Never was heard, a discouraging word; No negativity allowed!
(Image by ip.ographer from flickr) Details DMCA
In his critique of balanced approval voting (BAV), Clay Schontrup (a founding member of Election Science) places special importance on what he calls the Favorite Betrayal Criterion (FBC):
A voting method satisfies the FBC when a voter never benefits from giving a lower score to their honest favorite. Approval voting satisfies this criterion. So does score voting. IRV does not, and neither does plurality. [IRV is the voting system that is now often called "ranked-choice voting"]
Both (unbalanced) Approval Voting (UAV) and Balanced Approval Voting (BAV) are score voting systems and so, as Schontrup points out, both satisfy FBC. Consequently, FBC does nothing to distinguish UAV from BAV.
There are infinitely many different score voting systems and for these systems there is a conventional procedure for constructing a ballot for use in an election. Each line of that ballot looks much like that pain scoring system you often see in your doctor's office, assigning a score from one to five. But when there are only two or three scores, there are better ballot designs than the score paradigm would suggest. This is the case for both UAV (with two scores) just as it is for BAV (with three scores).
And for these two systems, even the improved ballot designs are quite similar. In each case there is a list of candidates along side check boxes to allow voters to specify all but one of the scores; when no box is checked, that is interpreted as if checking the missing (neutral) score. It is least confusing, though not technically necessary, for the scores to be selected so that abstention is scored as zero.
And fortunately, it is always possible to make such a selection of scores. This is because adding (or subtracting) a constant from all scores will never alter the outcome of an election. So, for example, a score voting system using scores [1, 2, 3] is equivalent to the score system using the scores [-1, 0, 1]; this is because the second system is derived from the first by subtracting 2 from each of the three scores. Vote tallies will change but the election outcomes for the two systems will be the same.
Ballot design can be psychologically important, despite a poor ballot design being mathematically equivalent to another one. For example, with BAV, voters are asked whether they support or oppose each candidate rather than (with the score ballot) being asked to assign a numeric score to indicate to what degree they approve of a candidate. It would seem a mistake to assume voters would generally be aware the two are mathematically equivalent.
But UAV and BAV both satisfy FBC, so why does Schontrup bring up the topic of FBC? Apparently, it is as support for the notion that satisfying FBC enough; given that UAV meets the FBC criteria, there can be no justification for considering whether there may be an even better voting system.
In principle, elections are held to find out what voters want and that objective cannot be met if ballots fail to allow voters to express how they honestly feel. FBC seems to have been proposed as criterion that would discourage voters from voting strategically. And this is surely a worthy goal since strategic encourages voters to lie about their true opinions about the candidates.
Unfortunately, a voter can sometimes feel forced to rely on some strategy simply because the ballot fails to provide the voter any way to cast an entirely honest vote.
Schontrup promotes FBC as a vitally important concept, so surely it deserves our consideration. Initially, we should probably consider the simplest and most familiar voting system, plurality voting (PV) regarding FBC.
PV works best if we could assume that a voter's only concern is to help elect a single, most-favored candidate. And, given that assumption, it is surely impossible for a voter to find any better alternative than to simply vote for that one favorite candidate. With that assumption, we could conclude that PV satisfies FBC.
But then why does Schontrup assert that PV fails to meet the FBC conditions? It must be that voters sometimes have concerns other than electing their single top choice.
For example, some voters may not prefer one single candidate above all the others. It is surely possible for a voter to feel that two or more candidates are equally suited for being the best candidate.
In such a case, voters feel forced to choose just one of the most favored candidates to support, and that is a textbook example of vote splitting, something well understood to lead to anomalous elections.
The problem is that support for each of several candidates is divided, possibly even equally between two or more equally qualified candidates. This leaves only half of the split votes for each one. A candidate with support from a minority of voters may then win even though a majority of voters would have preferred any of the other candidates. Such an anomalous election for Governor of Maine led that state to adopt ranked-choice voting (then called instant runoff voting, IRV).
(Note: You can view every article as one long page if you sign up as an Advocate Member, or higher).
