Last Primary Poll (of an electorate both energized & unaware)
Massachusetts primary voters are enthusiastic and turning out in droves - but most lack basic candidate information. It is time to invest in changing this sad reality.
Happy primary day, Massachusetts! For voters, there is a 100% chance of rain, a 50% chance you voted before election day - and a 50% chance you were undecided on at least one competitive statewide race.
Results of the final poll of this primary season at the bottom of this post (you need to eat your vegetables/civic responsibility info first). But first, PFP in the News:
“‘A record amount of unpredictability’ hangs over Massachusetts primaries” - The Boston Globe’s Matt Stout’s astute primary election preview features PFP/MassINC Polling Group data on how low voter knowledge is driving volatility heading into the Democratic primary.
OPINION on why Our Democratic Institutions are Failing Voters from PFP Policy Advisor John Griffin, former gubernatorial candidate Danielle Allen, and WelcomePAC co-founder Liam Kerr in CommonWealth. More voters will turn out for the Democratic primary this year than in any recent, similar election — but for the most part they are woefully uninformed. Our democratic institutions have failed to grab voters’ attention and have left most voters stating they lack the knowledge to make an informed vote.
Tune in to ‘GBH election night coverage to get PFP’s John Griffin on what this means for our democracy
Election season can be silly, but this primary election is sillier than most - although in a different way than last year’s municipal elections. There’s a lot of information being produced for voters, a lot of ads, and much higher primary turnout than in the last open gubernatorial election in 2014. But even with all this fervor, voters don’t seem to be getting the information they need to choose candidates in the cycle’s most competitive elections. With just days to go, many voters lack opinions of any candidate and still undecided who to support (some even if they’ve already voted!).
Looking into future cycles, there is an opportunity to connect high-quality journalism and information more directly to voters - or leave the gap open for self-funders or organized interests to define elections.
This lack of information is an even bigger problem given the large role that mail voting will play in this election. As reported by Steve Koczela and Rich Parr in a WBUR after the 2020 primary, communities that disproportionately voted on election day also voted at lower rates overall. In a low-information environment, early voters can take their time with vote choice - Googling candidates and weighing choices in real time - while Election Day voters may not. In 2020, communities where more people voted in person also tended to have larger populations of color and larger low-income populations. This election’s low-information dynamics, then, will only add to existing disparities in voter choice.
Analyzing the Electorate: What is the Threshold Problem?
So how can campaigns, journalists, advocates and civically-minded leaders in Massachusetts understand why the voters are so clued out of this race–and work to address the problem? Priorities For Progress (PFP) exists to analyze and engage the Massachusetts electorate with the goal of advancing pragmatic, results-oriented policies. As part of that project, PFP continually tests hypotheses about the dynamics in elections in the Commonwealth.
This election is also a good opportunity to illustrate how different elections in the same place and similar time periods can nonetheless suffer from very different threshold problems. This Tuesday’s threshold problem will be voters’ lack of knowledge or opinion about the candidates—which calls for one set of interventions. But other elections could have the opposite problem: low voter participation despite high levels of knowledge. PFP has a paper coming out soon on just such an election: the 2021 Boston mayoral election.
2021 Threshold Problem: Off-Cycle Dropoff
PFP’s upcoming paper from Policy Fellow Steph Domond features a randomized control trial (RCT) that tested interventions to increase Black voter turnout in the 2021 Boston mayoral final election in November. PFP undertook this experiment for three related reasons:
Most investment in getting out the vote focuses on highly-partisan, high-turnout November even year elections, like the tremendous work led by Stacey Abrams in Georgia to register and turn out the marginal voter in the most high-profile contentious elections on the calendar - gubernatorial, senate, and presidential elections in swing states.
There is little investment & research in addressing “dropoff” from high-turnout elections like the 2020 presidential election (despite few contested elections) to elections like the September 2021 municipal contests in cities throughout Massachusetts (despite many competitive elections that affect the daily lives of voters)
Public data shows that municipal elections suffer from low turnout generally. These elections are always held in odd-numbered years and therefore are not tied to races like those for Governor or Congress; turnout in municipal elections is higher in years that include a competitive mayoral election (like 2013 or 2021) but lower in years with no mayoral election (like 2017).
A PFP poll of Black presidential-year voters who did not vote in 2021’s preliminary election found room to increase turnout. To investigate the sources of lower turnout, PFP polled within the universe of roughly 36k Black Boston voters who had voted in the 2020 presidential primary but not the 2021 preliminary election for mayor of Boston. A plurality said they did not vote because they were busy, while just over a tenth cited a lack of familiarity with candidates.
Turnout was low despite high voter knowledge of the candidates. In PFP’s September 2021 poll right before the preliminary election, fewer than a quarter of voters said they had never heard of each of the four top candidates. In a Globe/Suffolk University poll conducted September 2-4, 2021, over 60% of respondents viewed Andrea Campbell and Michelle Wu favorably; Kim Janey was at 56%, and Annissa Essaibi George was at 44%. None of the four candidates had more than half of voters report neither a favorable nor unfavorable view. Fewer than 10% said they had never heard of any one of the four.
Taken together, these data points presented a puzzle: Unlike in this year’s election, voters generally had an opinion about the candidates on the ballot. But turnout did not reflect this high level of knowledge. So PFP tested how to increase turnout.
Testing Interventions to Increase Black Voter Turnout
To conduct that intervention, PFP conducted a randomized control trial (RCV) with a data analytics firm, to test the ability of a get-out-the-vote (GOTV) mailer and text messaging to increase Black voter turnout in Boston’s 2021 final mayoral election. A full description of the trial can be found in our upcoming paper, but to test the effects of these interventions, we compared the outcomes of voters who received:
An issue-based GOTV mailer encouraging them before the vote and a single text message on the day before the election, reminding the voter to vote the following day.
No treatment (control group)
Figure 1: An example of the mailer sent out to voters in the treatment group.
The takeaway: the mailer and text message increased turnout by 1.4 percentage points, with the 95% confidence interval falling between 0.5 and 2.2 percentage points. Overall turnout in the treatment group was 20.2%, versus 18.9% in the control group. While this may seem like a small difference, it does demonstrate at least one method to obtain a measurable increase in a voter’s likelihood to go to the polls.
Figure 2: 95% confidence interval for the treatment effect.
Figure 2: 95% confidence interval for the treatment effect broken out by political party of respondent.
2022 Statewide Election has the Opposite Threshold Problem
In 2021, PFP identified that low turnout and turnout disparities would be defining features of Boston’s mayoral election.
In this year’s elections, the problem is different. Turnout in the 2022 primaries will still be lower than in the general election or in presidential elections. Estimated at under 25%, this will still be the most politically tuned-in quarter of the electorate. At the same time, though, turnout will be twice as high as in the 2014 primaries. Those voters, though, have little knowledge or opinion of candidates for the competitive races on the ballot. While low turnout threatened democratic voice in 2021, low information is threatening the 2022 election.
Statewide Democratic Primary Turnout
2010-2018 Actual + 2022 Forecast
Source: Office of the Secretary of the Commonwealth
To that end, PFP is again researching interventions to address democracy challenges in the 2022 Democratic statewide primaries - but this time, it is voter knowledge. And polling in this election’s final days can help show how to understand the malleable and information-hungry electorate.
Final Poll of 2022’s Democratic Primary
While PFP partnered with MassINC Polling Group on three prior polls this year, this brief, large-sample tracking survey of 1,276 voters was conducted internally with the same live-caller method we used to accurately predict both the primary and general election results of last March’s 19th Suffolk special election.
Field Dates & Sample: 1,276 live calls over Labor Day weekend with statistically significant samples (303 on Friday, 511 Saturday, 462 Sunday) of randomized voters. To align with research uses this cast a wider universe that included ~5% ‘somewhat likely’ voters (similar to design of 2021 Boston mayoral preliminary poll that forecast Annissa Essaibi-George would win a low-turnout primary). Unfortunately, these are English-only calls.
Attorney General: Did Campbell Reverse the Surge?
After a summer of steady climbing, Shannon Liss-Riordan appears to have plateaued since late-August. The Liss-Riordan surge saw her jump from 6% in June to 26% in the PFP/MassINC poll released August 25 (and up to 29% and 30% in the two UMass polls released days later). All three polls were within 2 points, with Campbell leading two and Liss-Riordan leading one. In the final three days of August, the race momentum may have reversed:
Campbell 38%
Liss-Riordan 30%
This pattern may not hold. While Campbell appears strong in this large-sample survey (1,276 voters from 9/2-9/4), voters are still volatile and Liss-Riordan has continued spending aggressively (new filings show she will likely surpass $10m), and that volatility could still swing in the her direction among a low-information election-day electorate.
If the Campbell counteroffensive does hold, expect to see analysis of the final days of August:
August 29 - Globe endorses
August 30 - Palfrey drops out and endorses Campbell
August 31 - MassChoice (pro-Campbell) SuperPAC ad buy
Lieutenant Governor: Driscoll steady in front
Salem Mayor Kim Driscoll has been leading by small but stable amounts throughout the primary season. In 2018, more than 25% of Democratic primary voters left their ballots blank - a vote share higher than any individual candidate has right now.
Driscoll - 24%
Lesser - 16%
Gouveia - 8%
Auditor: Dempsey maintains lead
With 45% of voters undecided, the Auditor race surpassed LG in voter awareness. And Dempsey slightly expanded his edge - although nowhere near the divergence seen in the two UMass polls released in late-August (Amherst had the race tied, Lowell had Dempsey up 14).
Dempsey - 30%
DiZoglio - 25%
Information is Low. Volatility is High.
When presented with simple messages after the initial ballot test, voters move significantly - far more than expected at this late stage of a race.
For example, in the Auditor Race just one message - that The Boston Globe endorsed Chris Dempsey - shifts the race by 14 points and sharply reduces the share of undecided voters.
There will be three big winners on election night in the races for Attorney General, Lt. Governor, and Auditor.
Democracy in Massachusetts, however, is losing. Given how infrequently statewide incumbents lose, it may be years before we fully exercise our civic responsibilities again. Let’s be ready.