Voters use candidate slates in the Cambridge City Council elections
In 2023, Cambridge, MA used ranked choice voting (RCV) to elect nine city councilors out of 24 candidates. With such a large field, many voters used slates to identify candidates who shared their values – even though the ballot did not identify which slate(s) each candidate was part of. Voters’ use of slates indicates that more jurisdictions should allow candidates to identify with slates on the ballot.
What are candidate slates?
A slate is a group of candidates who are aligned on issues and may choose to campaign together. Slates can be formed by candidates, or can come in the form of endorsements from local organizations. In some jurisdictions, like Minneapolis, candidates can indicate which slate(s) they are part of on the ballot. In other jurisdictions, like Cambridge, slates are not identified on the ballot, and voters must determine which candidates are part of which slates before going to the polls.
The slates in Cambridge included a range of parties and coalitions, like Cambridge Voters for Good Government (CV4GG) and Cambridge Bicycle Safety (CBS). The Cambridge Citizens Coalition offered a helpful visualization of which candidates were part of which slates. Some candidates were part of multiple slates.
How do Cambridge voters use slates?
We analyzed Cambridge’s cast vote record – an anonymized digital record of how each voter ranked the candidates – to understand how candidate slates impact voter behavior. The record shows significant evidence that voters considered slates when deciding how to rank their ballots, even though slate identifications weren’t shown on the ballots.
Voters selected candidates from the same slate with their first and second choices 46% of the time. In some slates, it was as many as two-thirds of voters.
We compared the share of voters who stayed within a slate for their top two choices to two “baseline” measures that estimate how voters would rank candidates if slates had no effect. The first baseline was the percentage of all candidates who belonged to the slate. The second baseline was the share of total second-choices received by the slate.
| Slate | Voters who selected a slate candidate 1st | Voters who selected a candidate from the same slate 2nd | % of 1st choice voters whose 2nd choice was from the same slate | Baseline 1: # candidates | Baseline 2: % of total votes |
| A Better Cambridge (ABC) | 13,210 | 8,951 | 68% | 35% | 54% |
| Cambridge Bicycle Safety (CBS) | 14,006 | 9,272 | 66% | 39% | 55% |
| Cambridge Citizens Coalition (CCC) | 9,248 | 5,531 | 60% | 43% | 38% |
| Cambridge Residents Alliance (CRESA) | 10,560 | 4,584 | 43% | 30% | 40% |
| Cambridge Voters for Good Government (CV4GG) | 4,938 | 1,261 | 26% | 22% | 20% |
| Democratic Socialists of America (DSA) | 2,259 | 432 | 19% | 4% | 10% |
| Our Revolution (OR) | 3,440 | 1,354 | 39% | 13% | 16% |
| Average | 8,237 | 4,484 | 46% | 27% | 33% |
For example, consider the A Better Cambridge slate (ABC). For Baseline 1, we consider that if a voter ranked one ABC candidate first, there are eight remaining ABC candidates out of 23 remaining choices. If second choices were distributed evenly then ABC candidates would earn 8/23 second choices, or 35%. Baseline 2 considers the popularity of the candidates on the slate. Overall, ABC candidates earned 54% of total second-choice preferences. We use 54% as Baseline 2 to determine whether ABC’s share of second choices was higher than 54% among voters who chose an ABC candidate first. In fact, ABC’s slate cohesion was 68%, beating both baselines by a substantial margin.
The two baselines are simply two different measures of how well a slate might have performed if there was no effect from candidates being on slates together. Baseline 2 is usually higher than baseline 1, but not always.
We found that all seven slates outperformed both baseline expectations, indicating that slates are a strong predictor of voter behavior. Below, we display the amount by which each slate outperformed the two baseline measures.
What factors affect the strength of slates?
There does not appear to be a strong correlation between the number of candidates on Cambridge City Council slates and the share of second-choice rankings they received, as shown in the chart above. The cast vote records show a significant difference in the strength of the slates. The ABC, CCC, and OR slates appeared to have particularly loyal voters, while the CRESA and CV4GG slates seemed to have much weaker cohesion.
Among voters who ranked an A Better Cambridge candidate first, 68% also ranked one of their candidates second, despite only a third of all candidates belonging to the ABC slate. Similarly, among voters who ranked a Cambridge Citizens Coalition candidate first, 60% also ranked one of their candidates second, despite their candidates receiving only 38% of second preferences. 503 voters ranked all nine ABC candidates before ranking another candidate, while 340 voters did the same for all 11 candidates on the CCC slate and 379 voters did the same for the four candidates on the OR slate. In comparison, only 66 voters did so for the CRESA or CV4GG slates.
Candidate slates are a powerful tool to help voters make informed decisions, especially in nonpartisan races. Coupling slates with ranked choice voting makes it easy for voters to identify candidates who share their values and cast an impactful ballot. Jurisdictions should make this even easier for voters by letting candidates identify which slate(s) they belong to on the ballot itself.
