Re: Case for/data about elections

On Sun, 18 May 2014 16:26:18 +0200, Brian Kardell <bkardell@gmail.com>
wrote:

> Spinning off a new thread in order to keep the other about actually  
> voting on votes.

Thank you.

> To reiterate in order to pose my questions:  I agree there are potential
> biases in the first system, it has serious flaws.  I entirely support the
> idea that it is worth discussing and probably fixing.

> BUT - I am very dubious that THESE are the biases that have hurt things
> thus far and relatively confident that other biases (apathy/lack of
> participation or knowledge, who actually does the voting, etc) actually
> have had a big impact

That certainly was the case in the past. This is changing (in no small part
through your personal effort, which I applaud). And as it changes…

> AND changing the voting system does not address these.

No, but those are being addressed. E.g. by your efforts. And when that  
happens, the system we have effectively disenfranchises a lot of the  
membership.

> My assertions are easily validated with data.

Given sufficient data of the right type. Which we don't have.

The only data we have available are

the candidates
  The recent trend is for genuinely contested elections.
  AB candidates are more than twice the number of seats.
the winners of elections
the eligible voters

> So my question is:  Is there data to actually support the assertion that
> this has affected outcomes?

There is no data generally available that could clearly validate or  
invalidate that assertion.

W3C have very detailed data on each individual vote - the Team can find  
out who voted what.

However, it is also the case that the data collected for the majoritarian  
system we use are poor-quality for (in)validating an assertion that  
outcomes are affected.

The most reliable indicator we could get is whether tactical voting is  
likely to be reasonably common (i.e. many people vote for 1 or 2  
candidates when there are 4 or 5 seats up for election). W3C has that  
information readily available, but does not release it.

> On several occasions now i have heard people cite recent elections.
> The fact that candidates and folks like myself actively made an effort
> to turn out the vote and collaborated and discussed importance out in
> the open on issues is a perfectly rational explanation, but there is
> adamant insistence it seems that somehow the system is rigged or
> something.

"Something".

I strongly believe it is not "rigged" in the sense that "someone is
cheating".

The system we have now is known[1] to give slates of candidates landslide
victories, leading to a situation where elections tend toward two slates
of candidates (because only irrational actors would bother to run except
on a slate that was likely to win).

It *appears* that this is happening to the elected bodies of W3C,
meaning that they are the candidates selected only by the largest single
voting bloc within the membership. In elections where there are more than
2 candidates per seat, that may well be a minority of those who vote.

> Has there been a questionnaire to membership about whether they
> strategically voted

No. There may be some data available on that but we it is only
available to the W3C Team.

> or any effort to play it out another way that has shown otherwise?

No. There have been requests to do so, but they have been rebuffed.

> If not, let's please stop holding this up as an example if the data
> doesn't support it.

There is no hard data.  There are multiple requests to explain the
problem.

In a situation where hard data is actively withheld, explaining the
problem is difficult and essentially requires looking for indicative
data, which tend to be soft, and require interpretation that is
necessarily subjective (but not necessarily wrong).

> It's disheartening to people like myself who really care and all
> it's accomplishing is creating more "why bother" sentiment.

Yeah, I understand that :( We're between a rock and a crummy place.

> Again, in principle, theoretically - i fully support ranked voting.  I'd,
> love to see coordination with a linked, unofficial poll for studying
> impact, etc.  I'd also love high level preference data to be public -  
> it's useful for making sure membership is able to express itself and
> groups and W3C are addressing appropriately.

Yes. People have asked for such experiments, and been told to explain the
problem first.

Uncircling the wagons in this chain of argument and getting some real
data would indeed be helpful. Sadly, as far as I can tell it won't happen.

[1] http://en.wikipedia.org/wiki/Duverger%27s_law

If you want something shorter, the sections up to the next header at each  
of the following links are a total of 4 short paragraphs.

http://en.wikipedia.org/wiki/Voting_systems#Majoritarian_methods
http://en.wikipedia.org/wiki/Tactical_voting#Plurality_voting
http://en.wikipedia.org/wiki/Plurality_voting_system#Wasted_votes

Cheers

Chaals

-- 
Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
         chaals@yandex-team.ru         Find more at http://yandex.com

Received on Thursday, 22 May 2014 08:09:04 UTC