Talk:Wikimedia Foundation elections/Board elections/2011/Results/en

Interpreting results

Now begins the phase of trying to learn as much about the results as possible.

Right off the top of my head-- there are three main "voting strategies" for evaluating incumbents:

• Strategy 1: Re-elect incumbents as a group-- top 3 should be all incumbents
• Strategy 2: Re-elect some of the incumbents, elect some non-incumbents. Top 3 should be a mix of incumbents and non-incumbents.
• Strategy 3: Anti-incumbent. Top 3 are all non-incumbents.

Any chance we could get rough figures for what percentage of voting pattern fall into each of those criteria? The ratio of Strategy 1 to Strategy 3 is the best measure of "global voter foundation approval" we're likely to get any time soon, and my instinct is that it's quite high. --Alecmconroy 16:01, 17 June 2011 (UTC)

In the past the (anonymized) ballots have been made available. I've argued that there are risks in that, but the information is really invaluable to people trying to better understand the elections and voting systems in general. There are really an infinite number of questions you can ask that have implications on how the system is built. For example: How many people ranked only a single candidate, how many ranked only three? I hope the ballots are eventually made available. --Gmaxwell 17:29, 17 June 2011 (UTC)

I agree this would be most valuable. Also, a flag for "before the email" and "after the email" would be an additional highly useful bit of information, as it breaks the pool of votes in half with a clear intervention in the middle. –SJ talk | translate   23:30, 18 June 2011 (UTC)

You know, I think I preferred it when mvart4u was still included in the results :P --ErrantX 22:16, 20 June 2011 (UTC)

Please re-add mvart4u. Markus Schulze 07:22, 21 June 2011 (UTC)

Complete votes dump

Board elections/2011 says that «Detailed results will be available»; can we have the full dump as in 2008? Statistics are requested as well (please provide Pathoschild the data if you can't do it yourself). Thank you, Nemo 17:32, 17 June 2011 (UTC)

I would be fine also if you could just provide # of votes for all the candidates (just in order to be totally horrified :D ) --Elitre 19:14, 17 June 2011 (UTC)

What does that mean? It's a ranked ballot. All voters voted for all candidates (Though you might have voted for them all tied in last place). --Gmaxwell 19:22, 17 June 2011 (UTC)

Pairwise results, as already provided for the first 10 ranked. --Elitre 21:48, 17 June 2011 (UTC)

Agreed. The original "results" page definitely expects a dump, and we surely should have it for the sake of fairness. Deryck Chan 20:34, 17 June 2011 (UTC)

VERY happy to hear this! A complete ballot dump, properly sanitized, would be an _invaluable_ dataset. One of a zillion examples-- we could learn about improving the actual UI for voting-- How prevalent are ballots that fell victim to the "rank top, rank bottom, leave rest blank" potential-bug? That's how we'd gauge how important drag and drop is. We could look and see who had the time to rank all candidates individually vs tier voting. If you think I'm happy, the burgeoning community of social scientists who study us (the researchers) are going to be even happy to hear about this dataset than I am. --Alecmconroy 21:02, 17 June 2011 (UTC)

Alec: even if the ballot dump is available, I don't think you can tell whether a person voted A first, B [effectively] second and the rest don't care, or voted A first, B [assuming it is] last and the rest don't care! The voting page has already explained clearly that any candidate you voted blank is given an equal preference which is below all candidate which you've numbered. Deryck Chan 22:29, 18 June 2011 (UTC)

Such is the joy of statistics. You can never know with certainty that any GIVEN voter experienced the bug, but the totality of all the ballots will allow us to estimate the chance that a portion of the voters did experience it. It's not going to pop out as a field in the data-- but that sort of information is there, hidden in the data. --Alecmconroy 14:41, 2 July 2011 (UTC)

...how? Deryck Chan 06:53, 4 July 2011 (UTC)

{editprotected} request: typo

{{editprotected}} Typo: "3495 votes case" should be "3495 votes cast". ~ Ningauble 18:13, 17 June 2011 (UTC)

Fixed - Thanks for the info. -Barras 18:42, 17 June 2011 (UTC)

Translation

Would Jon Harald Søby care to explain who are Andrew, Tim (Starling?) and Philippe (Beaudette?) whom he referred to in the results? Deryck Chan 20:59, 17 June 2011 (UTC)

They all are employees of Wikimedia foundation whom helped us alot during the election, Andrew and Tim were were our technical people and Philippe Beaudette was coordinator Mardetanha ^talk 22:25, 17 June 2011 (UTC)

I linked to the userpages. Andrew Garrett provided technical help, Tim Starling did it in the past and was a resource, and I was the staff liaison to the committee (and a former three-term member of the committee). Philippe (WMF) 04:36, 18 June 2011 (UTC)

Could someone explain to clueless me what exactly "pairwise results" means? The number of people who ranked Ting over SJ and vice-versa? NW (Talk) 18:40, 19 June 2011 (UTC)

Yes. For example, here are the pairwise results for the 2009 elections. Markus Schulze 19:12, 19 June 2011 (UTC)

Post mortem

(Really, connecting people and pages does not look like our cup of tea). Board elections/2011/Post mortem. --Elitre 15:49, 21 June 2011 (UTC)

Mischa Vetere

He disappeared from the results. What happened to his candidacy? Deryck Chan 07:02, 4 July 2011 (UTC)

Mischa Vetere withdrew his candidacy on 10 June 2011 (link). However, in my opinion, he should be included in the results, since his name was on the ballot till the end of the voting period. Markus Schulze 07:47, 4 July 2011 (UTC)