Participation criterion

From electowiki
Wikipedia has an article on:

The participation criterion is a voting system criterion applicable to both single and multiple winner ranked voting systems. A method that passes this criterion ensures a voter that it's always better to cast a full honest vote than to not show up for the election at all. It does this by guaranteeing that adding a ballot can never change the winner from someone who is ranked higher on that ballot to someone who is ranked lower.

While the criterion ensures that a voter can't benefit from staying home rather than voting honestly, a voter may do even better by engaging in tactical voting; participation does not imply that the method is strategy-proof.

Definition

For deterministic single-winner methods, the criterion is defined as follows:

Adding one or more ballots that vote X over Y should never change the winner from X to Y.

For multi-winner methods and methods that involve an element of chance, the definition is:[1]

The addition of a further ballot should not, for any positive whole number k, reduce the probability that at least one candidate is elected out of the first k candidates listed on that ballot.

Variants

Semi-honest participation criterion

This is a weaker form of the participation criterion. It states that for any set of ballots, an extra voter with a given preference set must be able to cast a ballot which is semi-honest and meaningfully expressive, without making the result worse. Meaningfully expressive means that if the voter prefers some set of candidates to the winner, the non-harmful ballot must be able to express that preference.

Complying methods

This criterion is important in the context of the Balinski–Young theorem. Failing the participation criterion is an an example of failing population monotonicity.

Every weighted positional method that gives higher ranked candidates higher scores passes the participation criterion. In particular, Plurality voting and the Borda count both pass. Furthermore, Approval voting, Cardinal Ratings, and Woodall's DAC and DSC methods all pass the participation criterion. All Condorcet methods,[2][3] Bucklin voting,[4] and IRV[5] fail.

It's possible to pass both Condorcet and Participation for three candidates and any number of voters, or for four candidates up to 11 voters inclusive.[6] This result also holds for certain probabilistic extensions of the Condorcet criterion.

All Monroe type multi-member systems fail participation.

Notes

See Truncation for a Participation-like criterion for bullet voting.

Note that the Participation criterion doesn't say a voter should be able to benefit in some circumstances by voting, nor does it quantify such a thing. For example, a voting method which randomly chooses one of the candidates regardless of the votes would pass Participation, despite not giving voters any power. Voting methods like Score and FPTP can have this quantified because they are based on similar systems of increasing a candidate's "quality number", with each voter only being able to increase the number for a given candidate to a certain maximal amount.

See also

References

  1. Woodall, Douglas R. (1996). "Monotonicity and single-seat election rules". Voting matters. 6: 9–14.
  2. Moulin, Hervé (1988-06-01). "Condorcet's principle implies the no show paradox". Journal of Economic Theory. 45 (1): 53–64. doi:10.1016/0022-0531(88)90253-0.
  3. "Participation failure" is forced in Condorcet methods with at least 4 candidates". Retrieved 2014-12-24.
  4. Markus Schulze (1998-06-12). "Regretted Turnout. Insincere = ranking". Retrieved 2011-05-14.
  5. Warren D. Smith. "Lecture "Mathematics and Democracy"". Retrieved 2011-05-12.
  6. Brandt, Felix; Geist, Christian; Peters, Dominik (2016-02-25). "Optimal Bounds for the No-Show Paradox via SAT Solving". arXiv:1602.08063 [cs.GT].

Some parts of this article are derived with permission from text at http://electionmethods.org

External links

This page uses Creative Commons Licensed content from Wikipedia (view authors).