Survey of Cumulative Voting in Port Chester Shows 34% of Voters Gave All Six Votes to One Candidate

On July 20, a study of voting in Port Chester, New York’s recent local election was released.  Port Chester used cumulative voting for Village Trustee.  Voters were each given six votes, and they were free to distribute their votes as they wished.  A voter could give one vote to each of six candidates, or all six votes to one candidate, or any variant in-between.  The study shows that 34% of the voters gave all six votes to one candidate.  See this story.

Port Chester used cumulative voting for its Village Trustee election because it had been sued under the Voting Rights Act to stop using ordinary at-large elections.  Even though Port Chester had a large Hispanic population, no Hispanic had ever before been elected to a Village Trustee slot.  One was elected in the recent election, however.  The election was also helpful to minor party and independent candidates.  One independent candidate was elected, placing first; and one Conservative Party nominee (who was not the nominee of any other party) was also elected.  Thanks to Gene Berkman for the link.


Comments

Survey of Cumulative Voting in Port Chester Shows 34% of Voters Gave All Six Votes to One Candidate — 6 Comments

  1. At least cumulative voting is additive and even the layperson can tally it.

    But did CV make a difference or did the voter education at a cost of $100.00 per registered voter impact the outcome?

    Vote system that elected NY Hispanic could expand
    “We put so much emphasis on education – we may have spent $100 a voter – because we knew it would be critical to success,” said village spokesman Aldo Vitagliano
    http://www.washingtonpost.com/wp-dyn/content/article/2010/06/18/AR2010061803766.html

    Consider that there usually is very little voter education for these elections, maybe a few postings in local papers (that few read) and maybe a mention on the radio.

    $100 per voter is alot of education.

  2. I do think it’s a good eye to spot voter education as a possible confounding variable here. But I don’t think that’s what led towards the outcome. Let’s take a look.

    The article shows that six candidates ran. Cumulative voting has an estimated threshold here of (1/ (6+1))= 14.3%. The majority (74%) in Port Chester are white while 15% are Hispanic/Latino (1). The Hispanic vote just gets past the threshold. If voters voted anywhere near along racial lines prior prior, Hispanics wouldn’t have had a shot. That’s because multi-member plurality allows the majority to take all the seats.

    A vulnerability to cumulative voting is vote splitting. One has to be careful to run just the right number of candidates. But that wasn’t an issue in this race. There were three Hispanics running. The winner, Marino, was a Democrat. One was a write-in and the other was Republican (2). This may have provided a successful guard against vote splitting. Marino came in fourth out of six suggesting that voting wasn’t narrowly along racial lines, or that vote splitting occurred among the other group(s). I do not like that cumulative voting is susceptible to vote splitting.

    But I like that cumulative voting is simple. It’s summable and doesn’t need central tabulation. It’s also more expressive than limited voting (vote for less than the number of seats, but no stacking). It’s much, much better than single-member and multi-member plurality . . . awful, awful systems. Multi-member plurality is–to my knowledge–the worst system in existence regarding proportionality.

    But if voters are willing to accept some complexity, then there’s STV, reweighted approval voting, and reweighted range voting. Beyond looking at individual criteria, it’s hard to fully assess multi-seat systems just yet (unlike the currently available Bayesian Regret (3) for single-member systems). STV’s downfalls are not nearly as bad as its IRV cousin. Though, it’s still not monotonic. On the other hand, reweighted range (and I believe reweighted approval) is monotonic. Monotonic means you don’t hurt a candidate by rating/ranking that candidate higher.

    Obviously, party list PR is inappropriate here due to the small number of seats and its poor handling of independents.

    (1-Census Bureau) http://tiny.cc/qd81e
    (2-NY Times) http://tiny.cc/sf2rr
    (3-Rangevoting.org) http://tiny.cc/hmhsg

  3. Cumulative Voting is one more JUNK so-called *reform*.

    ONE vote per Voter. Cost – Near ZERO.

    Total Votes / Total Seats = EQUAL votes needed for each seat winner — via pre-election CANDIDATE rank order lists — to transfer surplus and loser votes.

    Way too difficult for armies of MORON lawyers and especially judges to understand.

  4. Here is the report from the study http://www.scribd.com/doc/34680247/PC-Survey-Report-July20

    19.3% of those who were contacted for the survey were Hispanic. An attempt was made by surveyors to contact each voter as they exited, and surveyors made a guess as to the ethnicity, age, and sex of each voter who refused the survey. About 10% of voters evaded contact. Hispanics were more likely among those who contact was made to participate in the survey, so even the 19.3% might be a bit high.

    The average number of votes cast was 5.81/voter. This is consistent with 96.2% of voters casting 6 votes and 3.8% of voters casting one vote, which is consistent with the number of voters who reported using all 6 votes. If there were a significant number of voters, using most, but not all votes, then there would be fewer voters using all votes. For example, if all voters had used 5 or 6 votes, then 19% would have used 5 and only 81% used 6, which is not consistent with the survey results.

    The survey did not delve into who voters had voted for, and the organization of party slates likely played a significant factor in whether voters plumped or not.

    The Republicans ran 6 candidates, and they very much ran as a team, encouraging a vote for each member. They could have been shut out if voters had reliably followed their advice. The 6 candidates received 1.31, 1.24, 0.97, 0.93, 0.84, and 0.71 of the average party votes.

    The Democrats ran 4 candidates and there wasn’t any evidence at the candidate forum that they were supporting one another. They received 1.58, 1.22, 0.79, and 0.41 of the average number of votes. Even if we assume Democrat voters cast 6 votes for Democrats, the last-place candidate would only have received a vote from 0.61 of the voters.

    The winning independent, Bart Didden, definitely encouraged plumping. He was a plaintiff in a lawsuit against the Village of Port Chester which was ultimately decided by an opinion written by then-Appeals Court Judge Sonia Sotomayor.

    The Conservative candidate was a former mayor (several decades ago), so it is possible that he may have picked up individual votes from long-term residents.

  5. “One person, one vote” means nothing here in this country. Look at the U.S. Senate. Wyoming has about 600,000 people and two Senators while New York has 30 times that number and has, you guessed it, 2 senators. Please don’t give me that “one person, one vote” junk anymore.
    The Port Chester election brought about a result that “looks” like Port Chester- mixed ethnically and politically and leaning to the right (even the African-American who was elected is a Republican!). Nothing wrong with that, I’m just saying.

  6. # 5 — Sorry — LOTS of EVIL timebomb ANTI-Democracy ROT in the U.S.A. and State regimes –

    1. unequal ballot access laws

    2. minority rule gerrymanders (including the U.S.A. Senate and Electoral College).

    3. Robotic party hack executive and judicial officers.

    P.R. and App.V.

Leave a Reply

Your email address will not be published. Required fields are marked *