Jump to content

Affiliate-selected Board seats/2019/Results

From Meta, a Wikimedia project coordination wiki

The successful candidates were Nataliia Tymkiv and Shani Evenstein Sigalov. A total of 122 affiliates voted, 85% of the 143 eligible to vote, which is a record.

As you know the election was conducted under a variation of the Single Transferable Vote, which meant that prorated votes were redistributed between candidates to come up with the final result. In the 10th step of counting the final place, after Nataliia Tymkiv was elected, was between Shani Evenstein Sigalov (40.519678) and Richard Knipel (40.480322).

We have put the full count narrative here on meta so that others can verify it if they wish.

It is the closest ASBS result for some time, and all candidates brought very valuable perspectives to the work of the WMF.

In the 9th step of counting Reda Kerbouche lost by a very small margin. Adding a ballot with rank #1 for Richard or Reda would result in them being elected instead of Shani. The same goes for removing a ballot. Changing the ranking on one of the ballots in a specific can way can result in a different outcome for the second seat. This is an election in which every vote counts.

As in any election, there is a chance that some voters misinterpreted the instructions and voted wrongly. We don't see a justification for an action as extraordinary and controversial as opening votes for review after the vote period is over. The instructions were visible and clear: "Rank any candidate from 1 (your preferred candidate) to 11 (your least preferred candidate)." After voting, voters received a confirmation email stating the name of each candidate they voted with the number of their rank: Rank 1, Rank 2, ... The agency of voters should be respected. As part of the retrospective we may identify areas of improvements on our side, but still the process was quite simple and documented.

Some voters realized they made a mistake and requested a new ballot. New ballots were issued in those cases. This choice was done because of the specific situation of this election, since the process was complex for new affiliates and participation, diversity and inclusion were a clear goal.[2] We have published on meta information about who got a new ballot within the voting deadline.[3] The Election Facilitators have been available nearly 24 hours a day monitoring the various communication channels to answer any questions affiliates might have. We did our best at answering all of them.

After our own scrutiny of the data, and based on our experience in community processes, we strongly advise the community to respect the integrity of the process, and advise against allowing any modifications of votes at this point. If the votes had been reopened for modification with or without publishing vote results, that would have caused significant confusion and criticism that could have jeopardized the entire election.

We will publish a debrief with recommendations for a next ASBS process on meta.[4] We invite all representatives of affiliates to a feedback session at Wikimania.[5]

We would like to congratulate Nataliia Tymkiv and Shani Evenstein Sigalov and thank everyone who stood.


Ad Huikeshoven, Lane Rasberry, Jeffrey Keefer, Neal McBurnett, Abhinav Srivastava, Alessandor Marchetti

Election Facilitators

[1] "Summary" form of the output from flexstv.py (verbosity of 1 via "-v 1") is given below. This uses a .csv file as input. The full output is written to Affiliate-selected_Board_seats/2019/Results/2019-initials

Counting the votes...

Total votes: 122 Quota: 41 Step 1. NT (29.07); SE (19.07); RKe (15.27); RKn (15.23); MM (10.33); YA (8.50); DS (7.60); TL (6.33); KY (6.07); GS (2.67); CH (1.87)

CH loses.

Step 2. NT (30.28); SE (19.07); RKe (15.32); RKn (15.28); MM (10.33); YA (8.50); DS (7.65); TL (6.50); KY (6.23); GS (2.83)

GS loses.

Step 3. NT (30.78); SE (19.57); RKn (16.95); RKe (15.32); MM (10.33); YA (8.67); DS (7.65); TL (6.50); KY (6.23)

KY loses.

Step 4. NT (32.83); SE (20.83); RKn (17.33); RKe (16.58); MM (10.67); YA (9.00); DS (7.75); TL (7.00)

TL loses.

Step 5. NT (37.31); SE (21.31); RKn (17.81); RKe (16.73); MM (10.81); YA (9.14); DS (8.89)

DS loses.

Step 6. NT (38.33); SE (23.33); RKe (19.33); RKn (19.33); MM (12.17); YA (9.50)

YA loses.

Step 7. NT (39.93); SE (24.43); RKn (22.93); RKe (22.10); MM (12.60)

MM loses.

Step 8. NT (46.58); SE (26.08); RKe (24.75); RKn (24.58)

NT is selected.

Step 9. SE (28.65); RKn (26.37); RKe (25.98)

RKe loses.

Step 10. SE (40.52); RKn (40.48)

RKn loses.

Step 11. SE (81.00)

SE is selected.

A readable pdf of the results file:asbs_presentation_matches_with_stv_py_results.pdf. Per page one step in the proces. In the blue table to left the ballots. In the green table to the right the vote value for each candidate from each ballot. On top of the green table the sum of the total vote value per candidate. Marked yellow is the minimum total value, the candidate to be eliminated for the next step in case no candidate met Droop quota in this step. Marked light red is the the total value that exceeds the Droop quota. That candidate is also removed in the next step. Total vote value could exceed Droop quota. Those excess votes are transfered in the next step by recalculating the vote value. The pdf was made using an Excel spreadsheet to calculate the results. That spreadsheet has been converted to a (read only) Google spreadsheet. The first sheet shows the raw output of Qualtrics.

[2] Affiliate-selected_Board_seats/Resolution_2019

[3] Affiliate-selected_Board_seats/2019/New_ballots

[4] Affiliate-selected_Board_seats/2019/Debrief

[5] wikimania:ASBS_Feedback