[Foundation-l] for the future...

[Andrew Gray]

On 05/07/07, Peter van Londen <londenp at gmail.com> wrote:
> offlist?
>
> The results for this voting will be biassed, because of this action of
> yours. I sincerely doubt if the results can be accepted. In most countries
> it is forbidden to campaign on election day.....

Yes. In those countries, it is forbidden for the candidates or their
parties to campaign. It is not forbidden to tell people there is an
election and encourage them to vote - indeed, it is encouraged, if it
can be done tactfully and neutrally.

Otherwise, one wonders how those countries have anyone remember to go
to the polls!

> I always thought that the best persons should be on the board, not the
> representatives of the strongest project; in the end it all turns out to be
> political.... I always hoped that wikimedia would have grown beyond that.
> Too bad.

It is worth, incidentally, noting that whilst enwiki is the largest of
all the projects, it is not immensely so - it only has around 50% of
the community members. It's very hard to get solid numbers on this -
if nothing else, if people are a member of more than one community,
enwiki is very likely to be one of those - but 50%-55% seems to be the
right sort of number.

[Greg: do we know what rough proportion of the overall votes are from
enwiki? I think you mentioned a number at some point, but I don't have
a note of it.]

Now, let's consider what votes-from-a-project means. Either:

a) People vote strongly based on project affiliation, so the enwiki
electorate will vote heavily for "internal" candidates. If we assume
this, then sharply increasing enwiki turnout will indeed sharply
increase the voteshare for enwiki candidates - but conversely, keeping
enwiki turnout artificially low will strongly *penalise* those enwiki
candidates. In this model, enwiki gets about its fair share of
"influence" if it has 50% of the final voters.

-or-

b) People vote only loosely based on project affiliation, so the
enwiki electorate will vote for a mixture of internal and external
candidates. If we assume this, then sharply increasing the enwiki
turnout will only result in a small proportional voteshare increase
for enwiki candidates, and keeping it artificially low only mildly
penalise them. In this model, again, enwiki gets about its fair share
of "influence" if it has 50% of the final voters, but the electorate
numbers from each specific project are much less critical to the
overall outcome of the election.

I am guessing you believe in model (a), from your remarks. In this
case, what do you think of the proportional turnout graph?

http://myrandomnode.dyndns.org/turnout.png

Note that when the email was sent, enwiki had an abnormally low
turnout rate - half that of nlwiki, a third of nowiki, and in the dust
compared to plwiki or itwiki.

As such, saying "enwiki shouldn't have had drastic measures taken to
get more people voting" is effectively the same as saying "enwiki
voters should have far less relative voice in the elections". Which is
a defensible position, in its way - rights of the minorities, etc -
but not one anyone seems to be actually putting forward in that form.

-- 
- Andrew Gray
  andrew.gray at dunelm.org.uk