The Legacy of Dennis S. Mileti and the Future of Public Alert and Warning Research

It was one of Mileti’s initial contributions to the field that sparked the focus of this special collection. The now foundational Communication of Emergency Public Warnings, which was co-authored with his long-standing collaborator Dr. John Sorensen, cataloged all available empirical research related to behavioral responses to alerts and warnings (Mileti and Sorensen 1990). The report offered a grounded summary of the state of the art, as well as an agenda for future research and practice. As we planned the collection, it became clear that something similar was called for in today’s world of escalating hazards. We felt an urgent need to take stock of Mileti’s research, as well as to chart a path to guide future efforts. We therefore issued a call for papers that invited authors to contribute theoretically grounded, empirical articles on innovative approaches, policies, and practices that have the potential to improve all-hazards public alert and warning systems. This particularly applies to those that will increase the access and efficacy of those systems and messages in the decades ahead.

Much has changed since Mileti and Sorenson wrote their seminal report, but their research still applies. In the decades since, we have experienced a technological revolution as the Internet and the World Wide Web have transformed the state of alerts and warnings. Just a few examples of these channels are the advent of 24-h cable news, mobile devices, streaming video, social media, the Internet of Things, and a Common Alerting Protocol that integrates alert and warning systems globally. With each new channel come new technological affordances including text, audio, video, icons, animations, and haptics. These innovations have made it possible to connect more readily with people who have access or functional needs, such as those with vision or hearing impairments, increasing the possibility that all members of the public can be reached under conditions of imminent threat. Furthermore, innovators continue to identify new markets and create new channels that can deliver alerts to people through cell phones, household appliances, vehicles, and the like. The increased reach and ability to alert people is truly astounding.

Even with rapid technological change, however, many of Mileti and his colleagues’ findings remain fundamental to today’s alert and warning research. Human beings are difficult to motivate, and it is even harder to get them to change their behavior (Mileti and Sorensen 1990). People will often mill about to confirm the message and the actions they should take (Wood et al. 2018). However, we know how to design effective warning messages that are actionable and that reduce behavioral delay (Wood et al. 2012). Not only has this body of work had a profound influence on risk communication research, scholars and practitioners have drawn from Mileti’s work to inform policy and practice. Indeed, a well-circulated checklist of the five key elements that make up an effective warning message (source, hazard, location, guidance, and time) has come to be known among some practitioners as the “Mileti Model,” which serves as a tool to guide their message writing (Sutton and Kuligowski 2019).

The findings in the papers of this special collection build on Mileti’s broad body of research. The call for papers sparked an outpouring of contributions that both acknowledge and underscore his legacy. They also offer new lessons in considering emergency communication via multiple channels, the effectiveness of images and maps in communicating complex information, the efficacy of nonimminent threat messaging, and the use of frameworks and tools to provide real-time and forecasted information during hazard events. Mileti (1999) was a longtime advocate of multidisciplinary and interdisciplinary scholarship, so it was encouraging to see the wide range of disciplines that the author teams in this collection represent.

James D. Goltz (Goltz 2023) describes how Mileti’s work has expanded our understanding of the social environment in which warnings are issued and provided clear guidelines on how warnings should be delivered to optimize behavioral compliance. Goltz situates Mileti’s work in the historical context of earthquake early warnings, showing how his observations and insights are foundational to contemporary hazard warning systems. Elizabeth Reddy, Julianna Valenzuela, Nicholas Yavorsky, Nina Guizzetti, and Cecelia Schroeder (Reddy et al. 2024) continue the discussion of Mileti’s contributions to earthquake early warning in their essay describing the “riskwork” involved in operating the ShakeAlert system. By capturing reflections from university, government, and industry partners, they identify the pervasive uncertainties in the institutional structures and processes of this early warning system.

Austin Harris, Rebecca Morss, and Pat Roebber (Harris et al. 2023) use agent-based modeling to investigate hurricane evacuation. They demonstrate how evacuation frameworks can be used to explore various scenarios by simulating the effects of vehicles, road networks, contraflow, evacuation order timing, and changing population characteristics. This work demonstrates that agent-based modeling can help diagnose potential challenges and implement effective intervention strategies. Using their research on environmental monitoring of roadway conditions during flooding, Pranavesh Panakkal, Elisa S. M. Fattoracci, Jamie Ellen Padgett, Danielle D. King, and Teddi Yoo (Panakkal et al. 2023) offer new strategies for real-time detection of hazardous events. Drawing from a series of stakeholder interviews, the authors develop a conceptual situational awareness framework and prototype map-based tool that provides real-time road condition data during floods. The framework facilitates emergency response decision-making during flooding—a key factor that leads to warning decisions.

Several essays contribute to our knowledge of messaging through content analyses, experiments, and mixed methods. Drawing from Mileti and Sorensen (1990) on the content required for an effective warning message, Erica D. Kuligowski, Nicholas A. Waugh, Jeannette Sutton, and Thomas J. Cova (Kuligowski et al. 2023) discuss their content analysis of Wireless Emergency Alerts issued for wildfires. They find that length limitations affect message completeness, with shorter messages being less complete. The authors also identify variations present in the “trigger language” used to motivate evacuation. The absence of consistent language across jurisdictions is identified as a future need for both research and policy. Elizabeth J. Carlson and Joshua B. Barbour (Carlson and Barbour 2023) use experimental methods to investigate the effect of messages delivered via mobile communication technology for risks that have an ambiguous level of severity to the message receiver. The authors find that messages containing content that promotes efficacy—that is, guidance and encouragement that the message receiver can use to make a good choice—perform better than messages that provide notification alone. Jeannette Sutton, Michele M. Wood, David O. Huntsman, Nick Waugh, and Savanah Crouch (Sutton et al. 2023) describe their study of how to improve location specificity in an earthquake early warning. They test the standard, text-only ShakeAlert earthquake early warning message and a message enhanced with additional location information via text and map. They find that, while participants prefer the text-and-map format, the enhanced text format is associated with significantly better outcomes. This finding demonstrates the need for evidence-based research rather than preference alone to inform message design for hazard warnings.

Populations with limited access to alerts and warnings are at the greatest risk when short-fuse, fast-moving hazards leave little time to prepare or respond. This is especially the case for those who may have limited English proficiency when the nation’s warning system is designed for English speakers. Joseph E. Trujillo-Falcón, América R. Gaviria Pabón, Justin Reedy, and Kimberley E. Klockow-McClain (Trujillo-Falcón et al. 2024) discuss interviews with Spanish-speaking community members following a tornado outbreak in Kentucky. Their aim is to learn how risk information is communicated via informal warnings and networks of trusted messengers. While milling serves as one strategy for information dissemination, the authors call for changes in warning systems to account for the growing demographic of non-English-speaking immigrants in the US. Rebecca E. Morss, Robert Prestley, Melissa Bica, and Julie L. Demuth (Morss et al., forthcoming) demonstrate how social media can facilitate informal information sharing through their study of tweets containing hurricane risk imagery. Using an integrated qualitative and quantitative approach, they find that information sharing online is one form of milling where people can engage in hazard response as they are exposed to, attend to, and make sense of evolving information.

This special collection is rounded out by discussion of multiphased, mixed-methods study conducted by Jeannette Sutton, Michele K. Olson, and Nicholas A. Waugh (Sutton et al. 2024) to create a lexicon for alerts and warnings. They build on Mileti’s hazard matrix, developed for Wireless Emergency Alerts, via content analyses, subject matter–expert interviews, and confirmatory analysis, leading to plain-language content to communicate protective action guidance and hazard impacts for 48 hazards. The Warning Lexicon also provides step-by-step recommendations for building complete, actionable, and consistent messages that draw from Mileti and Sorensen’s (1990) Warning Response Model.

Dennis S. Mileti died of complications from COVID-19 on January 31, 2021. He was 75 years old.

Before his death, Mileti made clear what he thought of messaging in response to the pandemic: “This is the largest public information mess I’ve ever witnessed… It just breaks my heart. We know how to do emergency planning better than anyone on Earth, and it’s not there” (Ripley 2020).

Mileti never stopped fighting for what he knew to be true: the best social science and interdisciplinary research improves alerts and warnings and ultimately saves lives. It was that commitment to applying science to the benefit of society that drew us and so many others into his orbit. Now, with this special collection, we see the field illuminated by an infusion of new scholars and novel approaches to understanding the state of the art. The essays in this special collection build on the knowledge base that Mileti helped to establish. But they also advance the field in important directions, just as he would have wanted and always encouraged. When viewed as a whole, this special issue is an important step in helping the next generation of hazard and disaster researchers continue building on his impact in the years to come. As we move forward, we hope that others will look to the past—reading pioneering works and engaging critically with all that has come before. This deep engagement and ability to question assumptions is just what we need.

We would like to thank Dr. Louise Comfort, who serves as Natural Hazards Review chief editor for the social sciences. Comfort was Mileti’s friend and colleague for many years, and has been an incredible champion and supporter of this special issue. We also thank the editorial team at the ASCE for production support, Jolie Breeden at the Natural Hazards Center for the editing expertise, and the anonymous reviewers who contributed so generously in providing feedback on the papers submitted for consideration. Finally, we offer much gratitude to the authors for their contributions to knowledge.