STAR voting (“Score Then Automatic Runoff”) combines score voting and two-round runoffs. Under STAR, voters assign a score to each candidate (typically between 0 and 5) and points are tabulated as votes over two rounds. The two candidates with the highest scores in the first round advance to a runoff. The finalist who is preferred by the higher number of voters wins the runoff.
STAR voting has not been used in any public elections for political office, but it has been used twice for political party elections in the United States. The Independent Party of Oregon used STAR in its 2020 primary election and Oregon’s Democratic Party used it to select delegates in 2020, though it used traditional, single-choice voting in the presidential primary .
STAR Voting has never been used in a public government election
RCV has a 100-year track record of delivering voter-preferred outcomes and improving representation. STAR Voting, by contrast, is untested in public government elections, yielding little evidence of how it would perform in real-world political elections.
STAR most likely leads to more inactive ballots in the final round
Both methods lead to some inactive ballots in the final round, but RCV appears to consider more ballots than STAR voting. In RCV, inactive ballots occur when voters don’t rank any finalists. In STAR, inactive ballots occur when voters don’t score either finalist or if they assign equal scores to both.
In RCV, 95% of ballots typically count in the decisive round. Early experiences with STAR voting suggest it may lead to many more inactive ballots in the final round.
For example, the Oregon Independent Party used STAR in their 2020 primary elections. The rate of inactive ballots in the final round, also known as “no preference” votes, for president, secretary of state, and treasurer were 5%, 30%, and 16%, respectively. In RCV, in comparison, a median 5% of ballots don’t count in the decisive round. At best, from what little evidence we have, it appears STAR can perform on par with RCV in high-profile elections where voters have strong opinions, like presidential elections. But in other elections, STAR considers fewer voters’ preferences in the final round.
RCV increases demographic representation
RCV has a demonstrated track record of improving representation for women and people of color. No evidence shows that STAR voting would have the same effect.
RCV is less susceptible to strategic voting
In STAR voting, votes for a back-up choice can harm voters’ first-choice candidates. Expressing support for a second-choice candidate — say, by giving them four stars — can propel them into the runoff round, ahead of the voter’s first choice. This may incentivize voters to strategically treat STAR ballots like “choose-one” ballots. They might give stars only to their favorite candidate — and not to other candidates whom they find acceptable. With few voters incentivized to give stars to a backup choice, this system essentially reverts to plurality voting, both in terms of which candidates can win and the incentive structure under which legislators operate.
Take the Oregon Independent Party primary for example.
Most voters “bullet-voted,” scoring only a single candidate. In the races for president and the secretary of state, only 45% and 49% of voters scored multiple candidates respectively. In both races, most voters treated the STAR ballot like a "choose-one" ballot, giving support to a single candidate.
The Secretary of State race stands out because six candidates — three Democrats, two Independents and one Republican — ran. In a crowded race including multiple candidates from the same party, voters should have had strong incentives to express multiple preferences, yet less than half engaged with the supposedly more expressive STAR ballot.
Ultimately, most voters have treated STAR voting like choose-one voting in its limited real-world tests, so it may not solve the problems of plurality voting.
In RCV, by contrast, voters’ backup choices are only taken into account if their first choice is defeated, eliminating the incentive to strategically deny support for second-choice candidates. In fact, 71% of voters in RCV elections rank multiple choices.
STAR voting might not elect a majority-preferred candidate
In STAR voting, backup preferences count at the same time as ratings for a voters’ first choice, and voters don't all use ratings the same way. For these reasons, the preference(s) of the majority of voters may be overridden, with majority-preferred candidates not advancing to the final round.
Take an election with three Democratic candidates and two Republican ones in a heavily Democratic city. STAR voting could lead to a Republican victory because it does not necessarily prevent vote-splitting between similar constituencies. This example shows how STAR voting fails the Mutual Majority Criterion. Star voting also fails the related Majority Criterion.
Consider this 2016 election for state treasurer in Washington state, which uses top-two voting. In this contest, a clear majority of voters preferred Democrats, but three Democrats split the vote, causing only Republican candidates to advance to the final round.
This was not a STAR voting election and each voter was permitted to vote for only one candidate in the first round. However, the results are instructive when considering the impacts of STAR voting and help to explain the questionable Independent Party of Oregon primary results below.
If voters in a STAR voting election don’t assign scores to multiple candidates, the results could be similar to the choose-one results in the Washington election above, in which a clear majority for one party fails to advance any of their preferred candidates to the final round.
In RCV, voters may rank backup choices without fear that doing so will undermine their first-choice candidate. Therefore, a majority faction can consolidate around a frontrunner candidate from their party without splitting the vote.
Subjectivity of preferences in STAR voting.
Different voters may interpret “five stars” in different ways, giving them different amounts of power over the election outcome. Voter A may think a neutral opinion equates to a 3, Voter B may consider a neutral opinion a 1 or a 2. Voter A may think a 5 means she agrees 100% with a candidate. Voter B may think a 5 means her favorite candidate, even if she doesn’t agree on policy entirely.
Consider Amy and David, neither of whom is excited about any candidate on the ballot. Amy gives three stars to one candidate and 2 stars to the others to express her middling support.
David feels the same way as Amy, but gives four stars to one candidate and zero stars to the rest. In this case, David has more impact on the outcome than Amy because they express their preferences differently. Just as rating a film on a scale of one to five stars is highly subjective, rating candidates on the same metric yields inconsistent returns.
With RCV, on the other hand, voters all interpret rankings in the same way - a straightforward expression of preference between candidates.
STAR voting rewards strategic voters and campaigns.
Savvy voters with robust understandings of STAR voting can ensure their vote is highly impactful, while other voters may miss out on this opportunity. Voters who are not excited about a candidate but want their vote to count more are incentivized to give their favorite candidate five stars and not rate any other candidates. Ironically, this system captures less nuanced preferences — the ostensible intention of STAR voting — and leads to fewer back-up choices on the ballot.
RCV, on the other hand, is less susceptible to strategic voting. Voters simply rank candidates in order of preference, knowing that a second-choice ranking will not harm their top choice, rather than worrying over exactly how much support to give to a second-favorite candidate.
STAR Voting could also incentivize hostile and polarizing gameplay between campaigns. Campaigns could encourage bullet voting to increase their chances of winning. For example, if a left, right, and center candidate (relative to the jurisdiction) ran in a STAR race, and all voters scored honestly, the center candidate would likely win, having gained many 3s and 4s from both sides. Sensing this, the left and right candidates could demonize the center candidate, so their respective bases don’t give scores to other candidates.
RCV, on the other hand, does not redistribute a voter’s ballot until his/her initial preference(s)
have been eliminated. This means a vote for one campaign does not equate to a vote against
another. In fact, candidates will rely on the re-distribution of ballots (the alternate preferences of
other campaigns’ supporters) to win in later round(s). Thus, candidates are incentivized to
coalesce with other candidates.
RCV is on solid legal ground
Jurisdictions that adopt STAR voting are rolling the legal dice. Unlike RCV and other systems, STAR voting has never been evaluated through the legal lens.
As noted above, voters have very different electoral power based on how they interpret and use their scores. Some voters will cast more first-round points than others, giving some voters more of a say in the outcome.
RCV, in contrast, has been upheld against every federal constitutional challenge brought to date.
Because STAR has never been used in public government elections, we have only limited evidence from party primary elections to gauge how it works in practice.
In 2020, the Oregon Independent Party used STAR voting in its presidential primary. Voter turnout was very low, with fewer than a thousand voters out of more than 125,000 voters registered in the party. The Independent Party looks to candidates of all affiliations as potential nominees, including Republicans and Democrats. The results were odd on their own terms: voters backed Joe Biden handily over Donald Trump for president, for example, yet none of the three Democrats (including the one ultimately elected) even made the top two in the Secretary of State race.
This is a deep dive into the Oregon Independent Party primary for secretary of state in 2020, which used STAR voting.
In the first round, Republican Kim Thatcher and independent Ken Smith earned the most points and advanced to the runoff round.
The three Democrats divide the vote and none of them make the runoff.
In the second round, Thatcher won with only 36% of the vote, followed by Smith with 34%. 30% of ballots were inactive because they either did not assign stars to either finalist, or assigned the same score to both.
Of the 30% of voters who expressed no preference, some could have assigned each the same score, and some voters may not have given either a score. Most of the voters whose ballots became inactive were those who preferred Democratic candidates.
According to STAR advocates, of the 30% inactive ballots:
In today’s highly polarized environment, it is highly unlikely that those Democratic-leaning voters genuinely had no preference between the Republican and Independent. Had they felt empowered to express sincere preferences — without fear that expressing support for the Independent would undermine Democratic candidates — the Independent would likely have easily overcome her two-point deficit in the final round and won the race, yielding a result that likely would better reflect voter preferences.