This week, while midterm votes are being counted, few observers would be surprised if someone attacks election officials. As many have observed, threats and harassment against elected officials and their families have increased dramatically over the past several years, including the recent violent attack on Paul Pelosi, husband of House Speaker Nancy Pelosi.
In Arizona, a judge restricted armed groups that had intimidated voters as they deposited ballots in drop boxes. State and local public officials have endured doxing, armed protests at their homes, personal and online threats, vandalism, and, of course, actual violence. Though few Americans say they support political violence, threats and harassment against elected officials are straining U.S. democracy.
But exactly how common have such threats against local public officials become? Without the answer, analysis and understanding are difficult. To help further public understanding, we have begun tracking that data systematically. We hope this empirical lens will provide greater insight so that policymakers, local leaders and others can craft mitigation strategies to support officials and their communities.
Introducing the Threats and Harassment Dataset (THD)
To create Princeton University’s Bridging Divides Initiative’s new Threats and Harassment Dataset (THD), we evaluated tens of thousands of news sources with a targeted LexisNexis search string and combined data from partners such as Anti-Defamation League (ADL), Armed Conflict Location & Event Data Project (ACLED), National League of Cities, Brennan Center and Prosecution Project. We define “threats” as instances in which one person communicates to another an intention to inflict pain, injury, damage or other hostile action. We define “harassment” as instances of knowing and willful conduct that a reasonable person would consider aggressively pressuring, intimidating, alarming, tormenting, or terrorizing, but stopping short of assault. We include both criminal and noncriminal incidents of threat and harassment.
Ours is the first ongoing longitudinal study that systematically evaluates such incidents across the United States using event-based, public data. As of this writing, we have recorded 400 cases across 43 states occurring since January 2020, all occurring at the local level. Moving forward, we will scale up tracking to include all local, state and federal incidents, using the same methodology.
Since there has been a limited amount of public reporting on threats faced by local officials, we are likely undercounting cases, particularly online threats. Local officials — especially women, racial and religious minorities — may often face additional targeting if they publicly report incidents. As our initial data collection relies on media reports, the undercounting is a known limitation. Reporting probably falters particularly with what are called “lawful but awful” cases, in which threats do not cross the line into crimes. Since law enforcement officers don’t take action on such cases, there’s little incentive for officials or the media to report publicly, and no prosecution to follow up on.
Who is being threatened?
To date, we have focused on election, education and health officials. We found that a plurality of threats targeted election officials or poll workers (35 percent), followed by school officials (31 percent), other locally elected or appointed officials (21 percent), and health officials (12 percent), as you can see in the figure below.
Women were targeted in nearly 43 percent of incidents, although the Rutgers Center for American Women and Politics reports that women make up only 30.5 percent of officeholders. Given that figure, we estimate that women are targeted up to 3.4 times more often than men, which is consistent with others’ findings on sexualized and gendered threats against women in politics.
Of the 400 cases in our data, 40 percent were related to elections, 30 percent to education, and 29 percent to public health, overwhelmingly covid-19. About 34 percent of all the threats mentioned death and gun violence; these spiked during November 2020 and January 2021, during and immediately after the presidential election and congressional certification of the results, and in August 2021, as officials issued coronavirus pandemic mandates for public schools. One such death threat arrived in Philadelphia in December 2020, as an email to City Commissioner Al Schmidt’s wife. With the subject line “Albert RINO Schmidt, committed treason,” the email warned that Schmidt should “tell the truth” or their three kids would be “fatally shot,” and then mentioned their children’s ages and address. The email was signed “Q.” Schmidt and his family were forced to leave their home and live under 24-hour police protection.
The education-related incidents mainly involved pandemic policies (61 percent), followed by “critical race theory” (7 percent) and LGBTQ issues (7 percent). Some of these also involved threats of violence. For instance, on June 14, 2021, a Brainerd, Minn., resident told School Board members that critical race theory was “demonic” and warned he would “dump hot coals” on all their heads.
What does this mean for voting and election officials?
While American election systems are safe and fair, according to numerous studies, unfounded claims of election fraud both discourage voting and undermine election officials’ work. We found incidents of threats or harassment targeting election officials or poll workers across 21 states, accounting for 34 percent of all incidents tracked. States that had the highest percentage of reported incidents targeting election officials include Pennsylvania (16 percent), Georgia (14 percent), Michigan (13 percent), Wisconsin (10 percent) and Arizona (6 percent). Combined, these swing states — and states with high levels of false election fraud claims — account for 59 percent of all threats or harassment incidents.
These findings align with the FBI’s recent warning that identifies heightened risk in the same states, based on direct reports from election officials. Fortunately, many of these states have robust safeguards for voters and election workers to deter aggressive electioneering, intimidation and the presence of firearms at polling locations.
THD is a “living” data set. We plan to expand the federation of civil society organizations and researchers tracking and reporting incidents. As we continue, we will analyze state and federal targets, release additional demographic information about threat targets, and examine the results of measures put in place to reduce threats.
Threats and harassment against local officials discourage Americans from getting involved in public service, undermine the work of those in public positions, and strain elections, education and public safety. All these endanger democratic stability. We hope that the Threats and Harassment Dataset helps those making evidence-based decisions to protect civic spaces.
Joel Day, PhD, is the research director of Princeton University’s Bridging Divides Initiative.
Aleena Khan is a PhD candidate in political science the University of Illinois at Urbana-Champaign and associate policy researcher at Princeton University’s Bridging Divides Initiative.
Michael Loadenthal, PhD, is the executive director of the Prosecution Project and a research team lead at Princeton University’s Bridging Divides Initiative.