Threats to and harassment of local officials present a significant challenge to American democracy by discouraging civic engagement, undermining the work of public servants, and creating unprecedented stress on the cornerstones of democratic society including elections, education, and public safety processes. A heightened environment of fear among local officials seems ubiquitous, but the data behind the phenomenon is opaque. This project is the first of its kind - an ongoing longitudinal study to systematically evaluate events of threats and harassment across the United States using public event-based data.

View Press Release | Download the Report

Why new data?

• At present, there are no other systematically tracked, public data available on a national scale. This dataset provides an initial assessment of known public threats to local officials across the US, as well as a clear agenda for expanding data collection efforts. 
• As opposed to surveys and stories, event-based data allow for empirical assessment of observable incidents. This allows analytical transition from perceptions to patterns, and the promotion of  more effective evidence-driven policy. These event data complement existing survey data showing how these issues are worthy of more attention and analysis. 
• Event-based data provide decision-makers with specific understandings of threat/harassment locations, targets, perpetrators, and the nature of threats in order to craft specific responses and mitigation strategies for local elected officials, law enforcement, community leaders, and public interest organizations.
• Impartial data about how threats and harassment function can reduce the politicization of events and provide clarity for policymakers seeking better intervention strategies. 

What does the dataset include?

• The Threats and Harassment Dataset (THD) is a product of collaboration across multiple organizations, integrating existing events collected by others and newly-coded events into a new dataset with 400 unique observations between January 1, 2020 and September 23, 2022.  
  • Incidents included are narrowly focused on threats and harassment to local officials, not state and federal targets. 
  • Initial coding focused on officials in three policy areas: election, education, and health. This broadens the scope of officials considered beyond mayors and other elected officials to include additional contested civic spaces.
• Details on threats and harassment of local officials specifically address:
  • Incidents beyond legal definitions of criminal conduct, encompassing incidents that a common person would find threatening or harassing. This acknowledges that even activity that is not illegal can impact democratic norms, continues to erode civic space, and create a climate of fear.
  • Information on both perpetrators and targets of threats and harassment. This allows concerned parties to understand more fully the impact on individuals and communities.
• Final incidents were produced after review of:
  • Over 10,000 news stories 
  • 3,000 evaluated incidents
  • Thousands of public protest events tracked in public datasets - the Crowd Counting Consortium (CCC) and the Armed Conflict Location & Event Data Project (ACLED).
• The Threats and Harassment Dataset is a living dataset, which will continue to grow with additional incidents from state and federal levels, new mechanisms for reporting incidents, and data collection on outcome variables. 

Findings: What does the initial data tell us?

The dataset offers a lens for understanding how threats and harassment are impacting all parts of local democracy--individual leaders, communities, and institutions. By observing incidents and trends in the data over time, researchers and policy makers can also better understand longer-term threats to our democracy. 

• The project specifically examined threats and harassment to local elections, health, and education officials, counting incidents across 43 states. Of the 400 cases observed, 40% were related to elections, 30% related to education (30%), and 29% related to health issues (overwhelmingly COVID-19 issues).
• Across issue areas, threats of death and gun violence are more than twice as common as any other form of threat (13%). Intimidation was the overwhelming form of harassment (29%). See Figure 4 for a full breakdown of tactics.
• Women officials were targeted at a higher frequency than others, totaling 42.5% of incidents. Adjusting for the proportion of women in local offices, we estimate that women are targeted 3.4x more than men.
• Approximately half of all recorded cases of threats and harassment targeted someone who had previously been targeted or who received multiple threats. The remainder were aimed at new targets.
• Education related incidents mainly involved the intersection of COVID-19 and education (61%), followed by so-called “critical race theory” (7%), and LGBTQ+ related issues (7%).
• Threats or harassment of election officials or poll workers span 21 states and make up about 35% of all incidents tracked. Of these incidents, the states with the highest percentage of threats or harassment incidents include Pennsylvania (16%), Georgia (14%), Michigan (13%) Wisconsin (10%), and Arizona (6%) which make up 59% of all threats or harassment to election officials or poll workers. These findings appear to reinforce the FBI’s analysis that threats are more frequent in states with contested election results and lingering election denial activism.

Recommendations

Based on initial data analysis, we offer five recommendations to improve data collection and better support community organizations and protect civic space.

1. Support robust, safe, and easily accessible self-reporting: Trusted avenues to receive reports to civil society monitors and partnership with law enforcement can both fill gaps in our understanding of chronically underreported events, as well as improve the field’s ability to support victims and targets of threats and harassment.
2. Diversify sources of reporting and incident collection: Data collection should be expanded to include reporting from journalists, social media, crowd-sourcing, and other sources in order to diversify and increase public reporting of observable cases. This will require careful documenting and clearer methodologies to verify non-traditional sources but may increase the accuracy of the total count. 
3. Increase data sharing and collaboration: Increased and ongoing data sharing and collaboration is necessary between civil society monitors, government, and civil society organizations. By finding secure and appropriate ways to share data about how communities are evolving in their response to threats and harassment, the government, analysts, and, most importantly, targeted communities, can improve understanding of the full picture. 
4. Elevate community responses: Future data collection can elevate community responses to threats and harassment, including the need to center additional future data collection efforts on tracking outcome variables, such as policy innovation, resignations, or civil and criminal penalties. This shift builds on the strength of longitudinal data structure and could perhaps create a typology of mitigation efforts to reveal more about the strength of democratic institutions and norms, beyond just incidents of threat. We could then ask questions about how effectively specific interventions can help protect communities against negative events. 
5. Invest in comprehensive policy frameworks to protect civic space: Stronger policy frameworks are needed to protect civic space. Analysis of the dataset shows how critical new policy coordination and new resources are to systematically address incidents of threats and harassment. These may include anti-doxing and privacy protections, better partnerships between sectors, and transparent reporting of incidents from social media platforms and law enforcement who already receive reports. Table 8 in the report provides a curated list of civil society and government resources and policy recommendations.

Download the report

To request the underlying data, please fill out this form.