Re: Disclosing election results -- a voice of caution

On 6/2/2014 6:24 PM, Michael Champion (MS OPEN TECH) wrote:
> Forgot to respond to one Brian’s question:
> > what sorts of questions would we be able to ask
> Assuming we collect more data than we use in the current vote counting 
> algorithm, questions might be:
> - What percentage of members voted?

98 Members voted which is more than 25% of the Membership. Personally, I 
am quite pleased that there was this level of interest.  While not the 
90+% that I would have preferred, the 25% is still larger than some 
democratic political elections in some locations.  They voted for 355 
candidates, an average of 3.6 votes per Member.

> -  What percentage of people voted for the maximum number of candidates?

51 of the 98 voted for 5 candidates.

> How about only 1 candidate?

22 of the 98 voted for 1 candidate.

> - Assuming that voting for one candidate is an indication of strategic 
> voting, were there obvious patterns in the voting blocs?  OK that’s 
> not a “statistical” question but we could ask the Team whether it was 
> obvious to them on what was most similar about the people who voted 
> for the same single candidate – Being in the same geographic region?  
> Position on a controversial issue?

I don't think that I can provide this information.  Voters have a 
reasonable expectation of anonymity.  Providing data about voting 
patterns jeapordizes anonymity.

> - What percentage of voters were willing to go to the trouble to rank 
> order candidates?  For those that did, would an STV system have 
> changed the outcome?  Which STV schemes would have created the most 
> diversity (geographical, member size, business type)?

I don't think we have this information.

> - For those who categorized candidates as acceptable / not acceptable, 
> what percentage of people listed more acceptable candidates than open 
> slots?  Did a majority of voters find all candidates acceptable or 
> were they more discriminating?

I don't think that people had the opportunity to categorize in this way.

> *From:*Michael Champion (MS OPEN TECH)
> *Sent:* Monday, June 2, 2014 2:18 PM
> *To:* 'Brian Kardell'
> *Cc:* Bassetti, Ann; Daniel Glazman; Charles McCathie Nevile; L. David 
> Baron;
> *Subject:* RE: Disclosing election results -- a voice of caution
> > Anything less than 'open' as elections in <stick just about any 
> country here>' seems like it yields a question of - why is that necessary?
> Let’s start with the reasons Ann gave … these are not positions that 
> offer power or financial rewards to the winners, so why drive the 
> inevitable losers away by exposing potentially embarrassing 
> information about how little support they got?
> *From:*Brian Kardell []
> *Sent:* Monday, June 2, 2014 2:11 PM
> *To:* Michael Champion (MS OPEN TECH)
> *Cc:* Bassetti, Ann; Daniel Glazman; Charles McCathie Nevile; L. David 
> Baron; <>
> *Subject:* Re: Disclosing election results -- a voice of caution
> On Mon, Jun 2, 2014 at 5:06 PM, Michael Champion (MS OPEN TECH) 
> < 
> <>> wrote:
>     Ann, if the Team kept the raw data confidential but answered
>     **statistical** questions from the AB/AC/Process CG would that
>     raise any concerns?
>     Brian, I know that’s not what you want, but would it address your
>     most important concerns?
>  If you can ask the right sorts of questions I suppose that would be 
> fine - but what sorts of questions would we be able to ask?  I mean, 
> wouldn't you expect me to immediately tend to just ask for the same 
> statistics?  What % of members voted/what % did each candidate get? 
>  :)  From that you could derive the suggested anyway, right?  Anything 
> less than 'open' as elections in <stick just about any country here>' 
> seems like it yields a question of - why is that necessary?
>     *From:*Brian Kardell [
>     <>]
>     *Sent:* Monday, June 2, 2014 1:32 PM
>     *To:* Bassetti, Ann
>     *Cc:* Daniel Glazman; Charles McCathie Nevile; L. David Baron;
> <>
>     *Subject:* Re: Disclosing election results -- a voice of caution
>     On Mon, Jun 2, 2014 at 3:59 PM, Bassetti, Ann
>     < <>> wrote:
>         Background information about me:  For the first time since
>         1999, I was not in this election.  I am American, hence
>         familiar with one-vote-per-person.  I always vote.  I am quite
>         outgoing and social, and not easily embarrassed. I like data.
>         I value openness.
>         Given all of that:  If I had been in this election and the
>         votes were made public, there are many scenarios I would find
>         embarrassing.  Such as: if I was not elected by a wide margin;
>         or, if I was barely elected; or, if I got very few votes at
>         all; or, etc.  Any scenario that would indicate I'm 'not
>         liked' by the group, would be embarrassing.  Further, I can
>         imagine my management would pay attention to that data, and
>         add it to my grade.
>     I'm not trying to belittle what you are saying, I appreciate your
>     comments - but I think that W3C is the exception rather than the
>     rule here and in a bad way - especially because we like to think
>     of ourselves as 'open'... It seems hard to see how it can even
>     'work' (and I would argue it hasn't worked nearly as well as it
>     could have) without this sort of information... Imagine for
>     example that all these years no one noticed that there is no
>     'minimum number of votes rule' and - because it is dysfunctional -
>     only 3 or 4 mega-companies even bother voting.  If you can see
>     that data there will be a clamor to fix it.  Without that, then
>     what?  Similarly, if 350 orgs vote overwhelming for a shared
>     position held by numerous elected candidates - that _means_
>     something... People should take notice.  But we really have no
>     idea about any of this other than some vague anecdotal evidence.
>      Open is better.
>     Plus, this isn't new - this is how elections work in every country
>     I know of - even in ones that have elections no one trusts - basic
>     results are known - in part for the opposite of the reason you
>     cite (and actually my own want to do it is based on this) - we get
>     some kind of indication about what members support and don't.
>      Even votes for elementary school president or something provide
>     this level of data in my experience.
>     I'm not sure how to get past that except to say that if someone is
>     worried that not getting votes will define them to that extent, I
>     would suggest that running probably isn't a good idea.  Assuming
>     the things you say - it seems like either way you'd be likely to
>     have similar feelings if you lose.  Personally - I *AM* actually
>     easily embarrassed and not especially outgoing, but I recognize
>     that there are any number of factors to an election besides being
>     'liked'.  I like people I wouldn't vote for at this juncture in
>     time, I prefer some candidates to others (some by a lot, some by a
>     little) - and we all know that, for example, you're more likely to
>     vote for someone you are reasonably familiar with than someone you
>     aren't regardless of what's said in a single statement.  You could
>     say the same about participating in standards in a way - if you
>     post something to a list, you might inadvertently illustrate your
>     ignorance - not because you are flawed as a person or in a mean
>     spirited way, but it is possible for you to build the same sort of
>     "I won't ever speak because it might embarrass me" argument,
>     employer and all.   I don't think that this is how _most_ people
>     look at it and I've never really seen that to be the case.
>         I suggest to all of you who are pushing hard on this, that you
>         should consider people's feelings; consider cultural values
>         other than your own; consider people who are quieter than you;
>         consider people's jobs; and so on.  While there may be value
>         in honing a better voting system (about which you already know
>         I am skeptical), I would not want that value to be at the
>         expense of the human 'costs' described above.
>         Much of the tenor of this voting 'push' makes me want to
>         withdraw, not participate. If *I* feel that, I can only
>         imagine others may feel the same.
>         For these reasons, if data is released, I strongly urge it be
>         anonymized.
>     DEFINITELY everyone is suggesting that it be anonymized in terms
>     of who voted for whom - it sounds like you are suggesting further
>     anonymity?  What would that data even look like?
>         -- Ann
>         Ann Bassetti
>         The Boeing Company
>         mobile: + <tel:%2B1.206.218.8039>
>         email: <>
>         *From:*Brian Kardell [
>         <>]
>         *Sent:* Monday, June 02, 2014 12:27 PM
>         *To:* Daniel Glazman
>         *Cc:* Charles McCathie Nevile; L. David Baron;
> <>
>         *Subject:* Re: Disclosing election results (was Re: Result Re:
>         Call for Consensus - "Use 'Schulze STV' for voting")
>         On Mon, Jun 2, 2014 at 3:21 PM, Daniel Glazman
>         <
>         <>> wrote:
>             On 02/06/2014 21:06, Charles McCathie Nevile wrote:
>             > I'd be happy to have the pattern data, but not the
>             candidate names - i.e. anonymize them so we can't figure
>             out who romped in, who scraped in, and who
>             > was beaten out by a single vote - or only got 1.
>             Then I disagree. Publishing anonymized data is not useful
>             to people
>             not drastically involved in W3C Process. I suggest then
>             W3M shares
>             *all* election data with the AB, in full confidentiality.
>             I don't
>             even know if it's already the case today or not, and that says
>             something about the opaqueness of our electoral system...
>             The AC would get, as I said earlier, number of votes
>             globally and
>             per candidate and that would be enough IMHO.
>             (please note that even if the votes are ballots, the
>             results are
>              counted per person)
>             </Daniel>
>         I assume that the actual system stores 'ballot' records, I'd
>         like to propose that those are exported anonymously - it is
>         possible to glean slightly more data that way and certainly no
>         more difficult for a reasonably intelligent person to create a
>         'count' for each candidate even in a simple csv which doesn't
>         provide that directly.  I'm reasonably sure that within an
>         hour or so of release, someone will re-post with counts if not
>         provided.
>         Let's not overcomplicate things, just keep it simple :)
>         -- 
>         Brian Kardell :: @briankardell ::
>         <>
>     -- 
>     Brian Kardell :: @briankardell :: <>
> -- 
> Brian Kardell :: @briankardell :: <>

Received on Tuesday, 3 June 2014 14:13:55 UTC