June 6, 2018 — Hallie Roberts believes that STAR (Score Then Automatic Runoff) Voting will revolutionize the way people vote in Lane County, and throughout the entire country.
“People are aware of the problems with our voting methods,” Roberts said. “We end up with results that are not really what the people wanted. If we can implement a system that will lead us to more broadly favored candidates winning, I think our political process will be reinvigorated and people will be inspired to be a part of the process once they see it’s working more smoothly.”
Roberts is the campaign manager for STAR Voting, which was created by political scientist Alan Zundel and Mark Frohnmayer, a Eugene entrepreneur and creator of the organization Equal Vote Coalition.
The three are looking for signatures for an initiative that would both change how county officer elections are held, and how Lane County residents would vote for those candidates.
Roberts said the system has the potential to save candidates money while campaigning, greatly reduce voters and candidates gaming the political system and, eventually, doing away with the need for partisan politics all together.
It’s a tall order, particularly when looking back at the difficulties that have faced the American electorate in the preceding decades that have created labels like “spoiler candidates” and “strategic voting” that many believe point to fundamental flaws in how Americans choose their elected officials.
But is STAR really the right choice to fix these electoral ills? Is it a fool proof system, or are there lingering issues about the process that still need to be resolved? To find that answer, it’s best to look at what the STAR system is and how it relates to the electoral problems of the past.
Roberts will tell anyone about STAR. In an attempt to gather signatures for the group’s initiative in hopes of getting STAR into the November election in Lane County, she has been using every trick in the book to get the word out, from online videos to newspaper interviews.
“I’ll even go to your house to help explain it to you,” she said.
The initiative she is advocating for is broken into two parts, the first of which is to do away with May primaries for the non-partisan Lane County races for commissioners, sheriffs, assessor and district attorneys.
Instead, the votes would take place during the main November election.
Currently, these races are mostly decided in May. If a candidate wins the race with a plurality of votes, that person becomes the winner.
In the example of the May West Lane County Commissioner race between Jay Bozievich, Nora Kent and Beverly Hills, Bozievich received 55.75 percent of the vote, compared to Kent’s 38.61 percent and Hills’ 5.39 percent. Bozievich received over 50 percent, so he will move alone to the November general election ballot in November, clinching the win.
But having the vote in May can have problems. First, voter turnout is lower in primaries than it is in general elections, so a smaller portion of the population picks commissioners.
And that population is generally partisan. Since it’s a primary, most voters are concerned with voting for their candidates in the general election. Independent voters, or those disinterested in who wins a party’s nomination, may not get their voice heard.
In addition, a winner isn’t always declared in May. This is what happened in the East Lane County Commissioner race. Out of six candidates, none received 50 percent. The two highest candidates received 31.25 percent and 30.69 percent. Per Lane County rules, those top two vote getters will spill into a runoff in the November election.
The problem is, those candidates didn’t get there based on any large swell of public support. Thirty percent is a far cry from a majority. Even though one will eventually win a plurality of votes in November, the candidates got to that opportunity through a minority of partisan electors in May.
Those pushing for STAR voting hope to fix this dilemma in two ways. First, hold the primary of six voters in November, which would widen the pool of potential voters and allow more diverse voices to vote.
But it’s still possible that only 30 percent of the population would choose the winner. Will that person be a true representative of the electorate in Eastern Lane County? To fix that issue, STAR voting will be applied to the county races. It won’t be implemented in state or city races, but Roberts hopes that if it’s successful in Lane County, the system will be adopted throughout the rest of the state.
The current method of voting used in Lane County, along with most of the U.S., is what is called plurality voting. That is, a person is given a list of candidates and the voter picks one.
STAR voting on the other hand, is an alternative, two step-voting process that involves voters scoring candidates.
The first step is the initial vote made by the public. Instead of choosing just one candidate, the voter gets to rate each candidate between 0-5, with 5 being a score in favor, and 0 being a score of no-confidence. It’s akin to rating a business on Yelp or giving a star rating for a movie.
“The main advantage is you get the opportunity to vote honestly for whichever candidates you like,” Roberts said. “You can give whichever candidate you want a five, and then give your second favorite candidate a four or a three. Or you can give your least favorite candidate a zero, and you give the one that’s just slightly better than that a one. So, every step along the way, you show your preference.”
The theory goes, if a person is faced with two candidates that they like, they’re not forced to vote one over the other. Doing that can have detrimental effects on elections. An example of this would be the 2000 election with Ralph Nader and Al Gore, which will be covered later in this article.
With STAR, a person can show approval for both candidates without compromising their vote.
After the initial vote is completed, the top two candidates are put into an automatic runoff. It’s there that votes will be reexamined, where the candidate who scored higher most often is given the win.
“It’s about voting honestly and non-strategically,” Roberts said about the runoff.
She stated that without the runoff, it’s possible that people would just rate who they like with fives, and who they don’t like zeros. This is known as “bullet voting.”
“They’ll just try to ‘bullet vote’ all candidates they want to advance, and zero for candidates they don’t want to advance,” she said. “So, the automatic runoff is a step that incentivizes honest voting. If you know there is going to be a runoff, and if you give one candidate higher than another, and that one person gets your vote, that will inspire people to vote more honestly.”
An example of how this works can be seen in the recent Whiteaker Community Council election in Eugene, where the system was given its first (and so far, only) live test.
Data to Choose Representatives
The Whiteaker Community Council used the STAR method for its non-partisan, at-large seats. There were 11 positions open, with 14 people running for the seats.
Did the voting process create any major shifts in how the vote ended up?
“It’s possible for there to have been some changes in the last seat or two, but the first nine or so all had very solid support,” Brad Foster of the Whiteaker council said. “If I had to bet, I’d say it ended up pretty much as it would have.”
Foster does see promise in the system and believes it’s ready for a larger trial in Lane County.
“I also think STAR voting might help bring more diversity into local politics,” he said. “Races with multiple candidates from the same party are somewhat rare and appear to be actively discouraged by party activists. Under the STAR format, it wouldn’t matter if several people with similar, but slightly different, platforms ran in those races since voters could fine-tune their votes. Right about now, Democrats in California are wishing they had adopted STAR voting instead of Top Two Primaries.”
Foster was referencing the current dilemma congressional Democrats are facing in California’s primaries. In that state, the top two primary-vote winners wind up on the November election, regardless of their party. But one party could see two Republicans running against five Democrats. Democrats run the risk of diluting their vote between too many candidates. If that happens, it’s possible the two Republicans could gain the most votes and head to November, leaving the Democrats without a nominee.
“Overall, it was great,” Foster said about STAR. “This was a big improvement on our prior system of casting votes for up to 11 candidates. The old system felt like we were voting someone off the island by not including them in the 11, but STAR allowed us to grade the candidates in a more refined way.”
Foster also found some unexpected results in the data.
“The software gave us a bit of interesting data,” he said. “It ranked the candidates by their total points. It also looked at the election as though each seat was independent and compared the top two point-getters in the pool head-to-head. Those two methods of ranking have a couple of differences. So, Candidate A might get more points than Candidate B in the score portion, but Candidate B might win the automatic run-off. I think it worked that way for two of the eleven.”
What happened in the Whiteaker race was that for two candidates, the initial voting score was higher than their adjusted runoff score.
To explain, we’ll call “Candidate A” dogs, and “Candidate B” cats.
Dogs were very polarizing to the voters in the initial voting stage. Out of six votes, two voters really loved them, giving them a score of five.
Two voters were rather lukewarm on the animals, giving them a three. Two more voters absolutely hated dogs, giving them a big zero. On the whole, dogs gained 16 points.
Cat support was a little broader. Two people gave them a four, two people gave them a three, one person gave them a one and only a single voter gave felines a zero. The total vote for cats was 15.
In plurality voting, dogs would have come up the winner of the race, 16-15.
But was that vote actually indicative of how voters were feeling about the choices of household pets? Yes, some people really loved dogs, but just as many people hated them. For man’s best friend, they were pretty polarizing.
Cats, on the other hand, actually had broader support of the public. Sure, people weren’t as passionate about cats, but people also didn’t hate them as much. Felines appealed to a broader population of voters.
This is where the importance of STAR’s runoff comes in. It takes the top two winners of the initial election, then counts how many times each voter scored one animal over the other.
In two instances, dogs scored higher over cats. In one instance, dogs scored lower. But in three instances, cats actually scored higher.
“We believe that STAR voting will help us elect representatives with a broader base of support,” Roberts said. “You’re going to see candidates winning that have lots of threes and fours, those candidates that everyone can say, ‘Wow, I think they’re good candidates,’ rather than the polar extremes.”
Is it possible that candidates who receive a majority of initial votes will lose an election? Yes, as was shown in Whiteaker.
Nonpartisan organization FairVote, which champions electoral reforms, stated that because of this possibility, STAR runs the risk of violating fundamental democratic principles.
STAR proponents believe that without the rating and runoff of STAR, the intricacies that go into a person’s reasoning when it comes to voting is lost in the numbers. Even though the initial tally may equal a majority vote, that doesn’t necessarily mean the majority wholeheartedly agrees with the choice.
“Using data to choose your representative leads to the more scientifically or mathematically based result,” Roberts said. “The data shows what you prefer compared to the other candidates, and that information is used to elect the representatives to support the people that they want the most.”
It should be noted that while FairVote had multiple concerns regarding STAR, it remained neutral on the system, neither condemning or endorsing it.
“We don’t see STAR Voting as politically viable nor likely to work like its advocates believe,” FairVote wrote in December 2017. Instead of continuing to look at STAR, the organization stated they would continue to look at Rank Choice Voting, another form of alternative voting.
No matter what type of alternative voting solution someone supports, the point is that, in many cases, existing plurality voting can inhibit people from electing representatives that voters like the most. This can be caused by political parties, candidates, or sometimes the voters themselves.
One of the driving forces of STAR voting is to eliminate the “spoiler” candidate, someone who cannot possibly win an election, but gain enough support to throw an election for a similar candidate. The most common example given is the 2000 presidential race, where Green Party candidate Ralph Nader was considered a spoiler.
The last, major battleground for that election was in Florida, where only 537 votes separated Al Gore and George Bush, who would go on to win the election. That contentious election saw arguments on multiple fronts regarding how votes were tabulated, how people were registered and the readability of the voting cards. But some of the blame was placed solely on Nader being in the race.
In a 2007 study in the Quarterly Journal of Political Science, Floridian votes were examined and found that if Nader had not been in the race, Gore could have carried the state, thus giving him enough electoral votes to win the presidency. In
all, 97,488 Floridians voted for Nader.
The study estimated that 60 percent of Nader voters would have voted for Gore, thus giving him in the presidency. If those estimates are true, Nader being in the race “spoiled” the election for Gore.
Nader was a true believer in his cause and was attempting to win the election outright. However, some candidates get into a race with the express goal of tearing down another candidate. It was thought that then “Never Trump” Republican Mitt Romney would run as a spoiler in 2016, but he felt he could not fundraise in good conscience to play spoiler.
Some candidates even attempt to spoil the entire system, or at least game it beyond recognition.
Also in 2016, many hoped that “Never Trump” Republican David Evan McMullin would act as a spoiler for the president. He actually ran in a number of states, with a big showing in Utah. Some hoped that he would win that state and deny Trump the needed 270 Electoral College votes to win.
But McMullin had loftier goals. Since he wasn’t on the ballot in enough states to win the 270 electoral votes outright, his eyes were set on winning enough votes the ensure neither Trump nor Clinton could reach 270, which would leave the decision up to the House of Representatives.
McMullin failed at being a spoiler. His highest showing was in Utah with 21.54 percent of the vote, compared to Trump’s 45.54 percent.
It’s these kinds of shenanigans that STAR voting hopes to put an end to.
“You can vote for Nader if you want,” Roberts said of STAR. “Why should you feel like you can’t express that on a ballot? You should be able to express, ‘This is my favorite candidate, period.’ Vote honestly. And if you have a ballot that doesn’t allow voters to vote honestly, you have to ask yourself ‘Why?’ — and ‘What can we do to fix it?’”
However, political parties and their candidates aren’t the only ones who can engage in gaming the voting system. STAR voting is not completely immune to voters taking part in tactical voting.
FairVote found that there were indeed examples of tactical voting within the system. Following an example from FairVote, the Siuslaw News conducted its own experiment to see if tactical voting could be used to manipulate an election’s outcome. The experiment, which utilized soda brands, parroted FairVotes concerns.
Imagine a contest where people are voting for their favorite cola soda drink. The four choices are Coke, Pepsi, RC and Shasta Cola. The frontrunners are Coke and Pepsi, with RC coming in a close third and Shasta pulling up the rear.
Let’s say the majority of people would be happy with Coke or Pepsi and would rate either soda a four or a five. They could rate RC a four or a three, and Shasta a three or a two. After the initial election, Coke and Pepsi would head to the runoff, and Shasta and RC would be dropped from the running.
However, some Coke voters don’t just like their soda of choice, but they really hate Pepsi. Not so much the taste of Pepsi, as they find it pretty palatable. It’s the brand they hate. They grew up in a Coke town, but the majority of restaurants in their new city only serve Pepsi. It’s against their hometown values and they want it out of the picture completely. So, they decide to sabotage Pepsi.
They want Coke to win, so they give it a five. Pepsi gets a zero, of course. But RC and Shasta? They get a four rating. Now, they’ve never actually had RC and they find Shasta only palatable. But their goal is to trounce Pepsi out of the runoff. By overvaluing their like of RC and Shasta, and undervaluing their like of Pepsi, they raise the probability that either RC or Shasta will make it to the final runoff along with Coke, leaving Pepsi in the dust.
“I don’t think there’s any safety in using that strategy when you could elect RC Cola, and you don’t want that to happen,” Roberts said. “It would be very unwise for voters to try and manipulate the system that way. I can’t imagine a Republican voting for (2016 Green Party nominee) Jill Stein just because they don’t want Hillary to win.”
The voters would be taking a gamble with this. A large block of voters, upwards of 30 percent, would have to tactically vote together to make that strategy a sure bet. That’s difficult to do in mass numbers. But in today’s social media word, is it that difficult?
In the 2016 election, it was postulated that Democrats were signing up in Republican primaries to vote for Trump, thinking that Trump would be an easy target for Clinton. But so far, there isn’t any evidence that it actually happened. A Washington Post analysis of voting data in March 2016 found that the margin of Trump’s primary victories, compared to the percentage of Democratic voters, showed no sign of mass strategic voting.
But such a thing is not unheard of.
One of the greatest examples of tactical voting is in the 2000 election, again with Nader.
In that election, even many of Nader’s supporters didn’t feel he had a chance to win. Instead, their goal was to have Nader get 5 percent of the popular vote, allowing the Green Party to receive federal funding in the 2004 election.
But Nader supporters didn’t want Bush to win.
To fix this problem, Gore supporters in solid Republican states that had no chance of having their candidate win electoral votes would swap votes with Nader supporters in swing states. A Nader supporter in swing-state Michigan would vote for Gore, with the express promise from a red-state Gore supporter in Texas would vote for Nader. Gore would get a vote in a swing state in a bid to win electoral votes, and Nader would get a vote in a red state that would give him a chance to get the needed five percent popular vote.
People organized. The internet, which was just coming into its own, saw trading sites begin to pop up, putting red-state and swing-state voters together. According to a 2000 article in Slate, one website got 90,000 visits in one day.
But there were problems. In California, vote trading was deemed illegal and a popular site was shut down.
The public wasn’t as tech savvy as it is today, either. This was before Facebook and Twitter, and people were still learning how to navigate the world wide web. Many had to go actively searching for these sites or hope for an email chain to get involved, far from the ease of a Facebook post today that can potentially reach 5,000 people with the click of a button.
And time wasn’t on their side. These websites only became prevalent just one month before the election.
The Nader Trader experiment ended in failure, with Nader only receiving 2.74 percent of the vote, well below the five percent threshold needed to obtain federal funding. But with more planning and a better social media, it’s possible a larger population could have taken up the cause.
A more recent example of tactical voting took place in the Illinois Third Congressional District race, held last March.
In that primary race, the Republicans only had one candidate, Arthur Jones. The Brookings Institute, which ran a report on the race results on April 18, described Jones as a Holocaust denier.
The Democrats had two candidates, incumbent Dan Lipinski and newcomer Marie Newman.
It was a hard-fought battle for the Democrats. While is it generally difficult to unseat an incumbent, Lipinski was unusual in his party. He was anti-abortion and had differing views on healthcare and funding for Planned Parenthood.
Newman, on the other hand, supported universal health insurance and keeping abortion legal, and considered herself a progressive like Bernie Sanders.
Polls had Newman leading the race in December by five points. But by March, she was down by two. She ended up losing the election by 2,000 votes. The likely reason? People who voted for Trump, according to Brookings.
One in five Lipinski voters stated that they had voted for the president in 2016, compared to one in 20 for Newman voters. Brookings postulated that because many Republican voters found their candidate so unelectable, they jumped ship over to the Democratic primary and voted for the more conservative choice, essentially choking out the progressive candidate for the general election, leaving a Republican and a centrist in the final election.
Brookings does point out that there could have been other reasons for the change, including the possibility that they just liked Lipinski. But the numbers hinted at a substantial population taking part in tactical voting.
It is these types of examples that proponents of STAR voting see their system actually helping.
For the Illinois race, getting rid of primaries, having all three candidates on the ballot, and having a more robust system allowing voters to express their opinions on a candidate would eliminate the need for tactical voting.
Of course, one way to fix the system would be for each person to simply vote for who they want to win and leave out the “politics” of politics.
While many Democrats were concerned with their perceived notion of electability in the 2016 primaries as they decided between Sanders and Clinton, Republicans went full steam ahead and voted for a change with nominating Trump, who bucked the norm. They brought substantial change to the system without implementing a new voting system like STAR.
“I think that’s a fair assessment, and it will always be based on the exact players,” Roberts said.
Of course, as seen with examples like Romney’s potential spoiler bid, not every Republican was happy with Trump. Both Clinton and Trump’s approval ratings were historically low prior to the election, with Clinton holding a 52 percent unfavorable rating compared to Trump’s 61 percent.
If the primaries were done away with, and the ballot had Trump, Clinton, Sanders, Rubio, Ted Cruz and others, perhaps the election would go beyond just tactical voting and become something more substantive.
“I think that psychology is a big part of it,” Roberts said. “By giving people a new method of voting that allows them more expression on each candidate, I think psychologically it will break some people out of that trap where they feel like they have to vote for that person because the system is rigged to elect that person.”
That’s not to say that STAR would create a Trump loss or give another candidate like Sanders a win. The results could stay the same. But Roberts believed the process could help quell the divisive political party politics that has enveloped the nation over the years. Instead of voting for a party, people could just vote for individuals they think would just do a good job.
A First Step
STAR wouldn’t be the end-all-be-all to fixing voting in America. It’s just a first step.
“I think it’s a powerful, revolutionary first step,” Roberts said. “I don’t think it’s a cure-all patch for everything, but I think it’s absolutely a step in the right direction for our voting method. Whatever changes need to be made to go along side of it, we can address those one at a time.”
Even if money could be saved by not doing primaries, how each candidate gets their name out to the public will still be based on campaign donations.
“Money will still be a huge component,” Roberts said. “We do run into people who say that the number one problem is money in politics, and ask why we aren’t working on that. We understand that, and we do see that as a huge problem.
“Because it’s such an enormous task to work on money and politics that, rather than redirecting energy into ending Citizens United, we believe that this is something that is a smaller chunk of change that we have the capacity to effect here locally in our area. Rather than just being overwhelmed by money in politics, we believe this is an achievable goal here locally. It is a step toward freeing people to vote for their favorite candidates, which is huge.”
And there are still questions about how STAR will be implemented in the future. Right now, it is for non-partisan races. What will it look like when it gets into more heated, and financially backed, races statewide?
Will people become confused in how to use the system? And will they trust the mathematical algorithms used in the runoff?
That’s something for the voters to decide in November, when STAR is expected to be placed on the Lane County ballot.
For more information about the process, or to sign the petition to get STAR on November’s ballot, visit www.starlane.us or www.equal.vote.
Beyond Lane County, STAR is also running a twin campaign in Multnomah County.