Introduction
Since 2016, election officials (EOs) across the United States increasingly invested in robust voter communication strategies on mainstream platforms such as Facebook, Instagram, and Twitter to restore the health of the election information ecosystem and build confidence in elections. In 2020, these efforts were supported through federal, state, and private funding, allowing state EOs to allocate resources to further develop and implement voter education strategies on social media (Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2023).
A key intervention to combat misinformation and build trusted networks of communication about elections is sharing one message in a consistent and repetitive manner: that EOs are trusted sources of information and that voters should turn to them for election-related information. To help with this effort, the National Association of Secretaries of State (NASS) launched the #TrustedInfo campaign, directing voters to use EOs’ websites and social media pages, so they receive accurate information about how to vote and look to EOs as credible, verified sources for election information. EOs bolstered this message with explicit references to the safeguards embedded in keeping elections secure in an effort to instill trust in the integrity of elections.Footnote 1
This article is part of a multi-year project that measures how EOs use voter education to position themselves as trusted sources of information, address misinformation, and inform voters about election processes. We introduce a dynamic dataset containing state EOs’ messages on Facebook, Instagram, and Twitter during the 2022 midterm election cycle. Using quantitative manual content analysis, we coded these messages for the presence of indicators we define as “trust-building,” in that they convey the message that elections are safe and secure and that EOs are trusted sources of information.
Our dataset represents the first effort to systematically track and thematically organize state EOs’ communications during an election cycle. With these data, we develop metrics to evaluate the role of EOs in improving the quality of democratic listening, illustrating how these data can be used to evaluate the relationship between EOs’ trust-building social media campaigns and public confidence in ballot counting. Finally, these data lend external validity to related efforts to causally specify the impact of messages from EOs on public confidence and other attitudes toward election administration.
Election officials’ trust-building strategies on social media
State EOs—Secretaries of State, Lt. Governors, or Elections Directors—are the authorities designated by federal and state law to oversee the conduct of elections.Footnote 2 By design, they are responsible for ensuring voters have positive experiences at the polls (Hale, Montjoy, and Brown Reference Hale, Montjoy and Brown2015). Responding to the increase in the public’s reliance on social media for information about voting and elections, EOs’ use of social media has increased significantly (Gross et al. Reference Gross, Baltz, Suttmann-Lea, Merivaki and Stewart2023), reflecting a commitment to building resilience against misinformation (Suttmann-Lea and Merivaki Reference Suttmann-Lea and MerivakiForthcoming).
Evidence from recent elections shows social media is an effective voter education tool. In states where EOs share information about election processes on social media, such as registering to vote and voting by mail, voters are more likely to navigate them successfully and participate in elections (Merivaki and Suttmann-Lea Reference Merivaki and Suttmann-Lea2023; Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2022). In 2020, voters whose state EOs invested in voter education communications on social media also expressed higher confidence in ballot counting. These effects held for individuals who voted for losing candidates and those concerned about voter fraud (Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2023).
Despite evidence of a positive relationship between EO voter education, voting access, and voter confidence, little is known about whether specific trust-building efforts are associated with greater trust in the election process nor which strategies are more effective in inoculating voters against misinformation. Research on political campaigns highlights the importance of “staying on message” so voters clearly connect the message to the messenger (Benoit et al. Reference Benoit, Glantz, Phillips, Rill, Davis, Henson and Sudbrock2011). In the context of election administration, consistent messages from EOs have the potential to build trust in elections and increase voter confidence (Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2023). Many voters view EOs as trusted messengers when it comes to election-related information, compared to political candidates or news media (Harbath et al. Reference Harbath, Fernekes, Orey, Suttmann-Lea and Wagner2022; Adona and Gronke Reference Adona and Gronke2018). It follows, therefore, that if voters are exposed to trust-building messages by their EOs, voters may form positive attitudes about election integrity.
To date, there is no established corpus of communication by EOs to their constituents across the United States on a medium of any kind, including social media. Messaging from EOs in the wake of the 2016 presidential election and in the lead-up and aftermath of the 2020 elections offers an opportunity to collect data on precisely what EOs are sharing with their constituents, which of their communications are “trust-building,” and whether their efforts have their intended impacts on the public. To support EOs in their campaign to combat election denial and misinformation in 2022,Footnote 3 the NASS, in coordination with other national professional associations like the National Association of State Election Directors (NASED), as well as federal agencies like the Cybersecurity and Infrastructure Security Agency (CISA), re-launched the #TrustedInfo campaign, a “public education effort to promote election officials as the trusted sources of election information.”Footnote 4
As the oldest nonpartisan professional organization of public officials in the United States, with 78% of its members serving as their state EO, NASS took the lead in providing resources on how to incorporate the #TrustedInfo2022 message on election websites and social media accounts almost a year ahead of the 2022 midterms. According to the NASS media toolkit, states were encouraged to officially announce their support for #TrustedInfo2022 and were expected to share trust-building messages on social media on a frequent basis, coupled with reminders for voters to look to official election sources, such as state and local election websites, for information about how to vote. This coordinated campaign establishes a baseline to measure the consistency of messages EOs share with voters and evaluate the effectiveness of these nonpartisan campaigns to shape public perceptions about election integrity.
The #TrustedInfo2022 dataset
Methodological approach and data collection
In this article, we introduce a dynamic dataset of all communications by state EOs on social media during the 2022 election cycle and develop metrics to assess the effectiveness of trust-building communication strategies on voter confidence. We first compiled a database of all states’ official EO accounts on Facebook, Instagram, and Twitter. In 2022, all state EOs were active in at least one social media platform, with Facebook being used by all states except Massachusetts. Thirty-nine states were active on Twitter, 17 on Instagram, and 13 operated an official page across the three platforms. We outline the data collection process for the EO database in Section 5.1.1 in the Supplementary Appendix.
Our next step involved the systematic collection of social media communications by all state EOs in our database. We partnered with the Algorithmic Transparency Institute (ATI.io), a program of the nonpartisan, nonprofit National Conference on CitizenshipFootnote 5, to collect organic content shared by a total of 118 state EO social media accounts between September 10 and November 30, 2022 (N = 10,000 posts).Footnote 6 We used ATI.io’s investigation and research platform Junkipedia, which is designed to monitor, track, and analyze misinformation on social media.Footnote 7 The platform facilitates the design of social media monitoring projects by allowing researchers to import their own social media account databases, create their own unique coding scheme, track, and code content shared on these platforms in real time, and export snapshots of coded data in csv format. For our project, we used the state EO social media database and imported our codebook, which was used to label content shared by state EOs on Facebook, Instagram, and Twitter during the target timeframe.
We employed quantitative manual content analysis and coded for the presence of variables that measure EOs’ efforts to combat misinformation and build trusted networks of communications. We coded each post for the presence of any of the 95 variables, which we aggregated into a hierarchy of seven thematic categories. We developed our taxonomy deductively based on our previous hand-coded content from the 2020 election cycle (Suttmann-Lea and Merivaki Reference Suttmann-Lea and MerivakiForthcoming). During this process, we refined the codebook deductively and inductively through pretesting with randomly selected samples of posts from narrow time periods during the 2022 cycle and through five rounds of coder training.Footnote 8
Our taxonomy, outlined in Table 1, includes seven thematic categories, and each thematic category has nested categories.Footnote 9 Type captures the purpose of a message shared by EOs: news stories about the EO or the jurisdiction which the EO serves; posts about deadlines as they are specified in the election code; and “here is how to register to vote,” “click here [link] to request a mail ballot,” or “here is how our office prepares for November 8, 2022” type of posts; outreach efforts to specific electorates such as language minority voters and the youth. We also capture platform-specific message types, such as replying to other people’s posts, sharing a thread, or retweeting.
Table 1. Coding taxonomy
Visuals captures how EOs utilize images and videos to convey their messages. Posts and videos were coded for the presence of human/non-human imagery. For non-human imagery, our coding scheme includes variables that capture how EOs customize graphics to communicate that the information shared comes from official sources, such as attaching the EOs’ logo on an informational flier. Access tracks the presence of languages other than English, and links that voters can click when instructed by EOs, or the presence of URLs/links that are unclickable.
Our next thematic variables track how consistently EOs signal to voters that they are trusted sources of election information by including the #TrustedInfo2022 hashtag in social media posts (Trust-Building). We also tracked the presence of words that EOs use to explicitly convey trust-building messages, such as “elections are safe and secure,” “visit [here] for accurate election results,” “your EOs is your trusted source for election information,” among others (Signaling).
Finally, we constructed eight nested categories within Theme to code posts with information about registering to vote, voting in-person early or by mail, and where to find a polling place or a sample ballot. We cover three election administration categories within the Theme category: ongoing, pre-election, and post-election procedures.Footnote 10
Strategies for effective social media use suggest that posting at least once a day on Facebook, at least twice on Instagram, and at least five times a day on Twitter is optimal for adequate content exposure.Footnote 11 Based on these recommendations, the minimum volume of posts for our study period of 81 days would have to be 81 for Facebook, 162 for Instagram, and 405 for Twitter, totaling 648 across all platforms. As we show in Figure 1, all states missed these thresholds for Instagram and Twitter, and 14 states exceeded the minimum number of daily posts shared on Facebook. Idaho, South Dakota, and Wyoming had the lowest social media activity (below 10 posts), with the latter sharing a total of two posts, one on Facebook and one on Instagram. The states with the highest social media activity were Maryland (784 total) and Iowa (734), with Kentucky (423), Tennessee (407), Illinois (359), Michigan (366), and West Virginia (344) following as the most active EO accounts.
Figure 1. State EO social media posts breakdown by platform.
Measuring election officials’ trust-building efforts
Recent scholarship suggests the absolute volume of posts may not be a good measure of effective social media usage and that prioritizing certain content prevents “drowning out” the core message that EOs want to communicate to voters (Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2022).
Since our focus is on trust-building communications, we measure state EOs’ adoption of the #TrustedInfo2022 campaign and explicit signals that EOs are trusted sources of information and/or elections are secure. According to the NASS media toolkit, a #TrustedInfo2022 pledge entailed sharing a statement, a press release, and/or a video from the EO’s official account, expressing support for the initiative on social media. We consider any of the three as “formal” pledges and construct our first metric as a binary measure (1 = official pledge; 0 = not official). States that used the hashtag but did not officially take the pledge are classified as informal users of the hashtag.
We also measure the usage of #TrustedInfo2022 hashtag by state EOs across the three platforms. Using the hashtag on social media was a key NASS recommendation, which is why we treat it as an appropriate metric of how consistently the initiative was implemented by states. Table 2 presents the variation in the volume of posts that included #TrustedInfo2022 among the states that shared at least one post with the hashtag by whether or not the state officially pledged to support the initiative. As Table 2 shows, usage of the hashtag is scant compared to the absolute volume of posts shared even among the most active state EO accounts (Figure 1). The Arizona Secretary of State was the most consistent user across all platforms, compared to all other states: about 50% of all Facebook posts included the hashtag, 53% on Instagram, and 30% on Twitter. The second most consistent usage was from the New Jersey Secretary of State’s Instagram account (66% of all posts) and the Delaware Secretary of State’s Facebook account, with 44% of all posts including the hashtag.
Table 2. #TrustedInfo2022 Pledge and usage by state EOs
Regardless of taking the #TrustedInfo2022 pledge, many EOs designed state-specific campaigns to send similar messages. Posts shared by the Colorado Secretary of State’s account, for instance, included #TrustedSource. All posts shared by the North Carolina State Board of Elections (NCSBE) included #YourVoteCountsNC. What is more, states would use the #TrustedInfo2022 hashtag when communicating election information not explicitly related to election security.Footnote 13 Finally, some states took the hashtag-less approach, such as the Wisconsin Election Commission, which launched an “Elections 101” video series, Georgia, which included a “Georgia leads” logo in all visuals shared on social media.
Coordinated and state-specific hashtag campaigns may not adequately capture the content and scope of trust-building efforts, so we also track explicit signals to voters that EOs are trusted sources of information and/or that elections are secure, what we describe as “trust-building messages.” As we show in Figure 2, there is variation in the volume of posts by state but also by platform, with Twitter being the most popular platform used by state EOs.
Figure 2. State EO trust-building social media posts by platform.
The Iowa Secretary of State was by far the most consistent messenger, followed by Michigan, Washington, Arizona, and West Virginia EOs. Many of these posts repeated the same message, that EOs are trusted sources of information (Figure 3, Iowa Secretary of State), with some posts making voters aware of election misinformation and to look to their EOs for accurate information (Figure 3, Illinois State Board of Elections).
Figure 3. Examples of posts coded as trust-building (Illinois) and trust-building—#TrustedInfo2022 (Iowa).
Using the #TrustedInfo2022 dataset to study voter confidence and trust in election administration
In this article, we present the first systematic measurement of EOs’ social media communications to build trust in elections during the 2022 US midterms (Merivaki Reference Merivaki2024). From a purely descriptive perspective, understanding how EOs operate within the election information ecosystem is important, particularly given their positions have become highly scrutinized, making them targets of online and offline harassment (Gross et al. Reference Gross, Baltz, Suttmann-Lea, Merivaki and Stewart2023). We show some states are more consistent messengers when it comes to communicating that elections are secure and EOs are trusted sources of information.
Our work builds on research demonstrating how EOs’ voter education efforts improve the quality of democratic listening (Merivaki, Suttmann-Lea, and Orey Reference Merivaki, Suttmann-Lea and Orey2023a); they enhance transparency about elections, increase voter exposure to accurate information from trusted sources, improve the voter experience (Merivaki and Suttmann-Lea Reference Merivaki and Suttmann-Lea2023; Mann and Bryant Reference Mann and Bryant2020; Atkeson, Alvarez, and Hall Reference Atkeson, Alvarez and Hall2015), and as a result, increase the likelihood of higher voter confidence (Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2023). Building from this research and political communication research on effective messaging campaigns (Benoit et al. Reference Benoit, Glantz, Phillips, Rill, Davis, Henson and Sudbrock2011), a key expectation is that consistent communication of trust-building messages will indeed increase trust in elections.
Since the 2020 US presidential election, EOs have invested in trust-building campaigns on social media, to communicate to voters that EOs are trusted sources of information and that elections are safe and secure. The 2022 midterms presented an opportunity to evaluate whether these efforts were effective, especially because they were designed to respond to the sharp decline in public confidence that votes are counted accurately. Using the dataset and measures presented in this article, we tested the relationship between trust-building efforts and voter confidence and found these efforts yielded positive outcomes; in states where EOs shared a higher proportion of trust-building messages, voters were more likely to identify their state EO as a top three information source about elections, and more likely to express higher confidence that ballots in their state were counted as intended. These findings hold even among voters who are more prone to election denial, such as voters who reported voting for Donald Trump in 2020 (Merivaki, Suttmann-Lea, and Orey Reference Merivaki, Suttmann-Lea and Orey2023a).
Our data show no clear relationship between the partisanship of the EO and social media usage in 2022.Footnote 14 However, it is noteworthy the most consistent trust-building messengers were Republican EOs, three of whom serve states where Trump won in 2020—Iowa, Ohio, and West Virginia. Research from the 2020 presidential election shows that voters living in states with Republican EOs expressed higher confidence in statewide ballot accuracy (Merivaki, Suttmann-Lea, and Orey Reference Merivaki, Suttmann-Lea and Orey2023b). These findings underscore that EOs are intentional in their efforts to build trust. They are consistent with extant research that finds these efforts directly respond to the national and state electoral context (Suttmann-Lea and Merivaki Reference Suttmann-Lea and Merivaki2023). Because state EOs miss the personal connection voters have with their local EOs (Merivaki, Suttmann-Lea, and Orey Reference Merivaki, Suttmann-Lea and Orey2023b), engaging in explicit trust-building campaigns and highlighting how these efforts are nonpartisan/bipartisan seems to be an effective strategy to improve public attitudes about ballot accuracy at the state level—that votes in one’s state are counted accurately—an outcome which voters may associate their state EO more directly than their local EO.Footnote 15
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/spq.2024.14.
Data availability statement
Replication materials are available on SPPQ Dataverse at https://doi.org/10.15139/S3/FLL3TR (Merivaki Reference Merivaki2024).
Financial support
The authors received financial support from the MIT Election Data Science Lab (MEDSL) for the research of this article. The authors received no financial support for authorship and/or publication of this article.
Competing interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Author biographies
Thessalia Merivaki (she/her) is an Associate Teaching Professor at the McCourt School of Public Policy, and an Associate Research Professor at the Massive Data Institute at Georgetown University.
Mara Suttmann-Lea (they/them) is an Associate Professor in American Politics at the Department of Government and International Relations, Connecticut College.
Mary-Catherine McCreary is a Research Program Specialist the the U.S. Election Assistance Commission.
Tyler Daniel is a Ph.D. Candidate in Public Administration at the Department of Political Science and Public Administration at Mississippi State University.
