Diving Into Our Straw Poll Data: Bullet & Anti-Voting

Our June 30 10th CD Straw Poll gave voters the opportunity to use three unique voting methods—score, approval, and plurality voting—simultaneously on the same four-candidate congressional race. This gave us the opportunity to directly compare how voters interacted with each ballot.

Here's some of what we found:

Bullet Voting

A bullet vote is a vote for a single candidate, even when the voting method allows people to vote for or score more candidates. Bullet voting essentially treats approval or score votes as if they were plurality votes. Plurality voting—our current voting method—is widely acknowledged by voting experts to be the worst voting method.

Because of this, one of the most frequently raised concerns with both score and approval voting is the assumption that voters will choose to bullet vote, whether that's for strategic reasons, or because they're used to plurality voting, or even just because they only like one candidate. If the vast majority of voters bullet voted for strategic reasons, they would undermine the positive properties of these alternative voting methods.

Examples of common plurality, approval, and score votes from our data

Examples of common plurality, approval, and score votes from our data

Our data do not show voters behaving that way.

Approval Voting

Comparing our straw poll's plurality and approval voting results, the plurality runner-up candidate—the incumbent—gained two approval votes. These additional approvals both came from voters who had voted for the plurality winner, moving the plurality runner-up from 21.1% of the vote (8/38) to 26.3% of the vote (10/38).

The plurality third-place candidate gained three approvals from the plurality runner-up’s eight supporters (37.5%) and seven from the plurality winner’s 28 supporters (30.4%), including two anti-votes.

An example of a bullet vote on an approval (left) and score (right) ballot

An example of a bullet vote on an approval (left) and score (right) ballot

An example of an approval anti-vote ballot (left) and hypothetical example of a score anti-vote (right). There were no actual score anti-votes in this straw poll.

An example of an approval anti-vote ballot (left) and hypothetical example of a score anti-vote (right). There were no actual score anti-votes in this straw poll.

An anti-vote is the act of voting to approve all of that candidate’s opponents. It's the mirror image of a bullet vote. One each of the plurality runner-up and the plurality winner's supporters cast anti-votes.

One of the advantages of cardinal voting methods like approval and score is that if a voter’s primary concern is opposing a particular candidate, they can do so without being forced to guess which of their opponents has the best chance of beating them.

The plurality fourth-place candidate gained only one approval.

Overall, every straw poll participant voted for their plurality choice again on the approval ballot. 39.5% (15/38) of participants chose to vote for at least one additional candidate when given the opportunity with an approval ballot. 5.3% (2/38) anti-voted.

And while a majority of ballots (23/38, 60.5%) approved only one candidate, the score section of the ballot indicates that most were honest in withholding approval; only 6 ballots did not approve of a candidate they scored higher than the halfway point. As such, it is fair to say that the remaining 22 voters (58.9%) were honest on their approval ballot.

Score Voting

When tallying our score voting data, only 10.5% (4/38) of our straw poll participants bullet voted, giving the maximum score of 5 to one candidate and minimum or blank scores to the others.

Overall voting behavior by voting method

Even with the loosest possible definition of bullet voting, where a voter supported only one candidate and gave the rest the same low score, our bullet voting rate increased to only 15.8% (6/38). No one truly anti-voted on the score voting ballot, giving one candidate a 0 and all others the same high score.

That means 84.2% of our participants used a broad range of scores, taking advantage of the expressiveness and nuance of the score voting method.

Takeaways

The proportion of straw poll participants who did not bullet vote is promising, especially considering there are only four candidates this year in the 10th Congressional District and the candidates have limited ideological overlap. We expect that with a larger or more similar pool of candidates, multi-candidate approvals would increase.

Voting behavior by voting method and WA-CD10 candidate

Interestingly, despite the blowout win for the first-place candidate, if our straw poll were the real election, the additional approval and score votes for the plurality third-place candidate would be enough to change which second-place candidate would advance to the general election under Washington state’s top two primary system.

Cardinal voting methods like approval and score voting free voters from worrying about vote splitting or which candidates are "viable." We at Counted want every voter to be confident they can express their true preferences and that the result will reflect the will of the people.


Read more about our straw poll here and check back soon for more analysis of our straw poll results, featuring min/max voting, subjective scales, and favorite betrayal.