[EM] Most important/used election methods?

[Kristofer Munsterhjelm]

On 15.10.2021 14:24, Bohdan Andriyiv wrote:
> Richard, thank you for the detailed list of (as I understood) the best
> / the most fair / the most supported by this forum election methods.
> It is very helpful and works for me as a good starting point to
> explore the election methods. Thanks to the others who made their
> inputs as well.
> 
> But my question is different. I guess it worked as a trick question
> for this community and I should have explained better the context
> behind it.
> 
> I am asking what election methods are the most important/popular in
> the current world? Not, which ones are the best, the fairest. BTW, by
> "important" I mean - the methods that might be not so popular, but are
> important in the current world for some other reason (for example,
> used in small, but important elections).

Plurality (FPTP), clearly, by sheer numbers, although the method is
nothing to write home about.

Probably also, if you're willing to bend the criteria of what counts as
voting: Approval with no opinion (thumbs up/down on FB, YouTube, etc.),
and Range (star ratings used on e.g. Amazon). I say "bend" because these
are not elections - they do not designate a given winner, but rather
they provide feedback on a per item (video, product, etc) basis.

The most popular multiwinner election methods are (to my knowledge)
party list, single member district Plurality, SNTV, and STV. They can be
combined as in MMP. Perhaps you could also include cumulative voting and
bloc voting, but they're not as common.

The Schulze method seems to be favored among open-source organizations
and pro open-source parties such as the Pirate Parties. See e.g.
https://en.wikipedia.org/wiki/Schulze_method#Users. It is also used in a
few cities for referenda or as part of a liquid democracy/delegable
proxy system.

The Condorcet Internet Voting Service supports minmax, Schulze, MAM, and
Condorcet-IRV: https://civs1.civs.us/rp.html

Approval (with or without runoff) is also in current use in the US. See
https://en.wikipedia.org/wiki/Approval_voting#Current_use.

FairVote is pushing IRV pretty hard, and it's used in some US states
(see
https://en.wikipedia.org/wiki/Ranked-choice_voting_in_the_United_States
) but it's not a particularly good method IMHO.

Finally, there's the meta-method "whatever election method, then the
president leans on the electoral commission and stays in power
regardless". This method is unfortunately all too common; I would not
suggest implementing it.

> In other words my question is the following.
> 
> What types of election methods are most likely to be used to organize
> different votings on the universal voting platform in the current
> world?
> Also, what methods, although not so popular, but important in the
> current world, should be supported by this platform?

I would not suggest implementing Plurality even though it's popular,
because it's just that bad. I'd probably say: Approval, Range, STAR, and
a Condorcet method (Schulze if you're going by popularity, though I'd
personally prefer Ranked Pairs for relative simplicity). This gives you
both the ranked standard (according to many people on this list at least
:), the cardinal standards (Approval and Range), and a compromise
between the two (STAR).

The only thing really missing is Majority Judgment; it's not used
anywhere but it has an interface unlike either Range or Condorcet, using
neither scores nor rankings but grades instead.

If you're going to implement proportional representation multiwinner
methods, then some form of STV. Probably also some kind of cardinal PR
method (although none are used for political elections at the moment) to
give the cardinal voters a multiwinner method too.

There's a similar argument to the "include IRV or not?" about what kind
of STV to use. Every form of STV used politically at the moment reduces
to IRV in the single-winner case, which means that its results get worse
the fewer winners you have. On the one hand, if you want to stick with
what's used in political elections, you should pick IRV-based STV. On
the other, if you want better outcomes, it would be better to pick a
Condorcet-based multiwinner method (e.g. Schulze STV, CPO-STV or CIVS's
PR method).

-km