User:BetterVotingAdvocacy/Big page of ideas: Difference between revisions

A way to measure how much voters are taking advantage of the ability to indicate maximal support in multiple transitive matchups in rated pairwise relative to ranking and scoring is to, for each voter, add up the margin they express in each transitive matchup from 1st choice to last i.e. add up 1st>2nd, 2nd>3rd+...+2nd to last>last. If, for a given voter, this cumulative margin exceeds 1 vote/100% support/(max score - min score) points, then that voter's rated pairwise preference can't be compressed into a scale. If this cumulative margin is added up for all voters, and is also calculated for each voter's ranked preference (this can be found from the data they submit on a rated pairwise ballot), then this also shows how much voters weakened their vote relative to if they had been forced to rank the candidates in Condorcet. One thing that confounds this analysis is that some voters might indicate a weak rated pairwise or rated preference between two candidates, but rank the two equally rather than differentiate the two, to avoid putting too much power in the matchup. There may be a way to do similar analysis simply by looking at the number of votes candidates have overall in rated pairwise versus ranked (compare the bottom 2 tables in <ref>https://forum.electionscience.org/t/how-should-transitivity-be-handled-with-rated-pairwise-preferences/693/34</ref> for an example).

== References ==