Talk:Combined approval voting

To do
— Omegatron (talk) 02:48, 1 July 2018 (UTC)
 * 1) Not sure if Ossipoff's endorsement is notable enough to mention in the text, or other people coming up with the same idea, like https://freethoughtblogs.com/singham/2014/11/19/a-very-modest-suggestion-to-improve-elections/ Nope, that's "balanced FPTP", not CAV.
 * 2) * Poundstone's book says "A three-valued system called "evaluative voting" has been proposed by D. S. Felsenthal, Claude Hillinger, and Mike Ossipoff."
 * 3) Do any of the researchers mention the "Unknown lunatic" averaging problem?
 * 4) * No because this isn't an averaged voting system, so it's not relevant.
 * 5) read about actual voter studies: "Nevertheless there is experimental evidence that people do use abstention if they are offered that option, e.g., the framed field experiment during the French presidential elections by Baujard et al. (2012)."
 * 6) * post analysis of Wikimedia elections on Wikimedia site and then reference it after others view it, so it's not as original researchy?
 * 7) ** No, Wikimedia's elections are averaged, so not the same system.
 * 8) ** Candidate A with 1 support, 99 neutral. Candidate B with 51 support, 49 oppose.  Wikimedia S/(S+O) averaged scoring selects unknown Candidate A (with 100% approval rating) over well-known but mediocre Candidate B (with 51% approval).   CAV's S-O summed scoring chooses Candidate B with 51-49 = 2 points vs Candidate A with 1 point.
 * 9) "Lepelley and Smaoui (2012) study some of its properties."
 * 10) Is Doge an example?  UN secretary general?  or are those aggregated differently?
 * 11) papers cite yahoo answers voting and eBay user ratings as examples. double check
 * 12) * Alcantud says "The thumb up or thumb down vote in Community Question Answering sites such as Yahoo! Answers illustrates that it is easy to ask users to cast a positive or negative vote (along with the possibility of abstaining)."
 * 13) * Poundstone says "eBay lets online buyers and sellers rate one another after each transaction. The three allowed choices are called positive, negative, and neutral. (This is not quite evaluative voting, as eBay ignores the neutral votes in computing the ratings. It's really approval voting with an option to abstain.)"

Definitions
Here I'm collecting evidence to confirm that all the various re-inventions are indeed referring to the same system (since there are other systems with the same {-1, 0, +1} ballot, such as Wikimedia's Board elections, which is instead tallied according to Explicit Approval Voting rules):


 * Felsenthal 1989 "In this paper I would like to investigate an extension of RAV, i.e., a voting method which allows each voter to combine approval with disapproval voting (CAV). Given n voters and the need to elect one out of k candidates (n, k 2 3), each voter under CAV has k votes and can, with respect to each candidate, either cast one vote in favor of this candidate, or cast one vote against this candidate, or abstain from voting for this candidate. The outcome of a CAV ballot is the candidate with the largest net vote total (algebraic sum of votes in favor and votes against), which may, of course, be negative."
 * Alcantud 2016 "In this paper we propose a characterization of a rule (that we refer to as dis&approval voting) that allows for a third level in the evaluation scale. The three levels have the following interpretation: 1 means approval, 0 means indifference, abstention or ‘do not know’, and -1 means disapproval." "We investigate the ‘dis&approval rule’, that selects the candidates who obtain the largest difference between the number of positive votes and the number of negative votes."
 * Hillinger 2004 "The theory leads me to advocate what I term evaluative voting. It can also be called utilitarian voting as it is based on having voters express their cardinal preferences. The alternative that maximizes the sum wins." "I define the evaluative voting (EV) rule as that which allows voters to express their cardinal preferences without restriction on a uniform, standardized scale. I propose that in a general election the scale should be EV-3 with the scale (-1 (against), 0 (neutral), +1 (for)). In a committee of experts a more differentiated rule, EV-5, with the scale (-2,-1,0,+1,+2) may be appropriate."
 * "In net approval voting, you vote up, down, or neutral on each candidate. The candidate with the most approvals minus specific disapprovals wins." https://news.ycombinator.com/item?id=17207095
 * Though in academia, "net approval" is a metric for use of approval voting in multi-winner elections? I don't see the term used for this.
 * "We have argued in the past that balanced approval voting is a good system and perhaps even the best system of voting. In that system, each voter has the option of voting approval (+1) or disapproval (-1) for each of the candidates. But necessarily also there is the option of marking neither approval nor disapproval (0). The sum is computed for each candidate and the winner is the candidate with the largest net vote. In this system the candidate with the largest net approval wins."
 * "After the polls close, for BAV the votes For and Against each candidate are tallied and a net vote for each candidate is computed as the difference. These net vote counts tells us the net satisfaction with electing that candidate. A positive net vote means that number more voters would feel pleased than would feel displeased with the election of that candidate.  Electing the candidate with the largest net vote would mean that the largest possible number of voters would be pleased rather than displeased with the outcome."
 * "We instead recommend Approval With Absention Option (AWAO), which is very easy to understand and implement, solves the Polarized-Environment Problem as well the Vote-Splitting Problem, and has numerous other benefits to be described in the full rollout. Voters approve any candidates they like, disapprove any they dislike, and may abstain wherever they may be neutral.  Disapprovals are subtracted from approvals for each candidate, and candidate with highest margin of net-approval wins."
 * "TWV (True Weight Voting) – This method simply asks Voters for their opinions (satisfaction or dissatisfaction) for each Candidate, ... the Candidate having the highest positive (or least negative) total is the winner. ... Here is how that would work using TWV1. Voters may mark any number of candidates they like as +1. They also may mark any number of candidates they don’t like as -1."
 * "TWV1 allows voters only three score values: -1, 0, and +1."

— Omegatron (talk) 15:46, 11 April 2020 (UTC)

Possible variants
What about a system with options for "approve" and "disapprove" that are interpreted as follows?
 * none => don't know => number of votes stays the same, score stays the same
 * approve => approve => number of votes += 1, score += 1
 * disapprove => disapprove => number of votes += 1, score -= 1
 * both => neutral => number of votes += 1, score stays the same

Then apply a Wilson score confidence interval (https://www.evanmiller.org/how-not-to-sort-by-average-rating.html) in favor of better-known candidates. Solomon Ucko (talk) 13:19, 9 September 2022 (UTC)


 * Haven't seen this suggestion before, but a common suggestion is to add a certain number of 0s to every candidate, which regularizes the scores of mostly-unrated candidates towards 0 (while not having much effect on any candidates who are rated by almost everyone). –Sincerely, A Lime 21:06, 15 May 2024 (UTC)