by James Brooks, Alaska Beacon
November 5, 2022
After Alaska’s Aug. 16 special election for U.S. House, mathematicians Adam Graham-Squire and David McCune noticed something strange: If 6,000 voters for Sarah Palin had switched to Mary Peltola or not voted at all, Peltola would have lost the election.
The two, who study spoiler effects in ranked choice elections, wrote their findings in a September paper, concluding that it was the first known example of a “no-show paradox” in an American ranked-choice election. The paradox? If Peltola had gotten more votes, she might have lost.
After the result, there was some criticism of ranked choice voting, including a comment from Arkansas U.S. Sen. Tom Cotton, who called it a “scam.”
“These comments raise the question: is RCV some kind of crazy scam to rig elections?” they asked in their paper.
“The short answer is: No. (The long answer is: Nooooooo.),” they wrote.
Looking ahead, the two say it is possible to make strategic choices on a ranked choice ballot, but only if there is good pre-election polling.
“As a generic rule of thumb,” Graham-Squire said, “there isn’t really any good way — with the exception of this particular election.”
McCune is a professor of mathematics at William Jewell College in Missouri, and Graham-Squire works in the math department at High Point University in North Carolina.
They both say that because the Aug. 16 special election was so close to the November vote, and because three of the four candidates in both races are the same, the special election amounted to an enormous opinion poll with almost 189,000 participants, giving unusually good data about Alaskans’ actions.
The detailed results of that earlier election showed that if Begich had been in a head-to-head race with Peltola or in a head-to-head race with Palin, he would have won either contest.
“If I was someone who had voted for Palin previously and then put Begich second and Peltola third or not at all, I would be thinking hard right now about strategically voting and putting Begich first,” he said.
He and McCune, caution that in a ranked choice – also known as an instant runoff – election, it’s pretty hard to vote strategically because polls are inexact. And in this case, the electorate in August and the one in November will be different.
“I can only speak for myself, but I think most voting theorists would say that instant runoff voting, top four or not, is much less susceptible to strategic voting than plurality is,” McCune said.
“It’s difficult because you essentially need really good poll data, and this Alaska House election is a good case study of that,” he said.
There are signs that Democrats are already voting strategically in the U.S. Senate election.
In the Aug. 16 primary and in public opinion polls since then, Democratic candidate Patricia Chesbro has received less support than similar Democrats in prior statewide races.
Conversely, Republican incumbent Lisa Murkowski has polled in first place despite being censured by the Alaska Republican Party, which is backing a more conservative Republican, Kelly Tshibaka.
Chesbro said by text message that she doesn’t believe her performance is about her specifically, and that another Democrat would be in a similar situation.
Because Republicans outnumber Democrats in Alaska, Democratic voters seem prepared to back a preferred Republican over a Democrat who might align more with their values but would be less likely to win in a head-to-head matchup. (A third Republican in the race, Buzz Kelley, has suspended his campaign and is backing Tshibaka.)
A similar situation exists in the statewide race for governor, but there, the choices are less clear. Incumbent Republican Gov. Mike Dunleavy was the frontrunner in August and has been the leader in opinion polls since the election.
Two challengers, Democratic candidate Les Gara and independent candidate Bill Walker, appear to be running close together in second.
Gara appears to have more first-choice support, but Walker could attract the votes of Republicans who are unwilling to support a Democrat.
“The decision to vote strategically really depends on how you think the election will go,” Graham-Squire said.
“Generally speaking, voting your actual feelings is the best bet,” he said, but if someone has doubts, break the race into head-to-head matchups.
In a head-to-head race, would Gara be more likely to beat Dunleavy, or would Walker be more likely to beat Dunleavy?
Finding what’s called the “Condorcet winner,” the person who wins all possible head-to-head matchups, can make sense, he said.
“In this way, instant-runoff elections are similar to a top-two primary system, where you may want to choose the candidate who is the most ‘electable,’ as opposed to the one who is your actual favorite,” Graham-Squire said.
What makes this — and other strategic-voting decisions — risky is that they are based on polling data that may be wrong. Voting strategically could mean accidentally contributing to the defeat of a candidate you prefer.
Gara and Walker have each urged their supporters to rank the other man second, and both believe they will need all of the votes of their friendly opponent in order to beat Dunleavy.
Walker has also released ads that repeat a statement from campaign poller Ivan Moore, who has said that in every one of his polls, Walker fares better than Gara in a head-to-head matchup against Dunleavy.
The ads do not show Moore’s other comments, which indicate Walker still loses that matchup.
Alaska Beacon is part of States Newsroom, a network of news bureaus supported by grants and a coalition of donors as a 501c(3) public charity. Alaska Beacon maintains editorial independence. Contact Editor Andrew Kitchenman for questions: [email protected]. Follow Alaska Beacon on Facebook and Twitter.
Be the first to comment on "Strategic voting is possible but risky on a ranked choice ballot, mathematicians say"