Re: [Gendispatch] New Version Notification for draft-knodel-nomcom-gender-representation-00.txt

[Christian Huitema <[e-mail]>]

.

On 11/22/2023 11:21 AM, Brian E Carpenter wrote:
> 
> On 22-Nov-23 17:38, Martin Thomson wrote:
>> On Wed, Nov 22, 2023, at 15:25, Brian E Carpenter wrote:
>>> Yes, but if (hypothetically) we had four possible responses: Male,
>>> Female, Other, Prefer-not-to-say,
>>> I really don't know what to do with Prefer-not-to-say.
>>
>> I have no strong preference here, but it could be reasonable to treat 
>> other and pnts as distinct categories.  Or both together.  Based on my 
>> understanding of demographics, it is unlikely that either (or both) 
>> would be selected for more than 8 or 9 out of 10, so while we need to 
>> write up firm rules, that would be the only reason we'd need that 
>> level of precision; it's unlikely to be a problem in practice.
>> .
> 
> As Mallory's draft is written, "other" would not be a problem, it would 
> simply be the third gender. Of course it conflates everybody that 
> doesn't identify as male or female, but it will serve its purpose of 
> favouring diversity.

Mallory's draft addresses a classic problem in statistics: the law of 
big numbers does no apply to small numbers. I heard 16.5% quoted. That 
would imply an average of 1.65 women selected, but of course the nomcom 
does not select fractions. Based on random chance, there will be 0, 1, 
2, 3, rarely 4, and even more rarely more women in the selection. 0 
would happen about once ever 6 years.

We may agree that 0 is undesirable. We can also observe that the cost of 
assuring at least one seat is not very high. The average number of women 
selected over many years would increase from 1.65 to 1.81, the number of 
men would decrease from 8.45 to 8.19, not a huge distortion. If we 
reached 25% or more female participation, the distortion would become 
barely noticeable.

The cost would be much higher if we started reserving seats for smaller 
minorities. The distortion is only limited because women, while being a 
minority of participants, are a substantial minority. Going beyond that 
would raise very divisive questions about fairness and equality of 
chances, the very issues that our random picking system was meant to bypass.

We could also decrease the "small numbers are not large numbers" effect 
by increasing the size of the nomcom. If there were 16 members selected 
for the nomcom instead of 10, the "zero women" effect would happen about 
three times less often. The representation of small minorities or small 
regions would also be more frequent. This might make the nomcom harder 
to manage, but that's a natural outcome of representing more points of view.

-- Christian Huitema