| Report |
|
________________________________________________________________________
MARCH 02, 2015
IntroductionThe following results are from a survey conducted between December 30 2014 – January 04, 2015. Respondents were asked to rank their priorities (i.e., values) for a new voting system. It was helpful to know the general priorities of The South Pacific to help ascertain which voting systems would meet our expectations better than others.
Results
The results suggest that The South Pacific is less concerned with the outcome of results of elections than they are concerned with the legitimacy of the democratic process. This means that having elections with more diverse candidates, or a more civil environment, is a secondary concern next to whether or not the electoral system's process is fair, open, and representative.
This is an important observation because one of the leading reasons why some have suggested we should maintain First-Past-The-Post or pursue Instant-Runoff Voting is because these systems deliberately limit voter choice and supposedly result in more "extreme candidates" winning through falsely divisive scenarios. The results of the survey would suggest that such a system would not meet the expectations of The South Pacific because the survey participants preferred a system that was fairer and did not limit voters and their ability to represent and express their views, as opposed to one that “forked” decisions and forced voters to support one candidate over another on a merely strategic basis. With regards to “majority rule”, the results were particularly interesting – majority rule was a divisive concept (reaching strong support and strong opposition) while consensus had a broader base of support. One possibility is that “majority rule” was interpreted differently by participants – while majority rule simply means that leaders must need the support of a majority to rule (a criterion that the First-Past-The-Post regularly fails to meet), some may have thought of majority rule in a more colloquial sense where minority views are disregarded in the face of a simple majority – The South Pacific, after all, broadly appears to value the idea of consensus and finding common ground between minorities. Comparing Voting Systems...
Throughout this report, the recent judicial appointment of Apad shall be used to compare different proposed voting systems. Some of the examples below are how said vote could have concluded using an array of different voting systems...
Assuming that the first preference of voters would be their choice in a First-Past-The-Post election, voters would have selected Sopo as the judicial appointment (see above - left). In all of the other voting methods that the commission tested, however, the election would have resulted in Apad winning the judicial appointment. For example, (as depicted above - center) under Instant-Runoff Voting (IRV), Henn would be eliminated in Round 1 and Henn's supporters would back Apad in Round 2. Meanwhile, Apad would also win under a traditional Borda Count (see above - right). Borda Count is a simple method where voters award candidates with "points" based on the preferences, like, for example if Henn was your favoured candidate, he'd be awarded two more points than your least favourite candidate (Sopo or Apad). The winner of a Borda Count is the candidate with the most points. Ultimately, the commission rejected First-Past-The-Post, Instant-Runoff Voting and Borda Count as proposals; however, the test comparisons have been provided here nonetheless for your perusal. The commission proposes three different voting systems, believing these voting systems best fit The South Pacific's preferences and values: Minimax, Condorcet and Approval voting. Proposal 1: Minimax
General Discussion When considering a system based on fairness, the committee was drawn to think of "Minimax" theory as a starting point for the discussion. When John Rawls, prominent political philosopher, was discussing what a "just society" might look like, Rawls proposed that justice was fairness -- this means to say that a just society was one in which everyone could agree the societal structure (e.g., social security, tolerance, freedom) was "fair". Minimax was a term which Rawls popularized: it means "minimizing the maximum" - in this case, a just society was one where the worst outcomes (e.g., abject poverty, authoritarianism) were avoided. In general, however, Minimax theory suggests that if you're given a set of options, you should choose the option that bests prevents you from receiving the worst of the possible outcomes. Minimax theory was later applied to voting theory; in terms of a voting system, a Minimax system means that the candidate who wins is the candidate who would minimize dissatisfaction. If for example, there are two candidates, Blue and Red, both are venomously opposed by either side of TSP, but Yellow, an alternative candidate, would be favoured by everyone in TSP over seeing their enemy win, then Minimax would support the compromise position (Yellow) over an alternative scenario where Red and Blue contest for supremacy and ultimately, a great number of voters are dissatisfied with the election. Minimax fits the region's preferences for an electoral system because Minimax's core values of fairness aligns with The South Pacific's core values -- the system is centrally concerned with, after all, pursuing a process of representation that is near universally acceptable and producing results which avoid wider dissatisfaction across the region. Technically speaking... The winner in a minimax election is the candidate whose worst result in a one-on-one election against every other candidate would be the best of the "worst results" for all other candidates. A Practical Example Using the recent judicial appointment election as an example - let's assume there are three candidates, Henn, Apad and Sopo and voters ordered their preferences for the appointment on their ballots from least preferred to most preferred. While it makes it a little more complex than it actually is, this graph shows you clearly how the votes breakdown and who would win under minimax theory... Essentially, we tabulate the votes. Then we assume that if, say eight voters prefer Apad than Henn and four voters prefer Henn over Apad then if an election were held solely between Apad and Henn, Apad would win. Henn and Apad aren't the only candidates though, so we have consider how things would work out between any of the candidates in potential one-on-one contests (i.e., Henn vs. Apad, Apad vs. Sopo, Sopo vs. Henn). Within these hypothetical scenarios, Henn's "biggest defeat" would be a loss of 66.66% (he'd lose to both candidates), while Sopo's "biggest defeat" would be 58.3% (a loss he sustained against Apad) - and Apad wouldn't lose against either candidate, so his biggest defeat would be 0%. In Minimax voting, the candidate with the smallest "biggest defeat" wins, so in this case, Apad would be declared the winning candidate - because he wouldn't have even been defeated. What would this system look like on the books? Section X - 1. Voters shall list candidates from their most to least preferred on their ballot. 2. The winning candidate shall be the candidate, when paired up against any other candidate, whose worst defeat (in terms of votes) would be the best of any other candidate. 3. For the purposes of reference, this voting system shall be referred to as "Minimax Condorcet". Proposal 2: Condorcet
General Discussion There are a lot of similarities between Minimax and Condorcet - they share the same values of fairness and consensus - however, Condorcet may be intuitively easier to understand for readers and may be more familiar too, because Condorcet is used regularly to decide contradicting votes. While Condorcet is regularly referred to as the "complicated" system - the Commission all agreed that Condorcet is far easier to calculate that alternatives like Instant-Runoff Vote (which requires votes to be divvied up and distributed with each round). Essentially, Condorcet is concerned with the Condorcet winner -- the Condorcet winner is like the winner of a sports tournament. In a single elimination tournament, the winner of a sports tournament is the team which defeats all of the other teams it faces against. Condorcet goes a little bit further than that, though. Not only does the Condorcet winner (the winning candidate) need to be the undefeated candidate, it needs to be the candidate who would defeat all of the other candidates in a one-on-one match. These contests all occur with one single vote, so we don't need to do a dozen little mini-elections to decide who is the supreme candidate. Technically speaking... The winner in a Condorcet election is the candidate who would defeat every other candidate in a one-on-one election against every other candidate. If no such candidate exists, the election is decided by Instant-Runoff (candidates are eliminated until one candidate attains majority support). A Practical Example Using the recent judicial appointment election as an example - let's assume there are three candidates, Henn, Apad and Sopo and voters ordered their preferences for the appointment on their ballots from least preferred to most preferred. In this case, the judicial appointment election actually was decided using Condorcet, so we don't have to hypothesize the results. While it makes it a little more complex than it actually is, this graph shows you clearly how the votes breakdown and who would win under Condorcet... Condorcet works very similar to Minimax. Essentially, we tabulate the votes. Then we assume that if, say eight voters prefer Apad than Henn and four voters prefer Henn over Apad then if an election were held solely between Apad and Henn, Apad would win. Henn and Apad aren't the only candidates though, so we have consider how things would work out between any of the candidates in potential one-on-one contests (i.e., Henn vs. Apad, Apad vs. Sopo, Sopo vs. Henn). Since Apad wouldn't lose against either candidate, Apad would be declared the winning candidate - because he wouldn't have even been defeated. What would this system look like on the books? Section X - 1. Voters shall list candidates from their most to least preferred on their ballot. 2. The winning candidate shall be the candidate, when paired up against any other candidate, who would surpass any other candidate in terms of support; if no such candidate exists, the winning candidate shall be determined by an instant-runoff vote. 3. For the purposes of reference, this voting system shall be referred to as "Condorcet-IRV". Proposal 3: Approval
General Discussion While Condorcet and Minimax are in fact fairly easy to use and produce fair and reliable results, Approval voting is probably the most easiest to conceptualize and understand for most readers. Approval voting works like endorsements in NationStates - you can endorse as many WA members as you like and the chosen delegate is based on who has the most endorsements. Approval voting works just like that! You state on your ballot which candidates you approve of, and the candidate with the most approvals is elected. Approval voting does not always produce a majority rule - in the sense that - Condorcet and Minimax will always elect candidates with majority support, whereas Approval voting elects someone with the most approvals (a plurality), but not necessarily the majority's candidate. However, approval voting is also able to respond to multiple candidates entering a race better than Condorcet or Minimax, because regardless of whether "similar" candidates or "big bad undesirable" candidates enter the race, it will not re-position your support for other candidates - you can still approve of X, Y, and Z, and you have no reason to not approve of any of those candidates, just because A, B or C has entered the race. Since majority rule was not as valued as other criteria, Approval voting might actually reflect The South Pacific's preference for fair, participatory systems that encourage people to run as opposed to other systems (like First-Past-The-Post) where someone may be discouraged to run because their candidacy might otherwise enable their opponent to win (by splitting their supporters's votes with another candidate who is running in the same election). Technically speaking... The winner in an Approval Vote is the candidate which receives the most approvals from voters. A Practical Example Using the recent judicial appointment election as an example - let's assume there are three candidates, Henn, Apad and Sopo and voters wrote down which candidates out of that list that they approved of; for example, they might approve of one, two or all three of the candidates. Since voters originally did not state which candidates they would approve of, in particular, the Commission set out to find which candidate would have won the election if it had been conducted via Approval voting which required the Commission to contact voters in the judicial appointment vote over the course of February to ask them which candidates they would have approved of, if given such an option. The Commission assumed abstentions in the judicial appointment vote would have remained abstentions, for the purposes of the experiment. We thank respondents for their cooperation and enthusiasm. With their participation we were able to hypothesise based on the results that they provided that, well... the Approval vote would be a damn close race - in fact, in our findings, Sopo and Apad tied. The reason why Approval voting might establish a tie here is it "smooths" out voters preferences and forces them to either approve or disapprove (a binary preference) -- what we found was there wasn't a strong disapproval of any of the candidates, so many voters simply approved the same candidates which, of course, lead to a tie. It's possible this could be a common phenomenon in The South Pacific, so it would be important for any implementation of Approval voting to have an established tiebreaker method (like a runoff, or you count the number of "disapprovals") in case of events like these. What would this system look like on the books? Section X - 1. Voters shall note on their ballot which candidates they approve of; they may approve of multiple candidates. 2. The winning candidate shall be the candidate who garners the most approval votes from voters; in the event of a tie... 3. For the purposes of reference, this voting system shall be referred to as "Approval Voting".
This is what I have so far...
Quote:The results suggest that The South Pacific is less concerned with the “quality” of results of elections than they are concerned with the legitimacy of the democratic process. This means to say that whether elections result in more diverse candidates or a more civil political environment is a secondary concern for the survey’s participants – instead, they’re more concerned with how their elections work and therefore, concerned with whether their electoral system is fair, open and representative – something agreeable and legitimate to everyone.
Proposal 1: Minimax
General Discussion When considering a system based on fairness, the committee was drawn to think of "Minimax" theory as a starting point for the discussion. When John Rawls, prominent political philosopher, was discussing what a "just society" might look like, Rawls proposed that justice was fairness -- this means to say that a just society was one in which everyone could agree the societal structure (e.g., social security, tolerance, freedom) was "fair". Minimax was a term which Rawls popularized: it means "minimizing the maximum" - in this case, a just society was one where the worst outcomes (e.g., abject poverty, authoritarianism) were avoided. In general, however, Minimax theory suggests that if you're given a set of options, you should choose the option that bests prevents you from receiving the worst of the possible outcomes. Minimax theory was later applied to voting theory; in terms of a voting system, a Minimax system means that the candidate who wins is the candidate who would minimize dissatisfaction. If for example, there are two candidates, Blue and Red, both are venomously opposed by either side of TSP, but Yellow, an alternative candidate, would be favoured by everyone in TSP over seeing their enemy win, then Minimax would support the compromise position (Yellow) over an alternative scenario where Red and Blue contest for supremacy and ultimately, a great number of voters are dissatisfied with the election. Minimax fits the region's preferences for an electoral system because Minimax's core values of fairness aligns with The South Pacific's core values -- the system is centrally concerned with, after all, pursuing a process of representation that is near universally acceptable and producing results which avoid wider dissatisfaction across the region. Technically speaking... The winner in a minimax election is the candidate whose worst result in a one-on-one election against every other candidate would be the best of the "worst results" for all other candidates. A Practical Example Using the recent judicial appointment election as an example - let's assume there are three candidates, Henn, Apad and Sopo and voters ordered their preferences for the appointment on their ballots from least preferred to most preferred. While it makes it a little more complex than it actually is, this graph shows you clearly how the votes breakdown and who would win under minimax theory... Essentially, we tabulate the votes. Then we assume that if, say eight voters prefer Apad than Henn and four voters prefer Henn over Apad then if an election were held solely between Apad and Henn, Apad would win. Henn and Apad aren't the only candidates though, so we have consider how things would work out between any of the candidates in potential one-on-one contests (i.e., Henn vs. Apad, Apad vs. Sopo, Sopo vs. Henn). Within these hypothetical scenarios, Henn's "biggest defeat" would be a loss of 66.66% (he'd lose to both candidates), while Sopo's "biggest defeat" would be 58.3% (a loss he sustained against Apad) - and Apad wouldn't lose against either candidate, so his biggest defeat would be 0%. In Minimax voting, the candidate with the smallest "biggest defeat" wins, so in this case, Apad would be declared the winning candidate - because he wouldn't have even been defeated. What would this system look like on the books? Section X - 1. Voters shall list candidates from their most to least preferred on their ballot. 2. The winning candidate shall be the candidate, when paired up against any other candidate, whose worst defeat (in terms of votes) would be the best of any other candidate. 3. For the purposes of reference, this voting system shall be referred to as "Minimax Condorcet".
Comparing Voting Systems...
Throughout this report, the recent judicial appointment of Apad shall be used to compare different proposed voting systems. Some of the examples below are how said vote could have concluded using an array of different voting systems...
Assuming that the first preference of voters would be their choice in a First-Past-The-Post election, voters would have selected Sopo as the judicial appointment (see above - left). In all of the other voting methods that the commission tested, however, the election would have resulted in Apad winning the judicial appointment. For example, (as depicted above - center) under Instant-Runoff Voting (IRV), Henn would be eliminated in Round 1 and Henn's supporters would back Apad in Round 2. Meanwhile, Apad would also win under a traditional Borda Count (see above - right). Borda Count is a simple method where voters award candidates with "points" based on the preferences, like, for example if Henn was your favoured candidate, he'd be awarded two more points than your least favourite candidate (Sopo or Apad). The winner of a Borda Count is the candidate with the most points. Ultimately, the commission rejected First-Past-The-Post, Instant-Runoff Voting and Borda Count as proposals; however, the test comparisons have been provided here nonetheless for your perusal. The commission proposes three different voting systems, believing these voting systems best fit The South Pacific's preferences and values: Minimax, Condorcet and Approval voting.
Proposal 2: Condorcet
General Discussion There are a lot of similarities between Minimax and Condorcet - they share the same values of fairness and consensus - however, Condorcet may be intuitively easier to understand for readers and may be more familiar too, because Condorcet is used regularly to decide contradicting votes. While Condorcet is regularly referred to as the "complicated" system - the Commission all agreed that Condorcet is far easier to calculate that alternatives like Instant-Runoff Vote (which requires votes to be divvied up and distributed with each round). Essentially, Condorcet is concerned with the Condorcet winner -- the Condorcet winner is like the winner of a sports tournament. In a single elimination tournament; the winner of a sports tournament is the team which defeated all of the other teams it faced against. Condorcet goes a little bit further than that, though; not only does the Condorcet winner (the winning candidate) need to be the undefeated candidate - it needs to be the candidate who would defeat all of the other candidates in a one-on-one match - and these contests all occur with one single vote, so we don't need to do a dozen little mini-elections to decide who is the supreme candidate. Technically speaking... The winner in a Condorcet election is the candidate who would defeat every other candidate in a one-on-one election against every other candidate. If no such candidate exists, the election is decided by Instant-Runoff (candidates are eliminated until one candidate attains majority support). A Practical Example Using the recent judicial appointment election as an example - let's assume there are three candidates, Henn, Apad and Sopo and voters ordered their preferences for the appointment on their ballots from least preferred to most preferred. In this case, the judicial appointment election actually was decided using Condorcet, so we don't have to hypothesize the results. While it makes it a little more complex than it actually is, this graph shows you clearly how the votes breakdown and who would win under Condorcet... Condorcet works very similar to Minimax. Essentially, we tabulate the votes. Then we assume that if, say eight voters prefer Apad than Henn and four voters prefer Henn over Apad then if an election were held solely between Apad and Henn, Apad would win. Henn and Apad aren't the only candidates though, so we have consider how things would work out between any of the candidates in potential one-on-one contests (i.e., Henn vs. Apad, Apad vs. Sopo, Sopo vs. Henn). Since Apad wouldn't lose against either candidate, Apad would be declared the winning candidate - because he wouldn't have even been defeated. What would this system look like on the books? Section X - 1. Voters shall list candidates from their most to least preferred on their ballot. 2. The winning candidate shall be the candidate, when paired up against any other candidate, who would surpass any other candidate in terms of support; if no such candidate exists, the winning candidate shall be determined by an instant-runoff vote. 3. For the purposes of reference, this voting system shall be referred to as "Condorcet-IRV".
Proposal 3: Approval
General Discussion While Condorcet and Minimax are in fact fairly easy to use and produce fair and reliable results, Approval voting is probably the most easiest to conceptualize and understand for most readers. Approval voting works like endorsements in NationStates - you can endorse as many WA members as you like and the chosen delegate is based on who has the most endorsements. Approval voting works just like that! You state on your ballot which candidates you approve of, and the candidate with the most approvals is elected. Approval voting does not always produce a majority rule - in the sense that - Condorcet and Minimax will always elect candidates with majority support, whereas Approval voting elects someone with the most approvals (a plurality), but not necessarily the majority's candidate. However, approval voting is also able to respond to multiple candidates entering a race better than Condorcet or Minimax, because regardless of whether "similar" candidates or "big bad undesirable" candidates enter the race, it will not re-position your support for other candidates - you can still approve of X, Y, and Z, and you have no reason to not approve of any of those candidates, just because A, B or C has entered the race. Since majority rule was not as valued as other criteria, Approval voting might actually reflect The South Pacific's preference for fair, participatory systems that encourage people to run as opposed to other systems (like First-Past-The-Post) where someone may be discouraged to run because their candidacy might otherwise enable their opponent to win (by splitting their supporters's votes with another candidate who is running in the same election). Technically speaking... The winner in an Approval Vote is the candidate which receives the most approvals from voters. A Practical Example Using the recent judicial appointment election as an example - let's assume there are three candidates, Henn, Apad and Sopo and voters wrote down which candidates out of that list that they approved of; for example, they might approve of one, two or all three of the candidates. Since voters originally did not state which candidates they would approve of, in particular, the Commission set out to find which candidate would have won the election if it had been conducted via Approval voting which required the Commission to contact voters in the judicial appointment vote over the course of February to ask them which candidates they would have approved of, if given such an option. The Commission assumed abstentions in the judicial appointment vote would have remained abstentions, for the purposes of the experiment. [Image: UBNMzVf.png] We thank respondents for their cooperation and enthusiasm. With their participation we were able to hypothesise based on the results that they provided that, well... the Approval vote would be a damn close race - in fact, in our findings, Sopo and Apad tied. The reason why Approval voting might establish a tie here is it "smooths" out voters preferences and forces them to either approve or disapprove (a binary preference) -- what we found was there wasn't a strong disapproval of any of the candidates, so many voters simply approved the same candidates which, of course, lead to a tie. It's possible this could be a common phenomenon in The South Pacific, so it would be important for any implementation of Approval voting to have an established tiebreaker method (like a runoff, or you count the number of "disapprovals") in case of events like these. What would this system look like on the books? Section X - 1. Voters shall note on their ballot which candidates they approve of; they may approve of multiple candidates. 2. The winning candidate shall be the candidate who garners the most approval votes from voters. 3. For the purposes of reference, this voting system shall be referred to as "Approval Voting".
Just needs an introduction.
Intro added. I'm ready to post if you are, Glen.
Just a few revisions...
Quote:The results suggest that The South Pacific is less concerned with the “quality” outcome of results of elections than they are concerned with the legitimacy of the democratic process. This means that having elections with more diverse candidates, or a more civil environment, is a secondary concern next to whether or not the electoral system's process is fair, open, and representative. to say that whether elections result in more diverse candidates or a more civil political environment is a secondary concern for the survey’s participants – instead, they’re more concerned with how their elections work and therefore, concerned with whether their electoral system is fair, open and representative – something agreeable and legitimate to everyone. Quote:This is an important observation because one of the leading reasons why some have suggested we should maintain First-Past-The-Post or pursue Instant-Runoff Voting is because these systems deliberately limit voter choice and supposedly result in more "extreme candidates" winning through falsely divisive scenarios. Quote:Essentially, Condorcet is concerned with the Condorcet winner -- the Condorcet winner is like the winner of a sports tournament. In a single elimination tournament, the winner of a sports tournament is the team which defeats all of the other teams it faces against. Condorcet goes a little bit further than that, though. Not only does the Condorcet winner (the winning candidate) need to be the undefeated candidate, it needs to be the candidate who would defeat all of the other candidates in a one-on-one match. These contests all occur with one single vote, so we don't need to do a dozen little mini-elections to decide who is the supreme candidate. Quote:What would this system look like on the books? |
| Users browsing this thread: |
| 1 Guest(s) |