On 30/04/2013 08:49, Sam Hartman wrote: ... > Statistical analysis is only useful if it's going to tell you something > that matters for your decision criteria. Yes. And I would like to know, in statistical terms, whether there are significant differences between (for example) the M/F ratios among (a) IETF registrants, (b) active participants, (c) WG chairs & secretaries and (d) I* members. [Discussion on the objective definition of active participation is deferred for now.] The null hypothesis would be that no significant differences exist. If that turns out to be true, we know that our problem is only lack of diversity among registrants. If it turns out to be false, we know that we have an internal problem of some kind as well. Brian