Many SF Voters Used Ranked-Choice Voting Incorrectly
Analysis finds voting patterns break down by neighborhood boundaries
Updated Dec. 2, 2011 at 12:09 p.m.
The results are in: San Francisco voters have trouble with ranked-choice elections.
Despite a $300,000 educational campaign leading up to last month’s elections, including a new smiley-face mascot, publicity events, and advertising on buses and in newspapers, only one-third of voters on Nov. 8 filled out all three choices in all three races, according to an analysis released this week by the University of San Francisco.
Under the city’s system, voters were asked to rank their top three choices for mayor, sheriff and district attorney.
Perhaps the analysis’ most troubling finding is that 9 percent of voters, mostly in Chinatown and southeastern neighborhoods like the Bayview, marked only one choice for each office, likely either because they considered only one candidate suitable or because they did not know how to fill out their ballot correctly.
“Some people just prefer to rank one,” said Corey Cook, a political science professor at the university who wrote the report with David Latterman. “But the geographic component suggests it’s more systematic.”
Cook and Latterman plan to do further research into the question of why some voters consistently chose only one candidate.
Although Ed Lee did not receive a majority of first-place votes, he became the city’s first elected Chinese-American mayor based on the ranked-choice system, which was first used in San Francisco in 2004.
Latterman, an associate director of the Leo T. McCarthy Center for Public Service and the Common Good at USF, said voters in neighborhoods with large black or Asian populations tended to vote for different candidates than residents in other parts of the city. But the Nov. 8 election was the first time researchers saw a geographic or perhaps ethnic difference in how people used ranked-choice voting.
The findings indicate one of two things, Latterman said: Either campaigns tried to manipulate the results by focusing on specific groups of people or there is not a clear understanding of how to use the system.
A recent Bay Citizen analysis revealed that 16 percent of ballots in the mayoral race — those of more than 31,500 people — were filled out correctly but were discarded when all of their chosen candidates were eliminated from the race. San Francisco does not allow voters to rank all the candidates on the ballot.
In June, a voting task force created by the Board of Supervisors recommended that the Department of Elections consider allowing voters to rank all the candidates to avoid this issue.
The panel urged the department to work with city supervisors to increase voter education.
Hence the mascot. “We made the conscious decision to have an image of a correctly marked ballot and to have a smiley face to draw people’s attention,” said John Arntz, the director of the Department of Elections.
When asked whether ranked-choice voting has worked well for San Francisco, Arntz said, “I guess it depends if your candidate wins or not.”
This article also appears in the Bay Area edition of The New York Times.
Related
alison sf
OK, let's see what commentary the professional defendants of ranked choice voting can provide to remind us how smart they are, and how stupid the rest of us are.
Mission Rosalind
ROFLMAO
alison sf FTW!!!
Mission Rosalind
(especially if you really meant 'defendants')
Charles Jencks
Whats your problem?
Robert Montgomery
A single vote on a ranked choice ballot is a vote for one person and a vote against everyone else. That is, if your candidate doesn't win on the first round, you are choosing to not help another candidate get more votes.
I don't see the added advantage of ranking all the candidates. IN the past, when we only voted for one candidate, our ballot was "spoiled" if that candidate didn't win. Now it's spoiled after 3- should you choose to vote for that many.
What I haven't looked at, which would be interesting, is whether San Francisco is spending more money funding the zombies with public dollars than if we just kept the old system around... you know, where candidates rounded up their own money and there was a run-off for the top two contenders.
MJP
There's a difference between a vote being spoiled and one's choice not winning.
In the past when voting for only one candidate, your vote counted in the final tabulation. If no candidate received a majority of votes, we had a runoff election between the top 2 candidates, so every voter had a chance to vote a second time. Their votes counted in the final runoff tabulation, whether their candidate won or lost.
In this case, they are reporting that 16% of votes did not count at all in the final tabulation. That is the difference.
The only mathematically fair way to use RCV is to allow as many rankings as there are candidates. But when there are 16 candidates, I for one, do not feel like undertaking that chore.
That issue of fairness versus the cost and convenience, as well as turnout, of second elections needs to be discussed. It's a question of balancing difference priorities, and different people are going to put different emphasis on different ones. I don't know right now how it should work out.
Mission Rosalind
"The only mathematically fair way to use RCV is to allow as many rankings as there are candidates. But when there are 16 candidates, I for one, do not feel like undertaking that chore."
Yes. One big problem with allowing (# of votes = # of candidates) is decision fatigue. Outside of political insiders and activists, many voters either won't bother learning about all the candidates or will only vote for a limited number. Of course, this may be what RCV proponents have intended all along. In an environment of high numbers of candidates, limited information about each candidate, and few expressed differences between candidates during RCV campaigns, slate card mailers and endorsements will drive a lot of voter behavior.
For those interested in the latest research on decision fatigue, this is a good starting point:
http://www.nytimes.com/2011/08/21/magazine/do-you-suffer-from-decision-fatigue.html
cornholio
Too many choices also will mean too many spoiled ballots, as rushed voters accidentally fill in the same choice multiple times, or get the order wrong and ask for new ballots.
Windy
The likelihood of ballot spoilage increases exponentially when users are asked to rank so many choices. Worse, people don't vote at all, like nearly 60% of eligible voters in this election.
Only voting for one candidate isn't a wasted vote any more than skipping a proposition is. That's neutral, not hurting anyone.
I used all 3 mayoral slots because I had 3 candidates I liked and knew enough about. Guessing isn't good on the SATs or at the ballot box.
Christopher Pederson
Buried deep down in this article is a potentially legitimate concern. Unfortunately, you have to dig through a bunch of nonsense to get to it. Simply because a voter doesn't vote for three different candidates in a particular race doesn't mean the voter is confused about how to vote. I know perfectly well how ranked-choice voting works, but sometimes I deliberately decide to vote for fewer than three candidates. That doesn't mean I'm confused or disenfranchised. It just means that there were fewer than three candidates that I wanted to vote for.
That said, if certain demographic groups consistently under vote, it's worth looking into whether that's just the result of members of those groups making similar decisions about how to vote or is instead the result of widespread confusion among those groups. Too bad this article didn't bother to even scratch the surface of that issue.
Gary M.
Chris, I think it is clear that the reporter SAID:
"according to an analysis released this week by the University of San Francisco."
and their analysis showed:
"Perhaps the analysis’ most troubling finding is that 9 percent of voters, mostly in Chinatown and southeastern neighborhoods like the Bayview, marked only one choice for each office, likely either because they considered only one candidate suitable or because they did not know how to fill out their ballot correctly."
USF RESEARCHERS see that two demographic groups of voters only mark one choice (and they were different choices all around, not rallying behind one candidate). Statistically, this would point that these demographics did not know how to fill out the ballot. That's very troubling....
James F
This article reads like it was written by a reporter who doesn't understand ranked choice voting themselves.
Just because a voter doesnt choose three candidates doesn't mean they didn't "use" RCV correctly. I know lots of people who only voted for one candidate - they knew full well what the system was and their options. The candidates were just not good enough to earn those other votes.
Please spare us the sanctimonious crying about RCV - its mostly crying by entrenched journalists and politicos having trouble with the strategic side of it. They are the ones who dont understand it.
Really, now - asking voters to rank their top three preferences is no more complicated than asking them to rank their three favorite ice cream flavors. If they only like vanilla, that's their right.
The major flaw of RCV is in the disenfranchisement of exhausted ballots. Fix that or go back to top two. And lets end this arrogant "voters are stupid" analysis.
Windy
No, there are lots of us who understand RCV and still think it's awful.
I agree with your point about the interpretation in the article being questionable.
b s
"Simply because a voter doesn't vote for three different candidates in a particular race doesn't mean the voter is confused about how to vote."
Completely agree with this statement. The author of this article is making conclusions from the data that logically do not follow. Please do not assume a causal connection when one does not exist (or at least when you have not demonstrated it to exist).
Gordon
There is nothing wrong with voting for only one candidate, unless there is something else fishy going one, like the "coaching" that we saw with voters in Chinatown...
Windy
Bingo.
Max Norton
Has the BC forgotten that Willie Brown wrote a column just a couple days out from the election specifically recommending people vote for only one candidate?
Howard Epstein
The RCV experiment has failed. Its time to back to one person, one vote.
David Cary
RCV is one person, one vote. Every voter gets one vote and three choices for how that vote is counted in the instant runoffs.
If you want each voter to have just one vote, ask to change the school board and community college board elections to multi-winner RCV. In those elections it is not uncommon for much steeper under-utilization of votes than what we saw in this mayoral election.
David Cohen
Several comments correctly make that point that to vote for only one candidate does not necessarily indicate that the voter doesn't understand the system. While increasing the number of choices on a RCV ballot and further voter education might be helpful, this doesn't mean that RCV should be tossed. The advantages are well known, not hard to understand, and have been explained many times, including in comments to an earlier Bay Citizen article on this subject. What is mysterious is why Bay Citizen and Mr. Shifflett keep flogging this issue in such a biased way. The anti-RCV tone of the article is plain in the sentence saying that "16 percent of ballots ... were discarded". That's not just true, and seems to reflect a stubborn imperviousness to reason similar to that of deniers of anthropogenic climate change. Maybe Mr. Shifflett should go back to writing about drug-addled prostitutes.
David Cary
I'm confident that the Department of Elections has not discarded any ballots, that all the ballots are being kept in secure storage and all ballots were counted and accounted for in all twelve rounds.
Unless Mr. Shifflett has found ballots floating in the bay or in a dumpster, the Bay Citizen owes its readers a retraction.
The target of the Department's education campaign was not voters who only ranked one or two candidates. The target was voters who invalidly marked more than one candidate for a choice or who irregularly marked the same candidate more than once. The Department has always educated voters that the number of choices they mark is each voter's perogative.
The Bay Citizen owes its readers a retraction for suggesting otherwise.
Gary M.
Am I the only one who finds it unbelievable that we would have to spend $300,000, or over $1.50 per voter, to "teach" voters a system which has been used for a decade???
And it still didn't work?
David Cary
RCV worked quite well, certainly better than the regressive alternative the supervisors are being asked to consider. Unless you like a process where only 61% of the voters who cast a vote would have fully participated in choosing the Mayor.
When the City has a former mayor who encourages voters not to fully vote, when the city had a mayor who faked not knowing how to vote, and when the city tolerates ongoing misinformation calculated to confuse voters, then some ongoing education is probably needed. When the efforts to confuse voters stop, there will be less need for education.
The extent to which this particular education campaign was effective is however still subject to evaluation.
Gary M.
"irregularly marked the same candidate more than once"
This does't make any sense. Isn't this the same as voting for one candidate? Why is it irregular?
David Cary
Voting
1. Leland Yee 2. Leland Yee 3. Leland Yee
gets counted the same as voting
1. Leland Yee 2. blank 3. blank
But for some observers, voting the first way raises some concern about whether the voter might have misunderstood and thought it was a stronger vote for Leland Yee.
As long as the Bay Citizen is telling people that voting the second way is incorrect, or that even voting the second way for Ed Lee is incorrect, that is only creating a greater need for education to counteract the attempts to confuse voters.
Gary M.
This still doesn't make sense. The posters around town showed a frown next to a vote like 1. Yee 2. Yee 3. Yee
But you say
1. Yee 2. Blank 3. Blank is OK?
They are the same are they not? Both acceptable?
David Cary
A ballot marked:
1. Yee 2. blank 3. blank
would be accepted by a precinct scanner without warning and would count 1 vote for Yee until he was eliminated. In subsequent rounds it would not count for any other candidate.
A ballot with
1. Yee 2. Yee 3. Yee
would generate a warning from the precinct scanner. If you chose to have the scanner accept your ballot anyway, you could and your vote would be counted with the same effect in the RCV instant runoffs as if you had marked choices:
1. Yee 2. blank 3. blank
With RCV you have one vote, but mark three choices for how that vote can be counted.
If you mistakenly thought you that marking three choices for Yee would triple the number of votes counting for Yee, the scanner warning would give you an opportunity to get a new ballot card and remark it, choosing different candidates for your second and third choices. For example, if you remarked your ballot as:
1. Yee 2. Chiu 3. Lee
then your one vote would count for Yee until he was eliminated, then it would count for Chiu until he was eliminated, and then it would count for Lee until the final round, since Lee stayed in the running for all of the rounds.
Look at the poster online at:
http://www.sfgov2.org/ftp/uploadedfiles/elections/VoterEducation/2011_RCV_Poster_EN.pdf
Look at the notations to the right of each of the three examples.
John Palmer
Pick your poison between exhausted ballots or "exhausted voters" who don't show up for (manual) runoff elections. The percent of ballots that end up exhausted is lower on average than the number of voters that don't return for a December runoff.
The "confusion" issue is entirely manufactured by the pundits. People get it. Pundits and the political elite hate RCV because it prevents them from making sure who wins, and it's tough to handicap outcomes. RCV doesn't work for the machine, but it works for the voters.
John E. Palmer
SF resident and business owner
Chris Jerdonek
Wow, the second RCV hit piece by the Bay Citizen in 9 days -- and by the same "reporter"! You guys are really on a roll. Keep showing us how biased you are.
This article is complete BS. Just because a voter doesn't rank "all three choices in all three races" doesn't mean they used RCV "incorrectly" or have "trouble" using it. That's like saying if someone skips a ballot measure, they don't know how to vote yes or no. But that's what this article is implying.
There were plenty of political organizations that told their members and supporters not to rank all three in some race, for example the San Francisco Democratic Party, the SF Bay Guardian, and the SF League of Pissed Off Voters to name a few. See, for example, http://www.theballot.org/2011/sf
Why not mention facts like 73% of voters ranked all three candidates in the mayor's race, or 84% two or more? Or that 99.6% of voters cast a valid vote in the race? Because as they have repeatedly shown, the Bay Citizen clearly has it out for RCV and has no interest in publishing balanced stories on this issue.
Mission Rosalind
The Bay Citizen should be embarrassed for leaving out a critical disclosure about Latterman that even that rag, the Chronicle, knew enough to include:
"It should be noted, by the way, that Latterman worked on Chiu's campaign."
http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2011/12/05/BAG41M73K7.DTL
David Cary
Latterman is also a professed opponent of ranked choice voting.