Proposed changes to the Spokesperson election procedures Report - - PowerPoint PPT Presentation

CMS week - CB 120 Proposed changes to the Spokesperson election procedures Report from the Reflection group

Report from the 
 reflection Group


2nd of October 2018

1

……

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

Reflection Group about 
 CMS Spokesperson election procedures

▪ In February the CB had a discussion 
 about the length of the SP’s term and 
 the election procedure.

▪ We decided to proceed with gathering 
 wider views on the SP election procedure. ▪ A survey was conducted, the results were 
 presented in the April CMS week and 
 CB meeting.

▪ A Reflection Group (RG) was created in April 
 and asked to consider the topic and work 


• ut a proposal.

▪ In June the RG presented some findings. ▪ Now the RG presents a proposal for 
 collaboration for wide discussion and to collect feedback.

2

Reflection Group Membership ▪ DE: Thomas Hebbeker ▪ France: Marc Besancon ▪ Italy: Paolo Giacomelli ▪ JINR: Igor Golutvin ▪ Other-member-states: Paris Sphicas ▪ Other-A: Brajesh Choudhury ▪ Other-B: A.Sanchez-Hernandez ▪ Switzerland: Ben Kilminster ▪ UK: Nick Wardle ▪ US: Greg Snow / Meenakshi Narain ▪ DO: Isabell Melzer-Pellmann ▪ YSC: Hannsjörg Weber + 
 Tamas Almos Vami ▪ Secretary: Quentin Ingram ▪ CB chair team

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

SP election process: proposed changes

The Reflection Group met over the summer and prepared a proposal. ▪ This should be discussed widely in the collaboration

• see mail to all CMS members last Friday
• a presentation in the CB meeting today

▪ CB decisions on each part are planned for December 2018 (earliest) Proposal: ▪ The collaboration should benefit from the participation of a broader constituency in the SP-election process. ▪ Areas where it is proposed to change our procedures are:

• 1. the nomination process
• 2. the interaction of the candidates with the collaboration
• 3. the election/voting procedure

3

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

Proposal: Change of SP election procedures

4

Bi Open CMS

Nomination Election current current current proposed

Ballot box CMS

• ther

PhD-M&O CBI Ballot box CMS

• ther

PhD-M&O CBI Nominations CMS

• ther

PhD-M&O CBI Nominations CMS

• ther

PhD-M&O CBI

proposed proposed Interaction

CMS candidates

• rganized

regional 
 meetings 3-4 weeks before election CMS candidates

• rganized

meeting at CERN 1 day before election

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

RG proposal 1: nomination

▪ Each CMS member, who has a PhD and is on the author list, should be allowed to nominate Spokesperson candidates. ▪ This is similar to the nomination procedure for L1 and L2 positions, in which the call for nominations is sent to the entire collaboration. ▪ Experience shows that this usually leads to well thought-out candidate lists. 


5

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

For the candidate list:

Current procedures include:

▪ Three months before the [SP] election: ▪ The Election Committee has the list of nominated candidates. These are then asked about their willingness to stand, and in the case of the election of the Spokesperson, CERN's agreement to the candidates must be secured. No change is proposed to this.

6

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

RG proposal 2: candidate’s campaigns

▪ Several open meetings with the SP candidates should be

• rganised 3-4 weeks before the election. Participation by all

regions should be made possible. ▪ Such organised meetings will help the candidates to inform the collaboration and increase transparency in this step of the election process. ▪ These meetings are necessary in order to have informed voters. ▪ The meetings should be organised and chaired by the election committee; the style of the meeting should be such that it remains unbiased. ▪ Candidates should present their statements, followed by time for questions and discussion. ▪ If possible, these meetings should be recorded and made available to those who are unable to attend. 


7

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

RG proposal 3: election procedure

▪ The proposal is guided by the following principles: ▪ The voting system should be simple and easy to understand. ▪ The special role of the CBIs is retained, who cast votes in parallel. ▪ Every CMS member for whom M&O-A is paid should be able to express his/her opinion. 
 ▪ The definition of voting rights for CBIs is defined in the CMS constitution, section 2.2. It should be applied in the same way as it is currently. ▪ CBI participation for institutes or clusters of institutes with 3 or more members.

8

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

RG proposal 3: election procedure

▪ The election is organized by a 
 web-based vote where all scientists 
 participate for whom M&O-A is paid. ▪ Team leaders (CBIs) vote N 
 times. ▪ We believe N=4 is a good value. ▪ Simulations suggest that the 
 value of N=4 would give the 
 popular and CBI components 
 approximately equal weight. ▪ This retains influence for the smaller institutions, while also enabling broader participation in parallel.
 Very small institutes (< 3 member) participate in the election. ▪ For large institutes the CBI vote does not play a large role, though the different opinions in the group can be represented.


9

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

RG proposal 3: election procedure

▪ The election is organized by a web-based vote where all scientists participate for whom M&O-A is paid. ▪ In our community satisfying experience exists using commercial systems: D0, CDF, in CMS, CERN developed a voting system in use at CERN and elsewhere. ▪ For the exact implementation details we propose to consider two

• ptions:
• 1. a single round of voting and the so-called preferential voting

system

• 2. two rounds of voting, where the top most voted candidates

enter into a run-off election.
 ▪ Option 2 is pretty much the current system. In a first round of voting the two candidates with the highest number of votes go forward to a second run-off vote. It is expected that a voting period of ~2 weeks is required.

10

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

Election options

Option 2: ▪ Election is done voting for the preferred candidate. ▪ The winner has to receive 
 > 50% of valid votes. ▪ If not achieved in round 1:

▪ 2 candidates with most votes participate in run-off election ▪ Winner is the candidate with 
 most votes.

11

Option 1: > Election organised as preferential (ranked) vote. > Voters places a candidate in first place, one in second place and so on, i.e. none, some or all candidates’ positions are assigned in a ranking

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

Option 1: Preferential (ranked) voting

Several ways exist to find the winner in an election using ranked voting It is proposed to use the “Instant Runoff” (IRV) procedure.

12

> Instant Runoff:

> The candidate with the fewest votes for first place is deleted from all ballots and the subordinate candidates move up. > The procedure is repeated from step 2 until only two candidates

• remain. The person with the

highest number of votes wins > This process is repeated until one candidate receives a majority of the ballots cast in that round. > The procedure can be terminated as soon as a candidate has more than half of the votes for the first place.

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

Work of the reflection group

▪ We have looked into several ways of defining the winner in a preferential vote. ▪ There is scientific literature about voting systems, based on mathematical and statistical analysis and simulations.

• It contains mathematical criteria by which voting methods

are compared. ▪ We looked at simulations ourselves. ▪ We looked at resistance to tactical voting.
 ▪ We believe that the ranked voting and the IRV procedure is robust for the CMS SP election. ▪ The advantage is: the result is derived in one round of voting. ▪ People may be hesitant to ‘downgrade’ colleagues by ranking them low.

13

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

14

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

15

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

16

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

Next steps proposed:

▪ The Reflection Group met over summer and prepared a proposal. ▪ It is presented to the collaboration for wide discussion and to collect feedback. ▪ Feedback, preferably in writing, is appreciated by the CB team and the RG members. ▪ The RG will meet again to consider feedback and suggestions received. ▪ Depending on the feedback received we foresees to put the proposal for decision in the CB meeting in December 2018 or February 2019.

17

CB Team - Matthias Kasemann | CMS CB 120 | 2nd of October 2018 | page

▪ Backup

18

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Preferential (ranked) voting - counting

Several ways exist to find the winner in an election using ranked voting Two very common ons are: “Instant Runoff” and “Borda counting”

19

> Instant Runoff:

> The candidate with the fewest votes for first place is deleted from all ballots and the subordinate candidates move up. > The procedure is repeated from step 2 until only two candidates

• remain. The person with the

highest number of votes wins > This process is repeated until one candidate receives a majority of the ballots cast in that round. > The procedure can be terminated as soon as a candidate has more than half of the votes for the first place.

> Borda counting

> for each ballot, a number of points corresponding to the number of candidates ranked lower. Example: 3 candidates

• 1st position: 3 points
• 2nd position: 2 points
• 3rd position: 1 point



 Other counting methods exist:
 2-1-0, 1-½-¼ > Once all votes have been counted the option or candidate with the most points is the winner.

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Preferential (ranked) voting - counting

Example D0, see document:

Borda counting tends to elect broadly-acceptable options or candidates, rather than those preferred by a majority. The Borda count is often described as a consensus-based voting system rather than a majoritarian one.

20

> Instant Runoff:

> Round 1: 
 C is eliminated for round 2


• votes for C are added

to A and B > Round 2: A wins 


• even though B had

more 1st votes


• C-voters prefer A over

B as 2nd choice

> Borda counting:


A: (50+60)*3+(78+43)*2+60+8=640
 B: (60+78)*3+(60+8)*2+50+43=643
 C: (43+8)*3+(50+60)*2+60+78=511


> B wins in the Borda score

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Preferential (ranked) voting - counting

Example 2:

21

> Borda counting:


> A wins in the Borda score

> Instant Runoff:

> Round 1: 
 B is eliminated for round 2


• votes for B are added

to A and C > Round 2: C wins 


• even though A had

more 1st votes


• B-voters prefer C over

A as 2nd choice

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Preferential (ranked) voting - counting

Example 2a: IRV is not fulfilling the Monotonicity criterion:

A voter can't harm a candidate's chances of winning by voting that candidate higher, or help a candidate by voting that candidate lower, while keeping the relative order of all the other candidates equal. A voter who places a candidate better on the ballot paper may not win the election while winning the election in a worse ranking.

• The order of the eliminations is decisive for the outcome of the choice.
• If it is possible to eliminate a candidate close to one's own favourite at an

early stage, one's own favourite can usually take over his or her votes.

22

> Borda counting:


> A wins in the Borda score

> Instant Runoff:
 Round 1: 


• C is eliminated for round 2

• votes for C are added to A and B


Round 2: A wins 


• 2 voters strategically decided to vote

B-A-C instead of A-B-C

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Comparison (from literature)

… “there is no "perfect" voting system”

23

Instant Runoff:

> improves the status quo in ways that are easy to see > balances the value of first choice support with the value of broad support > ranking additional choices never hurts your first choice 
 (this can be positive or negative) > favours polarised candidates with strong yet narrow support over consensus candidates with broader but weaker support.
 > Research concludes that IRV is one

• f the less-manipulable voting

methods

• it is vulnerable (see example)

• does not fulfil monotonicity

criteria

Borda counting

> elects broadly-acceptable options

• r candidates, rather than those

preferred by a majority > is often described as a consensus- based voting system > can elect candidates who might come in last in plurality votes, but be many voters' second or third choice. > could defeat a candidate who under today's rules would earn an absolute majority of 51 percent of the vote
 Many voters would react skeptically to such results > the Borda count is vulnerable to tactical voting

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Borda counting

24

Borda counting is vulnerable to tactical voting

> Compromising and burying:
 
 Compromising: voters can benefit by insincerely raising the position of their second choice candidate over their first choice candidate, in order to help the second choice candidate to beat a candidate they like even less. 
 
 Burying: voters can help a more-preferred candidate by insincerely lowering the position of a less-preferred candidate on their ballot. > An effective tactic is to combine these two strategies. 
 
 For example, if there are two candidates whom a voter considers to be the most likely to win, the voter can maximise his impact on the contest between these front runners by ranking the candidate whom he likes more in first place, and ranking the candidate whom he likes less in last place. 
 If neither front runner is his sincere first or last choice, the voter is employing both the compromising and burying tactics at once; 
 if many voters employ such strategies, then the result will no longer reflect the sincere preferences of the electorate.

CB Team - Matthias Kasemann | RG-10 | 9th of August 2018 | page

Graphical….

25