© PHOTOALTO AGENCY RF/GETTY IMAGES Communication Technology, Emergency Alerts, and Campus Safety GORDON A. GOW, TARA MCGEE, DAVID TOWNSEND, PETER ANDERSON, AND STANLEY VARNHAGEN Digital Object Identifier 10.1109/MTS.2009.932797 34 | 1932-4529/09/$25.00©2009IEEE IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 I n the wake of a report on the Virginia Tech shooting tragedy that criticized officials for being too slow to alert the campus community [1], universities across North America are now implementing mass notification systems, often using text messaging as a primary alerting channel for students, staff, and faculty [2], [3]. An effective campus alerting system, like other public warning systems, involves a complex sociotechnical assemblage that combines mass notification capabilities with extensive administrative, legal, training, and communications considerations [4], [5]. On the technical side, an effective campus alert system requires the integration of diverse information and communication technologies, including analog legacy systems (e.g., sirens and public address systems) and digital technologies such as email, Internet, and mobile phones. On the social side, campus alerting operates within the context of manifold policy and legal considerations, as well as specific, often complex, administrative and procedural requirements—all of which must be grounded in an adequate understanding of human responses to risks and warnings from the perspective of both the issuer and recipient of a warning message. There is concern within Canada, as in the United States, that the imperative to implement new mass notification capabilities on campus has overshadowed consideration of other factors necessary for well-integrated emergency messaging systems, especially with respect to policy, legal, and warning response factors (see, for example, Latimer [6]). Administrators and other members of the campus community may be working with a set of unexamined assumptions, leading to unrealistic expectations for these systems. For example, text messaging over mobile phones has become a popular choice among post-secondary institutions as a practical and cost effective solution for campus alerting, with a host of vendors now offering this service. However, a recent study by Georgia Institute of Technology demonstrated serious reliability issues with this method of mass notification, suggesting that it may be a poor choice when measured against a standards set for emergency alerting systems in the United States [7]. One foreseeable risk from unexamined assumptions and conjecture with regard to these systems is that their use could exacerbate an emergency incident by placing additional demands on first responders attempting to manage a situation. Congestion of telecommunications systems is a well-documented risk. Congestion problems result in part because mobile phone networks are not designed to handle the call volumes associated with significant national or localized emergency situations or even with “flash crowd” call attempts such as those experienced during major public events such as New Year’s Eve. A less clearly understood concern is the spread of rumors or deliberate misinformation campaigns by text message or email. Above all, the credibility of campus safety initiatives is also on the line when expectations are not carefully managed. This is often reflected in skeptical views about the use of text messaging among members of the campus community as characterized by this comment [8] posted online in response to a recent story on text message alerting: “These school alert systems are useless. We had a shooting alert and a few weather alerts like tornado warnings in the city area last year. When the messages came, most of the alerts already expired! They have to come up with a better way; for example, using digital message boards across campus or public computers, including those in the classrooms.” Although certain risks may be unavoidable, an integrated campus IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 alerting system, based on a clear understanding of the various sociotechnical factors involved, is an important step in reducing the impact of these kinds of risk and ensuring that the system remains credible and effective over the long term. To ward off unwarranted public criticism or avoidable legal action following an incident, it is vital to better align expectations with reality in order to mitigate potentially serious consequences for safety and security during crisis events. What follows is a presentation of several socio-technical factors implicated in campus alerting, initial observations based on a workshop held with campus emergency planners from across Canada, as well as highlights from a survey conducted with faculty, staff, and students at three Canadian universities. Workshop and Survey on Campus Alerting Researchers from the University of Alberta, University of New Brunswick, and Simon Fraser University formed the Campus Emergency Messaging Research Group (CEMRG) in the fall of 2007. In November 2007, CEMRG organized a Canadian National Roundtable on Campus Emergency Messaging that brought together diverse stakeholders to discuss campus alerting. Participants represented those who are responsible for implementing and managing campus alerting systems across the country. These included emergency planners from post-secondary educational institutions, government emergency managers, technical experts, and academic researchers. The agenda was organized around three elements: warning response, policy and legal issues, and new technologies. In March 2008, CEMRG conducted a survey among faculty, staff, and students at three Canadian universities. An invitation to complete an online survey was sent to a large sample of graduate | 35 and undergraduate students, as well as to all academic and nonacademic staff at the University of Alberta, Simon Fraser University, and University of New Brunswick. The survey consisted of 46 questions, only some of which are discussed here. The total number of responses across all three universities was 4529, with an overall average response rate of 16 percent. Questions dealt with access to communications technology while on campus, familiarity with campus security information and services, as well as views concerning campus alerting methods and procedures. Findings concerning the latter category of questions are discussed below within the context of a socio-technical framework. ties. Yet research findings from the social science literature suggests that human response to warnings is a complex social-psychological process, and that getting individuals to respond appropriately has remained an elusive goal even in settings where populations are familiar and experienced with a hazard. There are numerous documented cases where organizations and individuals have failed to understand the need to issue an alert message in the first place, whether out of fear or out of uncertainty concerning the anticipated response from the public [9], [10]. This amounts to a fundamental problem for those involved in campus alerting [11]: Warning Response as a Social Process Among the topics raised at the workshop was the fundamental expectation of those involved in campus alerting that anyone receiving an alert will understand it and then take appropriate selfprotective action, whether that action be to evacuate, to shelter in place, or to report suspicious activi- “… even in warning situations involving long lead times, clear evidence of a threat, and carefully executed warning dissemination plans, compliance with warning messages is by no means universal. … the first challenge any warning and alert system faces is simply achieving acceptable levels of warning compliance. Many systems Table I Expressed Preferences (Percentages) for Alerting Methods for Unexpected Severe Weather and Active Shooter Incidents. Unexpected severe weather (n= 4486) Telephone call (voice) UA* 12 SFU* 9 UNB* 11 SMS/Text message Email/instant messenger Campus/other radio or TV Siren/public address system Word of mouth (face-to-face) 12 34 4 35 3 35 43 4 6 3 8 67 9 3 2 Active shooter on campus (n= 4471) UA SFU UNB Telephone call (voice) 21 17 23 SMS/Text message 12 28 12 Email/instant messenger 17 11 14 Campus/other radio or TV 2 2 2 Siren/public address system 43 37 43 Word of mouth (face-to-face) 5 6 6 *Data shown are for University of Alberta (AB), Simon Fraser University (SFU), and University of New Brunswick (UNB). 36 | currently in operation fall far short of that goal.” The issue of warning compliance suggests that the effectiveness of campus alerting systems ought to be measured using both social and technical criteria. Whereas the technological elements of a campus alerting system may be capable of issuing campus-wide alerts within minutes, the question still remains as to how effective that system will be in prompting an appropriate warning response from the campus community. For example, fire alarms are a reasonably welldeveloped alerting technology, but there remains a widespread problem with getting people to take fire alarms seriously and respond accordingly. Along these lines, the term “warning compliance” was discussed at the workshop with several participants objecting to the term on the grounds that it was a simplistic conceptualization of the response phase during an emergency incident. Despite the apparent objections to this term, it did become clear that campus emergency managers are concerned with warning response on campus, and that little research is currently available to guide their planning decisions. The existing literature on warning response is concerned with community-wide emergencies with an emphasis on households as the unit of analysis. Moreover, previous studies have tended to focus on the use of traditional TV and radio broadcasting media for alerting [12]. By contrast, the warning response on a campus community may differ in certain respects, in part because of the mixed demographic characteristics, the diversity of spaces and activities that are found in these settings (e.g., classrooms, libraries, labs, phys ed facilities, and residences), the mobility of a campus population, and the growing IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 reliance on mobile phones, email, and other newer media as primary means of communications. Nonetheless, the existing literature on warning response does provide an important set of considerations to build upon. Moreover, emergency planners are probably justified in their concerns about warning compliance among members of the campus community. Social-demographic characteristics of warning recipients, situational factors, and the social context are known to influence warning response behavior [13]. For example, studies have found that women are more likely to believe warnings and take recommended protective action [14], [15]. Research has also discovered a “normalcy bias” during warning response that shows a tendency for certain individuals to ignore warning messages and continue to persist in their everyday activities, even in some cases where the risk to them may be evident [11]. On the other hand, Dow and Cutter [16] have contributed evidence to debunk the so-called “cry wolf” myth in finding that false alarms are not likely to decrease response rates if the reason for them is well communicated to recipients. Together, this range of factors presents a complex set of variables that may influence how members of a university campus will respond to a warning. In order to better understand these factors, it is important to analyze warning response as a socialpsychological process linked to issues such as trust, credibility, and risk perception [10]. Disaster sociologists have formalized this process using a multi-stage warning response model [17], [18]: 1) hearing the warning, 2) believing that it is credible, 3) confirming the warning, 4) personalizing the warning, 5) determining if protective action is necessary, 6) deciding if protective action is feasible, 7) determining what action to take (if any). This model is closely associated with factors that affect the clarity and credibility of the alert message itself, as well as various environmental and social factors that will influence an individual’s perception of risk. However, a major influence throughout each stage of the warning response process is receiving confirmation of the message from other sources. Campus emergency planners are well advised to take this factor into account because individuals will likely be very active in terms of seeking out additional information during a warning circumstance, and this social process is critically important in terms of fostering an appropriate warning response: “When warning information is received, most people try to verify what they heard by seeking out information in another warning message or from another warning source or person. … The single most important factor that influences public response to warnings is confirming risk information through interacting with others and searching for additional confirmatory information” [18, p. 3]. The clear implications for this response factor on a campus is that an alert will likely result in congestion of voice landline and wireless telecommunications networks, as staff and students attempt to contact one another for more information. Under these circumstances it may be difficult or impossible to issue subsequent alerts or updates to the campus community using the telecommunications or Internet infrastructure. Emergency planners will need to anticipate this type of situation and plan for alternative means of providing information such as sign boards, positioning of IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 campus security officials, and possibly prioritizing face-to-face contact for disseminating updated alert information. The survey conducted by CEMRG probed the relative importance of confirmation among members of the campus community by asking them about their preferred means of being alerted for a range of emergency incidents. Their choices included email, telephone (voice), text message, and broadcast media (radio/TV). In addition to these technological options, we also included the choice of “word of mouth (face-to-face).” Our initial expectation was that given this option, a significant number of respondents would prefer to be alerted through a direct encounter with a colleague or university official as this would correspond with claims about the importance of confirmation. The results in fact were quite different, with only six percent or less identifying this as a preferred method for all hazard types. (See Table I.) By contrast, most respondents preferred to be contacted by email or by telephone (the results for text messaging as a preferred method are described in more detail later). It is not quite clear yet what these findings suggest in terms of confirmation behavior. One possibility is that members of the campus community tend to associate “alerting” with a single act of communication while overlooking the follow-up process of confirming and corroborating information. If this is the case, it would then suggest that an integrated program for campus alerting might make a clear distinction between its “alerting” and “confirmation” components. Whereas the aim of alerting is to get the attention of the campus community and may require advanced technological systems to do so, confirmation is an intensely social process that is vital if people are going to take appropriate action following the alert. The confirmation stage may therefore require | 37 more interpersonal, possibly direct contact between campus officials and the community to reinforce the urgency of the incident and provide clear instructions as to the response actions being requested. Policy, Legal, and Planning Concerns Emergency communications have long been a concern to university officials, particularly with respect to alarm systems for fire, hazardous spills, and personal safety. With the passage of the Clery Act 20 U.S.C. §1092(f) (formerly, the Student Right to Know and Campus Security Act of 1990), most post-secondary institutions in the United States have also been required by law to provide “timely warning” of incidents that represent a threat to the safety of students or employees on campus [19]. No similar legislation exists in Canada but a range of important legal issues affecting campus emergency alerting systems must be considered, including duty to warn, liability for failure to warn or for false alerts, system security obligations, and the implications of privacy legislation on the content and distribution of warning messages. In addition, campus emergency planners face a complex set of planning and communications concerns with regard to alerting systems [20]. These are often complicated by a decentralized governance structure and the culture of autonomy typical of an academic setting. Consideration of external factors is also important. Of particular importance are the institution’s obligations with respect to provincial or state and federal emergency management legislation, and in relation to the jurisdictional responsibilities of local emergency management agencies and first responders (e.g., police, fire, emergency medical service). As post-secondary institutions come to terms with these concerns, various reports and plan38 | ning documents have been drafted and released [21], [22]. However, there remains a significant gap in Canada in terms of a systematic assessment of policy, legal, and planning concerns. A study on emergency preparedness of Canadian universities was published in 2004 but the findings are limited by the study’s specific scope and intent [23]. As such, another objective in conducting both the workshop and the survey analyzed here was to identify and better understand the kinds of socio-technical dynamics that are influencing the planning process. The workshop discussion revealed two competing perspectives in terms of broad policy and planning considerations for campus alerting. While most participants agreed that text messaging should be regarded as only one channel within a multi-channel strategy for campus alerting, there was also a clear difference of opinion around the use of this channel for general purpose messaging. One school of thought, illustrated by the University of Calgary, asserts that a text-based mass notification system ought to be reserved exclusively for critical, or so-called “short-fuse” emergency incidents that require the campus community to be alerted widely and quickly. A competing school of thought, and one illustrated by Concordia University’s text messaging initiative, asserts that emergency alerting ought to be a mandatory service bundled within a general purpose messaging strategy that permits students to elect to receive non-emergency notifications, such as class closures, bus schedules, administrative notices, and registration deadlines, etc. On the one hand, the potential advantage of the “general messaging strategy” approach is that routine communications between the university and its community build credibility and familiarity with the system, thereby enhancing the state of readiness among staff and students in the event it is activated during an emergency. On the other hand, such a strategy might be counterproductive as staff and students discount the value of the messages received through this channel and come to ignore or opt-out of the system altogether. This view is that people will eventually come to ignore alerts if they are over-exposed to warning messages. Various research studies, including the one noted above [16] seem to indicate that this is not likely, provided that the warning messages are distinct from other types of messages and that the campus community has been informed in advance as to the circumstances under which alerts will be issued. In order to examine these two views from the recipient’s perspective, the survey asked members of the campus community if they would want to receive information messages other than for urgent emergencies on their mobile phone. Most respondents indicated they would voluntarily subscribe to a text messaging alert service if it were offered, but the overall response was more varied when asked “what other types of information messages would you like to receive on your mobile phone either by voice or text message.” The results indicate that two-thirds of students would be receptive to the idea of receiving other types of messages, especially information about class cancellations, as well as registration and other university deadlines. However, about twothirds of faculty and staff are of the opposite view and would want the service to be reserved for emergency communications only. By contrast, most students, faculty and staff tended to agreed with the statement that “it would not be acceptable to receive alerts for incidents that are not urgent emergencies.” This result suggests that many students would likely be receptive to a campus IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 alerting system bundled to a general purpose messaging strategy, but that there is a need for university planners to carefully consider how to distinguish other types of messages from emergency alerts. Further research is needed to more clearly establish an acceptable definition for “urgent emergencies” by taking into account a wide range of factors such as the location, urgency, severity, and certainty of an incident. Workshop attendees also raised concerns about the reliability and the throughput capabilities of the mass-messaging products that are being marketed to Canadian postsecondary institutions by private information technology (IT) vendors. While there is growing consensus among campus emergency planners that using text messaging alone has serious drawbacks, it remains a relatively cost-effective means of quickly alerting the campus community, especially students. However, this new opportunity has also encouraged numerous thirdparty vendors to enter the market and offer universities the technical means to introduce large numbers of SMS messages into public mobile phone networks without the concomitant ability to guarantee that these messages can be processed and delivered to the end-user in a timely manner. Moreover, the impact of such a flood of messages on the public telecommunications system means that critical voice services, such as calls to 911, may also be negatively impacted due to network congestion [24]. Traynor [7] has documented several other drawbacks with text messaging, including problems with locating recipients, managing delays, and proper sequencing of multiple messages, as well as potential security breaches that could lead to fraudulent alerts. Further research is needed to determine if there are ways over the long term to improve the reliability of text messaging for The effectiveness of campus alerting systems should be measured using both social and technical criteria. public alerting and how best to proceed with the direct participation of the telecom network operators and third-party suppliers of mass notification systems [25], [26]. In the short term, however, we need to better understand the most appropriate role for SMSbased alerting in conjunction with other methods, as well as how to implement and manage a multi-channel alerting policy that best aligns with the needs of the campus community. To explore this matter further, we can return to the survey results about emergency alert preferences. (See Tables I and II.) Findings from the study indicate comparatively strong preferences among all three universities for the use of email for notification for severe weather, public health events, and theft incidents. Findings from the University of Alberta, where people are well aware of the threat of tornadoes, also showed a strong preference for the use of sirens or a public address (PA) system for severe weather incidents. Oddly enough, however, the results also show comparatively strong preferences across all three universities for using sirens or PA systems to notify for hazardous material/ chemical release; active shooter; and bomb threat/suspicious package incidents. The nature of such incidents is such that public announcements or siren blasts may in fact be contrary to emergency management best practices, suggesting the need for further study into the assumptions and expectations among members of the campus community about the role of an alert system in the broader context of emergency response and crisis management. IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 One surprise finding in terms of text messaging was that, in most cases, less than one-quarter of respondents chose this as a preferred method for any of the identified incident types. Even when results for the student population are examined separately the results do not change significantly. Given the apparent popularity of text messaging among the student population more generally, this finding is quite surprising. However, it is possible that actual experience with text-based “alert” services is modest and that results could change over time. For example, when asked if they subscribe to other types of non-emergency “alerting” services such as stock market, sport scores, etc., over 90 percent of students, staff, and faculty indicated they did not. As such, the low levels of expressed preferences for sending alerts by text message may be related more to current perceptions based on lack of experience with the use of a mobile phone as an alerting device. It is important to explore this possibility further and future focus group work is planned that will give participants an opportunity to experience receiving sample alert messages by mobile phone as part of an in-depth qualitative analysis of alerting preferences and warning response behavior. Emerging Technology and P2P Networks Although our survey findings seem to indicate that mobile phones are not the preferred method for campus alerting, their ubiquitous presence on campus is nonetheless a growing factor in campus crisis management. As such, there is a need to better understand how the widespread use of two-way text messaging, video messaging, and peer-to-peer | 39 Most respondents agreed that text messaging should be regarded as only one channel within a multi-channel strategy for campus alerting. (P2P) communications will transform campus crisis management during the immediate response stage following an alert activation. P2P communications now play an important role in emergency management, as observed during Hurricane Katrina, the London bombings of July 7, 2005, and the Virginia Tech shooting [27], [28]. The City of New York and Purdue University are also reported to be running trials that allow citizens to upload text and images from mobile phones to report crimes and other emergency situations [29]–[31]. Academic researchers have coined the term “crisis informatics” to describe this emerging field of study, and some early work has examined the use of P2P communications with both the Virginia Tech and Dawson College shooting incidents in 2006 [32], [33]. Early findings have led Palen et al. [34] to argue that emergency planners must now find ways to “incorporate citizen activity and citizen-generated information in formal warning, response, and relief efforts.” When we asked respondents if their university should consider social networking sites as part of its overall emergency alerting strategy, about one-half responded positively. However, when asked how often they checked these sites while on campus, most respondents (greater than 90 per cent) indicated that they do so only occasionally, rarely, or never. This suggests that while a social networking site like Facebook might be useful in terms of providing information updates concerning an emergency incident, it would likely not at this time be an effective means for alerting the campus community to an unfolding event. Nevertheless, there is growing interest in the role that social networking has played in recent disasters and the possibility for peer-based communications as both a threat and an opportunity to crisis management activities on campus. Table II “In addition to urgent campus emergency alerts, what other types of information messages would you like to receive on your mobile phone, either by voice or text message?” Yes. Would want to receive other types of information messages Class cancellations Special events University account Campus Traffic Registration and deadlines No. Would not want to receive other types of information messages Students Faculty (undergrad and graduate) 64.6% 35.3% Nonacademic staff 35.1% Total 64.9% 52.4% 47.6% 50.8% 8.8% 9.8% 13.3% 26.2% 35.4% 64.7% n = 4007 40 | Early work has examined the design of a social networking platform for campus crisis management [35] but more remains to be done. The annual conference organized by the International Community on Information Systems for Crisis Response and Management is one venue where this type of work will likely surface. In addition to the human factors research, further conceptual elaboration is needed to more closely examine the P2P phenomenon in light of previous sociological findings on “emergent response groups” during emergencies [36]–[38]. This existing body of literature provides an important conceptual foundation on which to begin to advance theoretical work in this area in order to examine how P2P communications might be harnessed to improve campus crisis management. Integrating Human Factors into Emergency Messaging An active campus emergency messaging strategy requires the integration of social and technical factors that account for human responses to warnings, institutional policy, and procedures, as well as the functional limitations of communications infrastructure and systems. Findings from the workshop held in Canada and from the survey done at three Canadian universities reinforce this view and clearly indicate that there is no “magic bullet” solution for campus alerting. In fact, there is considerable risk entailed in simply implanting a mass notification system on campus, especially one based on text messaging as a primary alerting channel. In the final analysis, an effective alerting system must be designed and managed as one component of a comprehensive program for campus safety with due regard for the interplay between new communication technologies and a complex array of social factors. Responsibility for this integration will extend to both IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 administrators and users of these systems. The research being done by CEMRG in Canada is intended to contribute to the development of a solid base of knowledge that will support the development of training materials and other resources. These will enable members of the campus community to effectively meet the challenge of campus emergency alert systems. Author Information Gordon A. Gow is Associate Professor, Graduate Program in Communication and Technology, at the University of Alberta, Alberta, Canada; email: ggow@ualberta.ca. Tara McGee is Associate Professor in the Department of Earth and Atmospheric Sciences, University of Alberta, Alberta, Canada; email: tmcgee@ualberta.ca. David Townsend is Professor, Faculty of Law, University of New Brunswick, New Brunswick, Canada; email: townsend@unb.ca. Peter Anderson is Associate Professor, School of Communication, at Simon Fraser University, Canada; email: anderson@sfu.ca. Stanley Varnhagen, Ph.D., is with the Faculty of Extension, University of Alberta, Alberta, Canada; email: stanley.varnhagen@ ualberta.ca. Acknowledgment The authors gratefully acknowledge the University of Alberta Faculty of Extension, the Faculty of Law at the University of New Brunswick, TELUS, and Bell Canada for their support of the National Roundtable on Campus Emergency Messaging. Ongoing research is being supported with funding from the Social Sciences and Humanities Research Council of Canada. References [1] Virginia Tech Review Panel, “Mass Shootings at Virginia Tech April 16, 2007: Report of the Review Panel,” vol. 2008, 2007. [2] Purdue University, “Text messaging to be tested as an alert method,” Secure Purdue, 2007. [3] University of Calgary, “University launches emergency text messaging,” News & Events, 2007. [4] P. Hall, “Early warning systems: Reframing the discussion,” Australian J. Emergency Management, vol. 22, pp. 32–36, 2007. [5] J. Sorensen, “Hazard warning systems: Review of 20 years of progress.,” Natural Hazards Rev. pp. 119–125, 2000. [6] D. Latimer, “Text messaging as emergency communication superstar? Nt so gr8,” Educause Rev., vol. 43, pp. 84–85, 2008. [7] P. Traynor, “Characterizing the limitations of third-party EAS over cellular text messaging services.,” 3G Americas, 2008. [8] “Report finds problems with text-message alert systems,” Chron. of Higher Education, 2008. [9] K. Dow and S.L. Cutter, “Public orders and personal opinions: Household strategies for hurricane risk assessment,” Environmental Hazards, pp. 143–155, 2000. [10] K.E. Rowan, “Why rules for risk communication are not enough: A problem-solving approach to risk communication,” Risk Analysis, vol. 14, pp. 365–374, 1994. [11] D. Mileti and P. O’Brien, “Warnings during disaster: Normalizing communicated risk,” Social Problems, vol. 39, pp. 40–57, 1992. [12] National Research Council, Facing Hazards and Disasters: Understanding Human Dimensions. Washington, DC: National Academies Press, 2006. [13] N. Dash and H. Gladwin, “Evacuation decision making and behavioral responses: Individual and household,” Natural Hazards Rev., vol. 8, pp. 69–77, 2007. [14] A. Fothergill, “The neglect of gender in disaster work: An overview of the literature,” Int. J. Mass Emergencies and Disasters, vol. 14, pp. 33–56, 1996. [15] B.D. Phillips and B. H. Morrow, “Social science research needs: Focus on vulnerable populations, forecasting, and warnings,” Natural Hazards Rev., vol. 8, pp. 61–68, 2007. [16] K. Dow and S. Cutter, “Crying wolf: Repeat responses to hurricane evacuation orders,” Coastal Management, pp. 237–252, 1998. [17] D. Mileti, Disasters by Design. Washington, DC: Joseph Henry Press, 1999. [18] D. Mileti, “Factors related to flood warning response,” presented at U.S.-Italy Research Workshop on the Hydrometeorology, Impacts, and Management of Extreme Floods (Perugia, Italy), 1995. [19] Office of Post Secondary Education, The Handbook for Campus Crime Reporting. Washington, DC: U.S. Dept. of Education, 2005. [20] Partnership for Public Warning, “A national strategy for integrated public warning policy and capability,” PPW rep. 2003-01, 2003. [21] Florida State University, “University campus emergency notification systems report,” State Working Group on Domestic Preparedness, Ad Hoc Committee on University and College Campus Emergency Notification Systems, July 6 2007. IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SUMMER 2009 [22] G. Bernstein, “A planned multimedia emergency communications strategy,” Network and Communications Services, McGill University, Montreal, Canada, July 10, 2007. [23] K. Friesen and D. Bell, “Risk reduction and emergency preparedness activities of Canadian universities,” Public Safety and Emergency Preparedness, Ottawa, Canada, 2004. [24] “SMS over SS7,” National Communications System, Technical Information Bulletin 03-2, 2003. [25] L. Wei, “Report on four technologies demonstrated at UCF on July 27, 2007,” University of Central Florida, School of Electrical Engineering and Computer Science, 2007. [26] X. Meng, P. Zerfos, V. Samanta, S. Wong, and L. Songwu, “Analysis of the reliability of a nationwide short messaging service,” in Proc. INFOCOM 2007, 26th IEEE International conference on Computer Communications, 2007. [27] N. Cohen, “The latest on Virginia Tech, from Wikipedia,” New York Times, 2005. [28] T. Douglas, “Shaping the media with mobiles,” BBC News Online, 2005. [29] “Facebook, text messaging helps universities alert students in emergencies,” Macleans.ca, 2007. [30] “Purdue uses Facebook as part of Emergency Alert System,” Frankfort Times, 2007. [31] “New York City to enable 911 and 311 Call Centers to receive digital images and video,” Government Technology, 2007. [32] L. Palen and S. B. Liu, “Citizen communications in crisis: Anticipating a future of ICT-supported participation,” in Proc. ACM Conf. Human Factors in Computing Systems, 2007. [33] P. Bérubé, “Social uses of new communications technologies in crisis situations,” presented at Industry Canada National Symp. Emergency Telecommunications (Montreal, Canada), 2007. [34] L. Palen, R.S. Hiltz, and S.B. Liu, “Online forums supporting grassroots participation in emergency preparedness and response,” Commun. ACM, vol. 50, pp. 54–58, 2007. [35] P.F. Wu, Y. Qu, J. Preece, K. Fleishmann, J. Golbeck, and P. Jaeger, “Community response grid (CRG) for a university campus: Design requirements and implications,” presented at 5th Int. ISCRAM Conf. (Washington, DC), 2008. [36] T.E. Drabek and D.A. McEntire, “Emergent phenomena and the sociology of disaster: Lessons, trends and opportunities from the research literature,” Disaster Prevention and Management, vol. 12, pp. 97–112, 2003. [37] A. Majchrzak, S. Jarvenpaa, and A. Hollingshead, “Coordinating expertise among emergent groups responding to disasters.,” Organization Science, vol. 18, pp. 147–161, 2007. [38] H. Rodriguez, J. Trainor, and E.L. Quarantelli, “Rising to the challenges of catastrophe: The emergent and prosocial behavior following Hurricane Katrina,” Annals Amer. Acad. Political and Social Science, vol. 604, pp. 82–101, 2006. | 41