One of the many blessings of the annual New England Political Science Association meeting is that it offers a chance for emerging scholars to present their work. Once again we are pleased to present a blog post based on the work of young scholars studying under the tutelage of the first-rate faculty at the University of Massachusetts at Amherst, this time on the efficacy of endorsements in the Boston mayor’s race.
Cameron Roche is a PhD graduate student at the University of Massachusetts, Amherst. He is an Assistant Director of the Umass Poll and is a research assistant on the NSF Cooperative Congressional Election Survey (CCES), working under the direction of Professor Brian Schaffner. His research interests include: American Politics, Congressional Elections, and Political Psychology. He can be reached at email@example.com.
Keith Lema is a Fulbright Scholar and graduate of the University of Massachusetts, Amherst. He was a Political Science and Economics dual major and member of Phi Beta Kappa. He was an undergraduate research assistant for Umass Poll.
The 2013 Boston mayoral election between Marty Walsh and John Connolly, provided politics-junkies with quite a close finish. While Connolly led his opponent for most of the campaign, his good fortune faded by the end of October. On October 30th, UMass Poll became the first organization to predict a Walsh victory outside of the margin of error. A few days later, Suffolk University released a poll confirming Walsh’s newly acquired lead. Less than a week later, Connolly was congratulating Walsh over the phone and preparing his concession speech. To further understand this outcome, we looked to media groups for their narrative surrounding the election.
Most media outlets issued similar accounts of the election. They consistently credited Walsh as the “union” candidate who beat Connolly because of his superior ground game. Certainly, Walsh’s relationship with unions provided an influx of money, organization, as well as boots on the ground for his campaign. However, UMass Poll data revealed evidence of a more nuanced narrative about Walsh’s win over Connolly: our analysis shows that endorsements also played a critical role in Walsh’s victory.
Voters, in every election and in their everyday lives, use heuristics to make decisions with minimal cognitive effort. In political science, party identification, candidate traits, ideology, and endorsements have all been identified as heuristics that voters use to in their decision-making processes. The Boston mayoral election provided the perfect, low-information setting for voters to use endorsements as a heuristic. First, Boston has a non-partisan mayoral election, which left voters without their most valuable heuristic: party identification. Second, the final segment of the election was composed of two white males with few, if any, stark differences on policy and personality. With few ways to identify a preferred candidate, voters had to rely on endorsements made by public figures, media outlets, or community leaders that they trust.
Figure 1: The distribution of primary candidate vote choice based on general election vote choice in the 2013 Boston mayoral election.
After analyzing our data, it became evident that the endorsements of former candidates may have influenced the votes of their former supporters. Preliminary voters often followed the direction of their former candidate’s endorsement. Walsh received the endorsements of preliminary opponents Felix Arroyo and Charlotte Golar Richie. As demonstrated in the graphic above, nearly all of Arroyo preliminary voters flocked to Walsh while Richie’s voters certainly preferred Walsh over Connolly. In contrast, Conley voters overwhelmingly followed his endorsement and supported Connolly in the final round. Beyond this pattern, Figure 2 shows that Walsh acquired the majority of endorsements in this election. He received more political, union, and community-based endorsements than his opponents. Many of these endorsements were certainly friends from the state legislature who have no jurisdiction in the Boston metropolitan area.
Figure 2: Types of Endorsement by Candidate
There is no doubt that unions greatly enhanced the efficacy and funding of Walsh’s campaign. Yet, endorsements also played a vital role. While not all endorsements are effective or positive for a candidate, in this election, courting former opponents clearly made a difference. The voters that these former candidates can deliver are highly motivated to incur the cost of voting and trust the judgment of their former candidates. To enhance their ground game and make inroads to new voting blocs, candidates should clearly spend time soliciting endorsements.
The data is available for analysis at thedata.harvard.edu/dvn/dv/umasspoll. The survey was administered to 600 likely voters in the Boston Metro Area between October 22nd and 26th by YouGov. Both landline and cell phone numbers were randomly selected to prevent an age bias in our sample. Sample data offered a 5.3% margin of error for the full sample of registered voters and a 5.9% margin of error for likely voters.