This chapter summarizes research on public response to alerts and warnings. It starts by reviewing results of research from the 1970s to the 1990s and then turns to more recent research, including work sponsored by the Department of Homeland Security (DHS) that explored public response in the context of Wireless Emergency Alerts (WEA) system and other emerging technologies.1 The research summarized in this chapter is only a subset of a large body of work done on emergency alerts and warnings. Summarized research was selected based on relevance, with an emphasis on work funded by DHS (in response to the committee’s charge). Attempts were made to summarize not only long-standing research but also research that looks at newer tools and technologies, such as mobile devices, WEA, and social media.
There is almost always a delay between when an alert is received and when the recipient takes a protective action. That time gap is known as the protective action initiation (PAI) time (Box 1.1 outlines the warning process). One major area of research aims to shorten that time period. This section starts by discussing factors that influence a person’s PAI time, including milling, reunification with intimates, and the time it takes to
1 Principal investigators funded by the Department of Homeland Security Science and Technology directorate were asked to provide brief summaries of their work to the committee. These summaries are printed in full in Appendix B.
prepare for the possible protective action. It then turns to how message content, message context, and message receiver characteristics can also impact PAI.
The key PAI question is “What delays people from taking a protective action upon receipt of a first alert/warning or observation of environmental or social cues?” Between the point of receiving a message and the point of taking a protective action, people generally engage in a variety of activities that reconstructs their perception of safety into a reception of personal risk, creating the delay between the alert and the action.
Milling, Reunification, and Preparedness
People often seek confirmation from others regarding alerts and warnings, which is a process referred to as milling. Through milling, people form ideas concerning personal safety, risk, and what to do about it. Individuals during this time engage in a series of activities designed to increase their comprehension of the event, which includes understanding, believing, personalizing, deciding, and searching and confirming (Box 1.2). Milling occurs regardless of the hazard type, the warning delivery technology used, or the source of the warning. Hence basic human nature creates a response gap for most people between getting an initial alert/warning and initiating a protective action.
In addition to milling, the response gap is also affected by two additional factors: reunification of intimates and protective action preparation.
People are less likely to initiate protective action until all members of their immediate family have reunified.2 This normally involves waiting
2 T. Drabek and J. Stephenson III, 1971, When disaster strikes, Journal of Applied Social Psychology 1(2):187-203; R. Mack and G. Baker, 1961, The Occasion Instant: The Structure of Social Responses to Repeated Air Raid Warnings, Disaster Study No. 15, Washington, DC: National Research Council, National Academy of Sciences; and R. Perry, 1987, Disaster Preparedness
for family members to assemble, going home from work, or picking up children from schools, day care, or other locations. Both early and recent research shows that households with children or pets are inhibited from
and Response Among Minority Citizens, pp. 135-151 in Sociology of Disasters: Contributions of Sociology to Disaster Research (R.R. Dynes, B. DeMarchi, and C. Pelanda, eds.), Milano, Italy: Franco Angeli.
There is also the protective action preparation, which is the time required to organize resources to implement the protective action. This period can involve a variety of actions, depending on the hazard. Actions can range from assembling resources such as emergency supplies, clothing, food and water; filling the evacuation vehicle with gas; or securing the home from hazard impacts or human intrusion. It is logical to hypothesize that people who are better prepared (e.g., they keep shelter kits in cars or have emergency food and water supplies pre-packed) will be less likely to delay taking the recommended protective action. Similarly, the more a message recipient knows about the hazard, the protective actions associated with that hazard, or how warnings about that hazard might be delivered to them in their specific location, the more likely they are to act promptly.5 Members of groups that are socially isolated are more likely to either fail to respond or not respond promptly to a message.
Message Characteristics Influencing Protective Action Initiation Times
Decades of work has identified that a variety of message characteristics—including content, style, length, delivery, and type of recommended protective action—influence public response.
3 T. Carter, S. Kendall, and J. Clark, 1983, Household response to warnings, International Journal of Mass Emergencies and Disasters 9(1): 94-104.; R. Lachman, M. Tatsuoka, and W. Bonk, 1961, Human behavior during the tsunami of May, 1960, Science 133:1405-1409; K. Wilkinson and P. Ross, 1970, Citizens Response to Warnings or Hurricane Camille, Report No. 35. State College: Social Science Research Center, Mississippi State University.
4 T.E. Drabek and K. Boggs, 1968, Families in disaster: Reactions and relatives, Journal of Marriage and the Family 30:443-451; S. Heath and M. Champion, 1996, Human health concerns from pet ownership after a tornado, Prehospital and Disaster Medicine 11(1):67-70; S. Heath, P. Kass, A. Beck, and L. Glickman, 2001, Human and pet-related risk factors for household evacuation failure during a natural disaster, American Journal of Epidemiology 153(7):659-665; A. Edmonds and S. Cutter, 2008, Planning for pet evacuations during disasters, Journal of Homeland Security and Emergency Management 5(1); L.K. Zottarelli, 2010, Broken bond: An exploration of human factors associated with companion animal loss during Hurricane Katrina, Sociological Forum 25(1):110-122; T. Litman, 2006, Lessons from Katrina and Rita: What major disasters can teach transportation planners, Journal of Transportation Engineering 132(1); R.J. Blendon, J.M. Benson, C.M. DesRoches, K. Lyon-Daniel, E.W. Mitchell, and W.E. Pollard, 2007, The public’s preparedness for hurricanes in four affected regions, Public Health Reports 122:167-176.
5 D. Glik, K. Harrison, M. Davoudi, and D. Riopelle, 2004, Public perceptions and risk communication for botulism, Biosecurity and Bioterrorism: Biodefense Strategy, Practice, and Science 2(3):216-223; J. Haas, H. Cochrane, and D. Eddy, 1977, Consequences of a cyclone on a small city, Ekistics 44(260):45-50; and M. Lehto and J. Miller, 1986, Warnings, Vol. I: Fundamentals, Design, and Evaluation Methodologies, Ann Arbor, MI: Fuller Technical Publications.
The research record provides repetitive evidence that public warning messages are more likely to motivate appropriate and timely public protective actions if the warning messages contain information on five topics: guidance, time, location, hazard and consequences, and source.6
Research has shown that people increase responsiveness when they receive guidance on exactly what to do to maximize their safety and how to do it, including information on what factors to consider when deciding on whether or not to stay.7 Time and location are important content for a message because it informs people when they should begin taking a protective action and by when they should have it completed,8 in addition to saying exactly who should and who should not take action in terms that the public can readily understand (e.g., physical geographical boundaries of those who will be affected).9 Messages also contain information about the specific hazard the warning is for and the potential consequences (e.g., hurricanes can cause flooding), which gives context to the guidance being given in message. To take that guidance, however, people must trust the source. Individuals will turn to their neighbors and family to verify warning messages.
Research from the 1990s provides strong evidence that public warning messages with certain style elements work best. Those factors are typically referred to as warning or message style and include clarity, specificity, accuracy, certainty, and consistency. People are more receptive to messages that are free of jargon and are written in words that most people can understand.10 People also want specific language that gives
6 D. Mileti and J. Sorensen, 1990, Communication of Emergency Public Warnings: A Social Science Perspective and State-of-the-Art Assessment, Oak Ridge, TN: Oak Ridge National Laboratory, U.S. Department of Energy.
7 T. Drabek, 1999, Understanding disaster warning responses. Social Science Journal 36(3):515-523; D. Mileti and C. Fitzpatrick, 1991, The causal sequence of risk communication in the Parkfield Earthquake Prediction Experiment, Risk Analysis 12(3):393-399; J. Sorensen, 1991, When shall we leave: Factors affecting the timing of evacuation departures, International Journal of Mass Emergencies and Disasters 9(2):153-165.
9 T. Drabek, 1999, Understanding disaster warning responses, Social Science Journal 36(3):515-523; and D. Mileti and C. Fitzpatrick, 1991, The causal sequence of risk communication in the Parkfield Earthquake Prediction Experiment, Risk Analysis 12(3):393-399.
10 L. Bellamy and P.I. Harrison, 1988, An Evacuation Model for Major Accidents, Paper presented at the IBC Conference on Disaster and Emergencies, London, April; J. Nigg, 1987, Communication and Behavior: Organizational and Individual Response to Warnings, pp. 103-117 in Sociology of Disasters: Contributions of Sociology to Disaster Research (R.R. Dynes,
precise and non-ambiguous information about the area(s) at risk, how much time they have to engage in protective actions before impact, and the source of the message.11 Timely and accurate information that is complete and free from errors becomes important during these events. People may disregard a message or consider the source(s) to be non-credible if individuals come to learn or suspect that they are not receiving the truth in its entirety.12 Transparency and honesty regarding a hazard enhances the perception of accuracy.13 As noted in a study published in Information Communication & Society, facts relating to the hazard need to be stated “authoritatively, confidently, and with certainty, even in circumstances in which there is ambiguity about message content factors and especially about the protective action the public is being asked to take.”14 These messages also explain that, even though physical details about the hazard are changing, experts agree on the protective actions people should take.15 Messages also need to be externally consistent, for example, by explaining any changes that may have occurred since the previous message, as well as be internally consistent.16
B. DeMarchi, and C. Pelanda, eds.), Milano, Italy: Franco Angeli; D. Mileti and E. Beck, 1975, Communication in crisis: Explaining evacuation symbolically, Communication Research 2:24-29; and B. McLuckie, 1975, Warning: A Call to Action, Washington, DC: U.S. Weather Service.
11 M. Lindell and R. Perry, 1987, Warning mechanisms in emergency response systems, International Journal of Mass Emergencies and Disasters 5(2):137-153; G. Rogers, 1985, Human Components of Emergency Warning, Pittsburgh, PA: Center for Social and Urban Research, University of Pittsburgh; P. Houts, M. Lindell, T. Weittu, P. Clearly, G. Tokuhata, and C. Flynn, 1984, The Protective Action Decision Model applied to evacuation during the Three-Mile Island crisis, International Journal of Mass Emergency and Disasters 2(1):27-39; and R. Perry, M. Lindell, and M. Greene, 1981, Evacuation Planning in Emergency, Lexington, MA: Lexington Books.
12 D. Mileti, 2012, Chapter 26, Public Response to Flood Warnings, in Coping with Floods (G. Rossi, N.B. Harmanciogammalu, and V. Yevjevich, eds.), NATO ASI Series, Series E: Applied Sciences, Vol. 257. Dordrecht, The Netherlands: Kluwer Academic Publishers.
13 D. Mileti, T. Drabek, and J. Haas, 1975, Human Systems in Extreme Environments: A Sociological Perspective, Boulder, CO: Institute of Behavioral Science, University of Colorado.
14 J. Sutton, E.S. Spiro, B. Johnson, S. Fitzhugh, B. Gibson, and C.T. Butts, 2014, Warning tweets: Serial transmission of messages during the warning phase of a disaster event, Information, Communication & Society 17:6, 765-787, doi:10.1080/1369118X.2013.862561.
15 D. Mileti and P. O’Brien, 1992, Warnings during disaster: Normalizing communicated risk, Social Problems 39(1):40-57; R. Perry, M. Lindell, and M. Greene, 1982, Crisis communications: Ethnic differentials in interpreting and acting on disaster warnings, Social Behavior and Personality 10(1):97-104; and R. Turner, J. Nigg, D. Paz, and B. Young, 1979, Earthquake Threat: The Human Response in Southern California, Los Angeles, CA: Institute for Social Science Research, University of California.
16 J. Nigg, 1987, Communication and behavior: Organizational and individual response to warnings, pp. 103-117 in Sociology of Disasters: Contributions of Sociology to Disaster Research (R.R. Dynes, B. DeMarchi, and C. Pelanda, eds.), Milano, Italy: Franco Angeli.; A. Chiu, L.E Escalante, J.K. Mitchell, D.C. Perry, and T.A. Schroeder, 1983, Hurricane Iwa, Hawaii, November 23, 1982, Washington, DC: National Academy of Sciences; R. Perry, 1983, Population evacu-
Recent research indicates that the length of a message plays a critical role in influencing people’s understanding, belief, decision making, risk personalization, and the amount of time they delay initiating a protective action.17
Individuals may have preconceived ideas about particular hazards that affect the type of message they are willing to receive and forward
ation in volcanic eruptions, floods, and nuclear power plant accidents: Some elementary comparisons, Journal of Community Psychology 11(1):36-47; and C. Flynn, 1979, Three Mile Island Telephone Survey: Preliminary Report on Procedures and Findings, Tempe, AZ: Mountain West Research.
17 D. Glik, K. Harrison, M. Davoudi, and D. Riopelle, 2004, Public perceptions and risk communication for botulism, Biosecurity and Bioterrorism: Biodefense Strategy, Practice, and Science 2(3):216-223; H. Bean, B.F. Liu, S. Madden, J. Sutton, M.M. Wood, and D.S. Mileti, 2016, Disaster warnings in your pocket: How audiences interpret mobile alerts for an unfamiliar hazard, Journal of Contingencies and Crisis Management 24(3): 136-147.
18 R. Perry and M. Greene, 1982, The role of ethnicity in the emergency decision-making process, Sociological Inquiry 52(Fall):309-34; G. Rogers, 1985, Human Components of Emergency Warning, Pittsburg, PA: Center for Social and Urban Research, University of Pittsburg; T. Saarinen and J. Sell, 1985, Warning and Response to the Mount St. Helens Eruption, Albany, NY: State University of New York Press.
19 S. Cutter, 1987, Airborne toxic releases: Are communities prepared? Environment 29(6):12-17, 28-31; R. Perry, 1987, Disaster Preparedness and Response Among Minority Citizens, pp. 135-151 in Sociology of Disasters: Contributions of Sociology to Disaster Research (R.R. Dynes, B. DeMarchi, and C. Pelanda, eds.), Milano, Italy: Franco Angeli.; R. Stallings, 1984, Evacuation behavior at Three Mile Island, International Journal of Mass Emergencies and Disasters 2:11-26.
20 M. Lindell and R. Perry, 1987, Warning mechanisms in emergency response systems, International Journal of Mass Emergencies and Disasters 5(2):137-153; R. Perry and M. Lindell, 1986, Twentieth-century Volcanicity at Mt. St. Hellens: The Routinization of Life Near and Active Volcano, Tempe, AZ: School of Public Affairs, Arizona State University; R. Perry and M. Greene, 1982, The role of ethnicity in the emergency decision-making process, Sociological Inquiry 52(Fall):309-34.
21 S. Cutter, 1987, Airborne toxic releases: Are communities prepared? Environment 29(6):12-17, 28-31; J. Gray, 1981, Characteristic patterns of and variations in community response to acute chemical emergencies, Journal of Hazardous Materials 4:357-365; R. Perry, M. Lindell, and M. Greene, 1981, Evacuation Planning in Emergency Management, Lexington, MA: Lexington Books.
to their networks. A study took subjects and had them look at a series of WEA and Twitter22 messages regarding five different types of disasters (flood, blizzard, hurricane, gas leak, and tornado). In the study, participants were more predisposed to share WEA messages or disaster tweets on Twitter expressing a dismissive sentiment about floods more than the other types of hazards, although overall, all subjects were highly responsive to the disaster messages and shared them a majority of the time.23 The study also found that subjects have different responses to different hazard types based on their perceived amount of danger or damage associated with that disaster, for example, on a psychological level,24 subjects perceive the threat posed by a flash flood differently than the other hazards in the study both while reading alerts about floods and when they were about to watch a video concerning floods.25
Context Characteristics Influencing Protective Action Initiation Times
Research shows that there are environmental and social cues that influence how people interpret alerts and warnings, which in turn, influences their PAI times. Environmental cues are indicators in one’s environment that reinforce the presence of the hazard. People are more likely to conclude the need to take a protective action stated in an alert or warning message if environmental cues are present that reinforce the presence of the hazard.26 Similarly, social cues are indicators in one’s social environ-
22 Twitter has recently increased the character limit for tweets from 140 characters to 280 characters for all users, with the exception of those tweeting in Chinese, Japanese, and Korean languages (A. Heath, 2017, “Twitter is turning on longer 280-character tweets for everyone,” Business Insider, http://www.businessinsider.com/twitter-280-character-tweets-for-everyone-2017-11).
23 C.D. Corley, Pacific Northwest National Laboratory, “Cognitive Modeling of the Impact of Wireless Emergency Alerts,” presentation to the committee on September 1, 2016.
24 The study collected 20-channel electroencephalography data in order to evaluate perception and responses.
25 C.D. Corley, Pacific Northwest National Laboratory, “Cognitive Modeling of the Impact of Wireless Emergency Alerts,” presentation to the committee on September 1, 2016.
26 J. Averill, D. Mileti, R. Peacock, E. Kuligowski, N. Groner, G. Proulx, and H. Nelson, 2005, Predicting Evacuation Delay in the World Trade Center: Occupant Behavior, Egress, and Emergency Communications: Federal Building and Fire Safety Investigation of the World Trade Center Disaster, NIST NCSTAR 1-7. Washington, DC: U.S. Department of Commerce, National Institute of Standards and Technology; C. Flynn, 1979, Three Mile Island Telephone Survey: Preliminary Report on Procedures and Findings, Tempe, AZ: Mountain West Research; M. Lindell and R. Perry, 2012, The Protective Action Decision Model: Theoretical modifications and additional evidence, Risk Analysis 32(4):616-632; R. Mack and G. Baker, 1961, The Occasion Instant: The Structure of Social Responses to Repeated Air Raid Warnings, Disaster Study No. 15, Washington, DC: National Research Council, National Academy of Sciences; G. Rogers and J. Nehnevajsa, 1987, Warning human populations of technological hazards, pp. 357-362 in
ment that reinforce the presence of the hazard and the need to take the protective action recommended in an alert or warning. People are more likely to initiate a protective action when social cues are present.27 These cues include seeing and hearing about others who are taking the recommended protective action.
How much time people have before a hazard strikes and the expected impact intensity also play a role in PAI times. Research has found that the amount of time people have to initiate an action before an event strongly influences PAI.28 As that time decreases, so does the PAI time. In hurricanes, people are quicker in taking protective actions as the impact time draws closer.29 Additionally, in hurricanes, more people decide to evacuate when strong storms are forecasted (impact intensity). People respond faster to hazards that are “dreaded” than ones that are “known.” For example, chemical accidents and nuclear power elicit a rapid response to avoid contamination.30 Hazards that are perceived as posing a serious threat elicit faster response.31
Research has also found that PAI times may be affected by the time of day and an individual’s location and activity when a message is received. However, there are currently no empirical studies on community-wide alert and warning events that establish or dismiss that initiation of a protective action, for example, evacuation, would take more time during the nighttime compared to daytime. Nor is there appropriate documentation to identify how a person’s location and activity impact protective initiation time. This includes if a person is sleeping, working, shopping, traveling, or engaged in recreation.32
ANS Topical Meeting on Radiological Accidents: Perspectives and Emergency Planning, Oak Ridge, TN: Oak Ridge National Laboratory.
27 S. Cutter, 1987, Airborne toxic releases: Are communities prepared? Environment 29(6):12-17, 28-31; and R. Dynes and E. Quarantelli, 1976, The family and community context of individual reactions to disaster, pp. 231-245 in Emergency and Disaster Management: A Mental Health Sourcebook (H. Parad, H. Resnik, and L. Parad, eds.), Bowie, MD: Charles Press.
28 M. Lindell and R. Perry, 1992, Behavioral Foundations of Community Emergency Planning, Washington DC: Hemisphere Press.
29 E.J. Baker, 1987, Warning and Evacuation in Hurricanes Elena and Kate, Tallahassee, FL: Department of Geography, Florida State University.
30 M. Lindell and R. Perry, 1992, Behavioral Foundations of Community Emergency Planning, Washington DC: Hemisphere Press.
31 D. Sorensen and D. Mileti, 1987, Decision making uncertainties in emergency warning system organizations, International Journal of Mass Emergencies and Disasters 5(1):33-61.
32 E. Baker, 1979, Predicting response to hurricane warnings: A reanalysis of data from four studies, Mass Emergencies 4:9-24; R. Clifford, 1956, The Rio Grande Flood: A Comparative Study of Border Communities, Disaster Study No. 17, Washington, DC: National Research Council, National Academy of Sciences; P. Houts, M. Lindell, T. Weittu, P. Clearly, G. Tokuhata, and C. Flynn, 1984, The Protective Action Decision Model applied to evacuation during the Three-Mile Island crisis, International Journal of Mass Emergency and Disasters 2(1):27-39;
Message Receiver Characteristics Influencing Protective Action Initiation Times
Over time, research shed more light on the human factors—such as age, gender, ethnicity, race, disabilities, and socioeconomic status—that influence response to warning messages. Early studies determined that status and role characteristics of individuals receiving warning messages influence protective action initiation time. It has also been found that those who are younger,33 have attained higher levels of education,34 and are employed,35 in addition to women as compared to men,36 are more likely to interpret alert and warning information better and take appropriate protective actions.
A number of recent studies have yielded additional insights building on the decades of earlier research on disaster response.
J. Nehnevajsa, 1985, Western Pennsylvania: Some Issues in Warning the Population Under Emergency Conditions, Pittsburgh, PA: University Center for Social and Urban Research, University of Pittsburgh; G. Rogers and J. Sorensen, 1991, Diffusion of emergency warning: Comparing empirical and simulation results, Risk Analysis 11:117-134.
33 S. Cutter and K. Barnes, 1982, Evacuation behavior and Three Mile Island, Disasters 6(2):116-124; H. Friedsam, 1962, Older persons in disaster, pp. 151-184 in Man and Society in Disaster (G. Baker and D. Chapman, eds.), New York, NY: Basic Books; R. Mack and G. Baker, 1961, The Occasion Instant: The Structure of Social Responses to Repeated Air Raid Warnings, Disaster Study No. 15, Washington, DC: National Research Council, National Academy of Sciences; R. Perry, M. Lindell, and M. Greene, 1981, Evacuation Planning in Emergency Management, Lexington, MA: Lexington Books; B. Phillips and B. Morrow, 2007, Social science research needs: Focus on vulnerable populations, forecasting, and warnings, Natural Hazards Review 8(3):61-68.
34 T. Drabek, 1986, Human System Responses to Disaster: An Inventory of Sociological Findings, New York, NY: Springer Verlag; D. Mileti, and C. Fitzpatrick, 1993, The Great Earthquake Experiment: Risk Communication and Public Action, San Francisco, CA: Westview Press; R. Turner, J. Nigg, D. Paz, and B. Young, 1979, Earthquake Threat: The Human Response in Southern California, Los Angeles, CA: Institute for Social Science Research, University of California.
35 C. Flynn, 1979, Three Mile Island Telephone Survey: Preliminary Report on Procedures and Findings, Tempe, AZ: Mountain West Research; R. Lachman, M. Tatsuoka, and W. Bonk, 1961, Human behavior during the tsunami of May, 1960, Science 133:1405-1409; R. Perry, 1987, Disaster preparedness and response among minority citizens, Sociology of Disasters 135-151; R. Stallings, 1984, Evacuation behavior at Three Mile Island, International Journal of Mass Emergencies and Disasters 2:11-26; Y. Yamamoto and E. Quarantelli, 1982, Inventory of the Japanese Disaster Literature in the Social and Behavioral Sciences, Columbus, OH: Disaster Research Center, Ohio State University.
36 T. Drabek, 1986, Human System Responses to Disaster: An Inventory of Sociological Findings, New York, NY: Springer Verlag; D. Mileti, and C. Fitzpatrick, 1993, The Great Earthquake Experiment: Risk Communication and Public Action, San Francisco, CA: Westview Press; and R. Turner, J. Nigg, D. Paz, and B. Young, 1979, Earthquake Threat: The Human Response in Southern California, Los Angeles, CA: Institute for Social Science Research, University of California.
Communicating Time Until Impact
WEA messages contain several elements—hazard, location, source, guidance, and time until impact. The Study of Terrorism and Responses to Terrorism (START) research team37 found that both guidance (what to do and how to do it) and time until impact (how much time people have to take the recommended action) play major roles relative to other message elements in the outcomes of public understanding and belief of the protective action recommendation and the ability to decide how to respond. Importantly, the START research team found that WEA messages would be more effective if they were to state how much time remains until impact rather than use time to indicate when the message expires, as is the current practice.
WEA messages are designed to provide alerts about imminent threats; the START research team conceptualized imminent as occurring within one hour. Other research has extensively examined the optimal timing of warnings for other alerting systems. For example, research on tornados finds that the optimal lead for issuing a tornado warning is from 15 minutes to just over 30 minutes.38 If too much lead time is provided, people are less likely to follow the protective guidance in a timely manner. Therefore, understanding how to best express lead time in WEA messages and the ideal lead times by hazard type are important areas for future research.
Including Protective Guidance in Web Links
Prior research called for including URLs in WEA messages to provide access to more complete information on the hazard and recommended protective guidance.39 Christopher McIntosh from Esri, a geographical information system firm, observed in his testimony to the committee that “without context, alerts are just noise.”
At this point, it is unclear what information is best included in a WEA message and what information is best included in Web pages the mes-
37 M. Wood, H. Bean, B. Liu, and M. Boyd, 2015, Comprehensive Testing of Imminent Threat Public Messages for Mobile Devices: Final Report, College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism.
38 S. Hoekstra, R. Butterworth, K. Klockow, D.J. Drotzge, and S. Erickson, 2011, A social perspective of warn on forecast: Ideal tornado warning lead time and the general public’s perceptions of weather risks, Weather, Climate & Society 3(1):128-140; and K.M. Simmons and D. Sutter, 2009, False alarms, tornado warnings, and tornado casualties, Weather, Climate & Society 1(1):38–53.
39 M. Wood, H. Bean, B. Liu, and M. Boyd, 2015, Comprehensive Testing of Imminent Threat Public Messages for Mobile Devices: Final Report, College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism.
sage links to. Prior research on WEA did not examine 360-character messages because the research was conducted before the Federal Communications Commission rulemaking that extended WEA messages from 90 to 360 characters.40 Furthermore, concerns remain as network congestion could be caused by people accessing an included link within seconds of receiving a WEA that includes a URL.41 Interestingly, some research found that message recipients are unlikely to open linked content and that instead they read only a few words of WEA-like messages owing to stress responses.42 Therefore, research is needed to understand how to most effectively craft 360-character messages as well as under what circumstances and what message content should be included in linked media. Also unknown is how to craft WEAs so that they galvanize people to read the entire message, including potentially life-saving linked content. Research is needed on how to best convey protective action guidance in 360-character messages vs. linked media. A consistent finding across research on WEA messages is that the public needs education on what the WEA service is as well as what protective actions to take during a variety of hazards.
More precise geotargeting that leverages the information that smartphones have about one’s location could be used to deliver more accurate and relevant alerts. Research has shed some light on the possible impacts of geotargeting on PAI time.
Some research has looked at the impact of reducing the size of the zone receiving an alert.43 It found that individuals receiving these more targeted messages would nevertheless forward these messages to individuals outside the zone. In some cases, this may decrease PAI time because people receive an alert from additional sources. In other cases, people may end up receiving forwarded messages intended for different zones that may call for the wrong protective action. For example, someone
40 Federal Communications Commission, “FCC Strengthens Wireless Emergency Alerts as a Public Safety Tool,” release date September 29, 2016, https://apps.fcc.gov/edocs_public/attachmatch/DOC-341504A1.pdf.
41 Federal Communications Commission, “Improving Wireless Emergency Alerts and community-initiated alerting,” release date November 19, 2015, https://apps.fcc.gov/edocs_public/attachmatch/FCC-15-154A1.pdf.
42 D. Glik, University of California, Los Angeles, “WEA Messages: Impact on Physiological, Emotional, Cognitive and Behavioral Responses,” presentation to committee on September 1, 2016.
43 A. Parker, RAND Corporation, “Exploring the Effect of the Diffusion of Geo-Targeted Emergency Alerts,” presentation to the committee on September 1, 2016.
intended to get a message to shelter in place may receive a forwarded message to evacuate.
Other research has looked at the location-proxy fallacy—that only those in the alert zone area would be interested in received messages for those zones.44 In fact, alerts may be of value to not only people in the alert zone at a specific time but also those who are contemplating entering or frequently enter that zone.45 Furthermore, individuals outside the geo-targeted area may share information with those at risk, adding credibility to the alert.
Constructing geotargeting algorithms that use cell site propagation footprints has been found to be the best method for sending out these messages as it can result in much smaller areas regardless of the physical location of cell towers and improve granularity, allows for monthly tests of the system without impacting the general public, enhances the reachability to people in harm’s way, and requires no change to the current WEA network.46 Another method of distributing these messages is a mechanism called Arbitrary-Size Location-Aware Targeting (ASLAT). Using this technique does not consume excessive mobile device power or radio resources, but increases geotargeting accuracy by utilizing the location awareness of mobile devices. It also increases delivery time as the phone will know its location before processing a received alert and will be able to determine if a certain message needs to get to the person or not. Implementation of ASLAT would require some changes to existing WEA standards for specific functionality in the cellular networks and mobile device behaviors.
Additionally, the use of geotargeting could address the issue of over-alerting. In one study, two different methods were used to distribute WEA messages in order to calculate geotargeting performance (GTP) estimates under two imminent threat scenarios: tornado warnings and earthquake early warning. In the first method, only cell towers within the warning area were directed to broadcast WEA messages, while in the second method WEA messages were broadcasted by cell towers within the warning area as well as towers adjacent to the warning area. It was found that over-alerting rates were lower when using the first method, but that if using alert failure rate (AFR) as the primary metric of GTP, the second method provided superior GTP in urban and mixed areas.47 The
44 B. Iannucci, Carnegie Melon University, “Opportunities, Options, and Enhancements for the Wireless Emergency Alerting Service,” presentation to the committee on September 1, 2016.
45 A. Parker, RAND Corporation, “Exploring the Effect of the Diffusion of Geo-Targeted Emergency Alerts,” presentation to the committee on September 1, 2016.
46 D. Ung, TeleCommunication Systems, Inc., “Geo-Targeting Method Using Cell Radio Frequency (RF) Propagation,” presentation to the committee on September 1, 2016.
47 D. Gonzales, 2016, Geo-Targeting Performance of Wireless Emergency Alerts in Imminent Threat Scenarios – Volume 1: Tornado Warnings, Washington, DC: Department of Homeland Security.
study noted that for tornados, the warning polygon is fixed for hours, although the path of the tornado may change. Initiatives like Threats In Motion (TIM) may be of value, since the tornado warning geotargeting would be improved by updating the portion of the warning polygon more rapidly. However, it is acknowledged that trying to integrate this system into WEA may present many challenges, including the transmission of more WEA messages and testing would be needed to ensure that WEA could handle TIM-based tornado warnings.48
Message Delivery Method
How an individual receives an alert and warning message can influence an individual’s perception on the risk and threat and therefore affects their PAI time. Earlier research on alert and warnings messages could not, of course, have factored in the opportunities and challenges presented by the recent dramatic changes in how people receive information. There are now a variety of new ways that an alert and warning message can reach an individual, including WEA messages, social media, phone applications, and online messaging among friends and family. Together with traditional media, such as television and radio, these constitute a complex, evolving ecosystem with many interacting systems.
There are unexplored opportunities to utilize social media as complementary channels for emergency alerts49—including uses related to both incoming and outgoing information. This aligns with a “ubiquitous alerting” strategy50 that includes all available channels and devices.
In the United States, 69 percent of adults use some type of social media.51 Social media are widely used during disaster events by emergency responders, people in the affected community, and global onlookers52
49 Communications, Security, Reliability and Interoperability Council V (CSRIC V), Working Group 2, Emergency Alerting Platforms, 2016, Social Media & Complementary Alerting Methods – Recommended Strategies & Best Practices: Final Report & Recommendations, Washington, DC: Federal Communications Commission.
50 R. Wimberly, “New Age of Alerting Coming: Ubiquitous Alerts,” release date February 9, 2016, http://www.emergencymgmt.com/emergency-blogs/alerts/new-age-of-alerting-coming--ubiquitous-alerts.html.
51 Pew Research Center, “Social Media Fact Sheet,” release date January 12, 2017, http://www.pewinternet.org/fact-sheet/social-media/.
52 A.L. Hughes and L. Palen, 2009, Twitter adoption and use in mass convergence and emergency events, International Journal of Emergency Management 6(3-4):248-260.
who converge there to seek and share information.53 While social media platforms can be used to gather and disseminate information, it can also be leveraged to help organize response efforts54 and share messages of support.55 Researchers have noted the potential for these platforms to contribute (as an incoming information source) to enhanced situational awareness of emergency responders and affected community members56—by aggregating information from distributed users who may have access to different perspectives of the disaster events. There is also opportunity for these platforms to be used as real-time communication tools for official responders to distribute messages to their various publics—and indeed many alert originators (including several regional offices within the National Weather Service) are already utilizing social media as part of their alerting activities.57
However, there are also many challenges related to the use of social media in the crisis context. For those monitoring social media during disaster events for safety-critical and/or actionable information, it can be difficult to identify the signal from within the noise, due to the volume of information shared. Significant for conversations about social media as an emergency alerting platform, emergency responders face several concerns regarding use of social media:
53 L. Palen and S.B. Liu, 2007, Citizen communications in crisis: Anticipating a future of ICT-supported public participation, pp. 727-736 in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, https://dl.acm.org; J.N. Sutton, L. Palen, and I. Shklovski, 2008, Backchannels on the front lines: Emergent uses of social media in the 2007 Southern California wildfires, pp. 624-632 in Proceedings of the 5th International ISCRAM Conference, http://www.iscramlive.org/portal/node/2236.
54 K. Starbird and L. Palen, 2011, Voluntweeters: Self-organizing by digital volunteers in times of crisis, pp. 1071-1080 in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, https://dl.acm.org; H. Gao, G. Barbier, and R. Goolsby, 2011, Harnessing the crowdsourcing power of social media for disaster relief, IEEE Intelligent Systems 26(3):10-14; J.I. White and L. Palen, 2015, Expertise in the wired wild West, pp. 662-675 in Proceedings of the 18th ACM Conference on Computer Supported Cooperative Work & Social Computing, https://dl.acm.org.
56 S. Vieweg, A.L. Hughes, K. Starbird, and L. Palen, 2010, Microblogging during two natural hazards events: What Twitter may contribute to situational awareness, pp. 1079-1088 in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, https://dl.acm.org.
57 Communications, Security, Reliability and Interoperability Council V (CSRIC V), Working Group 2, Emergency Alerting Platforms, 2016, Social Media & Complementary Alerting Methods – Recommended Strategies & Best Practices: Final Report & Recommendations, Washington, DC: Federal Communications Commission.
- They are challenged to keep up with the volume and to find relevant and actionable information.
- Social media platforms have opened up two-way channels of communication with the public. Members of the public expect to be heard on social media—i.e., for emergency officials to respond to what members of the public share on social media.58
- Though several early adopters have had some success with utilizing these platforms, the “rules for engagement” and best practices for emergency responders using social media are still evolving.59
- Social media are not geographic-specific and therefore it is difficult for responders to distinguish between their local community members and the global audience.
- Emergency responders fear the spread of misinformation on these platforms, and many have reported a reluctance to adopt social media in part due to this concern.60,61
Another challenge, which suggests the need for more research, is the varied use of different social media platforms across different demographics, a problem that could create new “digital divides” in the accessibility of information if alerts are shared via social media.62 This latter point suggests an “all channels” strategy, which positions social media in general as a complementary alerting source and not a primary one, and points toward the use of many different platforms (e.g., Twitter, Facebook, and Snapchat) at once.
Until recently, little was known about the relative effectiveness in prompting protective action or complementarity of alerts distributed through WEA or social media platforms, such as Twitter, which has been used by emergency managers since its debut in 2009.63 In one study,
58 S. Bernier, CEM, CBCP, MBCI SB Crisis Consulting, 2013, “Social Media and Disasters: Best Practices and Lessons Learned,” presentation at the Disaster Preparedness Summit, August 21, American Red Cross, Chicago, IL.
59 K. Starbird, D. Dailey, A.H. Walker, T.M. Leschine, R. Pavia, and A. Bostrom, 2015, Social media, public participation, and the 2010 BP Deepwater Horizon oil spill, Human and Ecological Risk Assessment: An International Journal 21(3):605-630.
60 S.R. Hiltz, J. Kushma, and L. Plotnick, 2014, Use of social media by U.S. public sector emergency managers: Barriers and wish lists, pp. 602-611 in Proceedings of the 11th International ISCRAM Conference, http://www.iscramlive.org/portal/node/2236.
61 A.L Hughes and L. Palen, 2012, The evolving role of the public information officer: An examination of social media in emergency management, Journal of Homeland Security and Emergency Management 9(1):1-20.
62 K. Crawford and M. Finn, 2015, The limits of crisis data: Analytical and ethical challenges of using social and mobile data to understand disasters, GeoJournal 80(4):491-502.
63 A.L. Hughes and L. Palen, 2009, Twitter adoption and use in mass convergence and emergency events, International Journal of Emergency Management 6(3-4):248-260.
which used mock WEA messages (90 characters) and mock Twitter-length messages (140 characters), the most common initial reaction was confusion and fear accompanied by a desire to acquire more information about the apparent threat.64 The lack of clarity about the hazard, the protective action guidance and how long they had to complete it, the time of the incident, and the affected area led to confusion and frustration. Because of the limited information and lack of personalization in these messages, most participants in the study could not tell whether the hazard would impact them specifically.65 Consistent with public warning response models,66 participants would have sought out additional information before taking any protective action.
Like most technology, social media platforms are continuously evolving, as are the practices related to their use, both generally and in the context of emergency alerting and response. Thus, there is a need to continue research in this area to better understand changing practices in order to more fully realize the acknowledged potential of these platforms. Specifically, there are unexplored opportunities to utilizing social media as complementary channels for emergency alerts—including uses related to both incoming and outgoing information. This is especially important since the research record provides strong evidence that how warning message(s) are delivered to the public influences public warning message response (referred to as warning or message delivery factors). Message delivery factors can help people confirm the warning message and personalize risk, for example, which are both important intervening factors between getting a message and taking a recommended protective action. It has been found that messages frequently repeated work best to decrease PAI time, and messages delivered over multiple and different communication channels work best in comparison to those delivered over fewer or single channels. In the context of social media, familiar accounts (e.g., friends, family, and celebrities), have the ability to share emergency alert and warning messages from official accounts, which adds to the credibility of the message. However, this can also lend itself to adding credibility to incorrect or misleading information.
64 H. Bean, B.F. Liu, S. Madden, J. Sutton, M.M. Wood, and D.S. Mileti, 2016, Disaster warnings in your pocket: How audiences interpret mobile alerts for an unfamiliar hazard, Journal of Contingencies and Crisis Management 24(3):136-147.
66 D. Mileti and J. Sorensen, 1990, Communication of Emergency Public Warnings: A Social Science Perspective and State-of-the-Art Assessment, Oak Ridge, TN: Oak Ridge National Laboratory, U.S. Department of Energy; R. Perry, 1979, Evacuation decision-making in natural disasters, Mass Emergencies 4:25-38.
Crowdsourcing tools, which include social media platforms such as Twitter and navigation services such as Waze,67 could leverage the capacity of the online crowd to help process and distribute emergency messages. While emergency messages disseminated through crowdsourcing technologies may not be more accurate than government emergency alerts themselves, they may be a trusted source because they are used on a day-to-day basis. Such has been the experience of Humanity Road, a virtual volunteer organization that assists with humanitarian response and disaster preparedness education. Humanity Road takes information from a variety of sources,68 analyzes it, and standardizes it for dissemination on Twitter. It developed what it calls Twitter Commandments (Box 1.3) that reflect lessons learned on how to most effectively collect, analyze, standardize, and disseminate disaster information.69
The crowdsourced navigation service Waze collects data, including traffic conditions and road hazards from its users, to generate optimal routes. To combat misinformation, user reports have to be validated by other users. Apps like Waze can be used during events to collect and distribute information about hazards such as road closures, floods, and accidents. Moreover, this information can be supplemented by alert originators who can transmit updated information on hazards or suggest safe travel routes as information is gathered throughout an event. In Texas, the Center for Collaborative Adaptive Sensing of the Atmosphere (CASA) Dallas Fort Worth Living Lab has already begun to experiment with how to provide information about urban flooding hazards.70 The CASA Alerts Application introduced personalization and allowed users to control thresholds and preferences, which determined which alerts reach the user. Users, in turn, can provide feedback by verifying data related to the event, providing information about their experience during the event,
67 While Waze relies heavily on volunteer information, reported incidences are validated through other users. Additionally, during a disaster much of its information (e.g., shelters) is confirmed by official sources.
68 Sources used by Humanity Road include government sources, although it is noted that there is an inherent distrust of the government among a large portion of the population.
69 K. Starbird and L. Palen, 2013, “Working & Sustaining the Virtual ‘Disaster Desk’,” University of Washington, http://faculty.washington.edu/kstarbi/cscw2013_final-2.pdf.
70 The CASA Dallas Forth Worth Living Lab brings together the NCTCOG, NWS, DFW Airport, Fort Worth, Grand Prairei, Midlothian, and other public safety officials, in addition to bringing in the user. They do experiments in the community rather than doing experiments in a lab setting (E.J. Bass, B. Hogan, D. Rude, C. League, P. Marsh, L. Lemon, B. Philips, D. Westbrook, J. Brotzge, and R. Riley, 2011, “A method for investigating real-time distributed weather forecaster-emergency manager interaction,” pp. 2809-2815 in 2011 IEEE International Conference on Systems, Man, and Cybernetics, doi: 10.1109/ICSMC.2011.6084098.).
and providing information on what action, if any, they took during the event. By getting feedback from users, more information can be gathered relating to what areas are being impacted and how, and gives researchers the ability to gather information on how many individuals took the recommended protective action(s).
Message Receiver Characteristics
Additional factors that could influence an individual’s PAI time are an individual’s characteristics. These characteristics include an individual’s age, gender, ethnicity, socioeconomic status, race, and disability status.
A recent study looking at the effects of weather on activities by the elderly using the University of Michigan-Thomson Reuters Surveys of Consumers helps shed some light on the vulnerability of the elderly in adverse conditions. This study found that icy conditions were the hazard type that most caused participation restriction (where people took the
protective action of not going outside or engaging in certain activities) in day-to-day life,71 as compared to cold temperatures, snow, rain, fog, and wind.72 Adults 65 or older reported greater difficulty leaving the home in icy conditions, especially in cases where these adults had mobility impairment or self-imposed driving restrictions.73 Other work has found that, owing to higher rates of chronic illnesses, older persons are more susceptible to the adverse effects of psychological and physical stress, including disasters.74 Although there are resources and guidelines for older adults, those with mobility impairment continue to have difficulties both during and after a disaster, and the older an individual is, the less they are prepared during and after a disaster. Another study, which sampled adults aged 50 and above from the Health and Retirement Study, found that two-thirds of the study population lacked an emergency plan, had never participated in a preparedness program, and were not aware of the resources available to them.75 In addition, 15 percent of respondents reported having medical equipment that needed electricity, making them susceptible to power outages.76 An additional problem with older adults is that social isolation may prevent them from receiving warning messages or asking for help when needed, which may render them invisible to rescue teams during times of disasters.77 The problem of social isolation is compounded by differences in mobile device ownership, where a small subset of elderly individuals do not own a cell phone or they own a device that is not a smartphone (which impacts the types of messages they can receive). In the United States, 97 percent of older adults between the ages of 50 and 64 and 80 percent of older adults aged 65 and above were found to own a mobile device. In the group between the ages of 50 and 64, 74 percent owned smartphones while that percent dropped to 42 percent for those 65 or older.78 In an effort to mitigate the issue of social isolation, the Department of Health and Human Services has created the emPOWER map tool, which maps multiple populations with various
71 This included participation in grocery shopping, playing sports, and driving.
72 P.J. Clarke, T. Yan, F. Keusch, N.A. Gallagher, 2015, The impact of weather on mobility and participation in older U.S. adults, American Journal of Public Health 105(7):1489-1494.
74 T.M. Al-rousan, L.M. Rubenstein, and R.B. Wallace, 2014, Preparedness for natural disasters among older US adults: A nationwide survey, American Journal of Public Health 104(3):506–511.
77 D.P. Eisenman, K.M. Cordasco, S. Asch, J.F. Golden, and D. Glik, 2007, Disaster planning and risk communication with vulnerable communities: Lessons from Hurricane Katrina, American Journal of Public Health 97(Suppl 1):S109-S115.
kinds of vulnerabilities that could potentially be used to identify vulnerable, at-risk populations during disasters.79
Gender, Ethnicity, and Socioeconomic Status
Those perceived to be the most vulnerable during natural disasters are older adults, women, and minorities. While looking at perceptions and circumstances by gender and ethnicity, a study conducted in Rhode Island found that women were more likely to have children or rent a home, while being less likely to earn an income of $100,000 or greater as compared to men. Minority populations, in comparison to white men and women, often have nowhere to stay if a hurricane hits, and are less likely to own a car or have an income greater than $100,000. It was found that it is not always the case that women and minority populations act differently than men or white men and women, respectively, but what influences perceptions of hazard (as subject to race and gender effects) are a person’s financial resources, home/car ownership, and relationships (marriage or children).80 Additionally, resilience is hindered for those with limited social connectedness.81
This emphasis on relationships, specifically within minority communities, was seen following Hurricane Katrina. In Houston’s major evacuation center, 58 qualitative interviews were conducted with individuals—most of whom were from New Orleans, low income, and African American. The study concluded that overcoming shelter and transportation obstacles would not have been sufficient to significantly improve disaster response. More important were a person’s ties to family, friends, and the community, suggesting that disaster preparedness strategies emphasize better community-based communications.82
Language and Culture
The U.S. population is becoming increasingly racially and ethnically diverse; it is predicted that by 2055, the United States will not have a
80 D.M. West, and M. Orr, 2007, Race, gender, and communications in natural disasters, Policy Studies Journal 35(4):569–586.
81 D.P. Aldrich, 2014, Building Resilience: Social Capital in Post-Disaster Recovery, Chicago, IL: University of Chicago Press, August.
82 D.P. Eisenman, K.M. Cordasco, S. Asch, J.F. Golden, and D. Glik, 2007, Disaster planning and risk communication with vulnerable communities: Lessons from Hurricane Katrina, American Journal of Public Health 97(Suppl 1):S109-S115.
single racial or ethnic majority.83 The resulting linguistic and cultural diversity can present a challenge for emergency managers who may not have the resources to communicate effectively with these populations. For example, in New York City, there are 208 spoken languages, but emergency alert and warning messages are available only in 18 languages.84 This leaves a large portion of the population susceptible to not receiving or to misunderstanding WEA messages. A study conducted in three Gulf Coast counties in South Mississippi with highly diverse communities85 found that 47.3 percent of respondents indicated they wished to receive alert and warning messages in a language other than English. Although there are a variety of technologies (further discussed in Chapter 3) that are available to disseminate alert and warning messages, including WEA, TV, and radio, it was found that authorities had to depend on posters and pictures to disseminate information to Vietnamese and Hispanic communities.86 The respondents to the study indicated that they trust and are more likely to act on information received from their family and friends. The second trusted source of information was information received via TV, radio, and WEA. Respondents in the study were also more likely to trust sirens than social media.87
Many regions in the United States have large populations for whom English is not the primary language, and the number of different languages spoken in some areas can be very large. However, even as language diversity is increasing, the use of smartphones and other technologies (e.g., Internet, broadband, tablet, and social media), which provide valuable translation capabilities, is also growing (Table 1.1). Smartphone adoption has more than doubled since 2011, when the Pew Research Center started surveying the topic.88 With smartphones, individuals (assuming
83 D. Cohn and A. Caumont, “10 Demographic Trends That Are Shaping the U.S. and the World,” Pew Research Center, release date March 31, 2016, http://www.pewresearch.org/fact-tank/2016/03/31/10-demographic-trends-that-are-shaping-the-u-s-and-the-world/.
84 B.J. Krakauer, New York City Emergency Management, “Vision for the Future of Public Alerts and Warning: New York City’s Perspective,” presentation to the committee on August 9, 2016.
85 Hancock, Harrison, and Jackson counties in Mississippi. The Mississippi Gulf Coast is susceptible to natural disasters like tropical cyclones and coastal flooding, in addition to being a landfall location for Hurricane Katrina. These highly populated counties have culturally diverse communities, including Anglo-Americans, African Americans, and Vietnamese and Hispanic immigrants.
86 B. Kar, University of Southern Mississippi, “An Integrated Approach to Geo-Target At-Risk Communities and Deploy Effective Crisis Communication Approaches,” presentation to the committee on September 1, 2016.
88 A. Smith, “Record Shares of Americans Now Own Smartphones, Have Home Broadband,” Pew Research Center, release date January 12, 2017, http://www.pewresearch.org/fact-tank/2017/01/12/evolution-of-technology/.
TABLE 1.1 Cell Phone Ownership
|Owns Any Cell Phone||Owns Smartphone||Owns Cell Phone That Is Not a Smartphone|
|Less than $30,000||92%||64%||29%|
SOURCE: Pew Research Center, “Mobile Fact Sheet,” release date January 12, 2017, http://www.pewinternet.org/fact-sheet/mobile/.
they know how to use the phone and have a data plan or WiFi access) have many tools at their disposal. For example, Google Translate could be used to translate an alert into a person’s native language, and Facebook has an integrated translation capability. However, despite rapid improvements in recent years, such tools are imperfect and may in some cases cause further confusion. A particular challenge is that such tools do not account for differences in usage among different dialects.
Culture can also affect how people respond to various states of a disaster. For example, the October 17, 1989, Loma Prieta earthquake caused extensive damage in Watsonville, Santa Cruz, and Los Gatos, California. Families with experience with earthquakes in Mexico preferred to camp outdoors rather than stay in possibly damaged buildings, yet it took time for authorities to agree to open public parks as official shelters. Those camping in the parks were also motivated by a desire to stay close to their homes in order to protect personal possessions.89 The challenge of sheltering all the disaster victims was further compounded by the lack of preplanning for the situation, which stemmed from a lack of community participation during the preplanning despite prior research calling for such engagement.90 Another cultural issue arose regarding shelters when victims who were refugees from Central America found the tents and fences set up by the
89 B.D. Phillips, 1993, Cultural diversity in disasters: Sheltering, housing, and long term recovery, International Journal of Mass Emergency Disasters 11(1):99-110.
90 E.L. Quarentelli, K. Green, E. Ireland, S. McCabe, and D.M. Neal, 1983, Emergent Citizen Groups in Disaster Preparedness and Recovery Activities, Columbus, OH: Ohio State University.
American Red Cross and National Guard too reminiscent of concentration camps and government-backed death squads. These differences in perceptions highlight the importance of involving the community in the creation, planning, and execution of emergency practices.
In a study conducted to understand how disability factors into the response to WEA messages, it was shown that individuals familiar with WEA were more likely to take immediate action,91 but individuals with a disability are only half as likely to have heard of WEA. Like other populations, the overwhelming majority of those who participated in the study (98 percent) reported owning a mobile phone. As a result, these populations stand to benefit from using devices that provide appropriate attention-getting, sound, and display affordances—capabilities that can also help nondisabled populations.92
Physiological and Mental Models
When it comes to forwarding or sharing messages, the type of hazard and associated biases affect what individuals share. One study found that subjects (shown both WEA and Twitter messages) were more predisposed to share dismissive messages and tweets about floods as compared to blizzards, gas leaks, hurricanes, and tornados.93 When deciding to share a message, there seemed to be a more deliberate thought process, as the subjects in the study typically had a higher level of brain activity in the frontal lobe compared to when they chose not to share a message. The study showed that, overall, subjects shared disaster messages a majority of the time and were highly responsive to all types of disaster messages (WEA messages and tweets).94 However, the study also suggested that there are perceived differences in the threat or urgency posed by various disasters on a physiological level, which explains why the subjects reacted to floods differently than the other hazards.95 With certain hazards (such as floods) there seems to be an optimism bias, in which a person believes that the probability
91 H. Mitchell, “Optimizing Wireless Emergency Alerts for Sensory Disabilities,” presentation to the committee on September 1, 2016.
93 C.D. Corley, N.O. Hodas, R. Butner, J.J. Harrison, and C. Berka, 2016, “Modeling Cognitive Response to Wireless Emergency Alerts to Inform Emergency Response Interventions,” Pacific Northwest National Laboratory, https://www.dhs.gov/sites/default/files/publications/WEA%20-%20Modeling%20Cognitive%20Response.pdf.
of the event actually happening is low, that influences their decision of whether or not to engage in the recommended, protective actions.
Part of increasing public participation in protective actions is educating the public about alert and warning systems. Although the Federal Emergency Management Agency (FEMA) partnered with the Ad Council to create an education campaign about WEA, there seems to be little knowledge among the public about what WEA is, what its purpose is, or how and why alerts are issued.96 Indeed, participants from a variety of studies have questioned the validity of the emergency alerts they have received. In addition, some participants do not know the meaning of acronyms used in messages97 and in one study, participants asserted that information from local sources would be more believable than WEA alerts.98 Education of the public is just one of the avenues that may be required to enhance the effectiveness of alert and warning messages.
Alert originators also need to be knowledgeable about the capabilities and shortcomings of the WEA systems.99 This can include understanding WEA local area coverage, as cellular networks can vary from one region to another, especially in rural areas.100 In order to provide alert originators with information concerning the implementation and utilization of WEA, the Software Engineering Institute published papers on integration,101 best practices,102 and security.103 Alert originators also need to know how to write effective messages under the time crunch of emergencies.
96 H. Bean, B.F. Liu, S. Madden, J. Sutton, M.M. Wood, and D.S. Mileti, 2016, Disaster warnings in your pocket: How audiences interpret mobile alerts for an unfamiliar hazard, Journal of Contingencies and Crisis Management 24(3):136-147.
99 C. Woody, Software Engineering Institute of Carnegie Mellon University, “SEI Wireless Emergency Alerts (WEA) Research 2013 through 2016,” presentation to the committee on September 1, 2016.
100 D. Gonzales, 2016, Geo-Targeting Performance of Wireless Emergency Alerts in Imminent Threat Scenarios–Volume 1: Tornado Warnings, Washington, DC: Department of Homeland Security.
101 Carnegie Mellon University, 2014, Commercial Mobile Alert Service (CMAS) Alerting Pipeline Taxonomy, CMU/SEI-2013-TR-019, Pittsburgh, PA.
102 Carnegie Mellon University, 2013, Best Practices in Wireless Emergency Alerts, CMU/SEI-2013-SR-015, Pittsburgh, PA; Carnegie Mellon University, 2013, Wireless Emergency Alerts New York City Demonstration, CMU/SE I-2013-SR-tbd, Pittsburgh, PA; Carnegie Mellon University, 2014, Maximizing Trust in the Wireless Emergency Alerts (WEA) Service, CMU/SE I-2013-SR-027, Pittsburgh, PA; Carnegie Mellon University, 2014, Wireless Emergency Alerts: Trust Model Technical Report, CMU/SE I-2013-SR-021, Pittsburgh, PA.
103 Carnegie Mellon University, 2014, Wireless Emergency Alerts (WEA) Cybersecurity Risk Management Strategy for Alert Originators, CMU/SE I-2013-SR-018, Pittsburgh, PA; Carnegie
If new capabilities, such as more precise geotargeting are made available, alert originators will need to understand those capabilities and how to best use them. Alert originators can also learn from the experiences of nongovernmental organizations like Humanity Road and operators of commercial services like Waze that may provide valuable insights on how to use two-way communications to get real-time information about what is happening in their area. By understanding how various platforms work, including social media as previously discussed in this chapter, alert originators may be able to leverage them more effectively in times of need in order to disseminate alerts and warnings to the public.
Mellon University, 2013, Best Practices in Wireless Emergency Alerts, CMU/SEI-2013-SR-015, Pittsburgh, PA; Mapping WEA Security Requirements and Guidance to Cybersecurity Risk Mitigation Recommendations (delivered separately to DHS - some requirements are restricted); INCOSE Insight Essay, 2013, Evaluation of security risk for WEA alert originators using mission threads, Insight 16(2).