influences how I'd think about deciding on proposals,
there's no answer to that question that has a clear set of decisions for
me, even ignoring questions about methodology, definitions of
participants, etc, etc.
1) I may believe that increasing diversity among leadership so that the
leadership is more diverse than the population as a whole will help
increase diversity of the population.
2) I may believe that the diversity of the leadership is more of a
problem in terms of either quality of spec or credibility of
organization than diversity of the participants/registrants.
But you've definitely started to get into a realm where the statistics
are more interesting to me.
And i'll drop this now, because I realize I'm only one participant and I
discussion that doesn't provide helpful information for me may well
provide useful information for others.