5 page literature review

Emergency Alert Communications on College Campuses: Understanding Students’ Perceptions of the Severity of a Crisis and Their Intentions to Share the Alert With Parents and Friends Pavica Sheldon

The purpose of this research is to experimentally test the impact of emergency alert communications on college students’ threat assessment and resulting behavior. Four different stimuli are created to resemble an actual warning message about a tornado and a shooting crisis on a college campus. Results of the repeated measures MANOVA reveal that warning messages sent through text messages are perceived to be more serious than warning messages sent through social media. This is the case in both the shooting and tornado scenarios. In addition, an alert about a shooting is perceived to be more serious than a tornado alert. In terms of channels used to notify others about a disaster, word of mouth is the top choice for students.

Keywords: Cell Phones; College Students; Crisis Communication; Social Media

One of the central missions of all institutions of higher education is to provide safe and secure learning environments. Multiple tragic events, including school shoot- ings, terrorist attacks, and natural disasters, reveal that campus communities are vulnerable and that technology may play a key role in better protecting them. After the shooting at Virginia Tech in 2007, a number of colleges started imple- menting their own emergency alert systems that could be used for any type of

Pavica Sheldon is an Assistant Professor and a Graduate Program Director in the Department of Communication Arts at the University of Alabama in Huntsville. This paper was presented at the Global Communication Association conference in Berlin in 2015. Correspondence to: Pavica Sheldon, Department of Communication Arts, 342-C Morton Hall, The University of Alabama in Huntsville, Huntsville, AL 35899, USA. E-mail: [email protected]

Western Journal of Communication Vol. 82, No. 1, January–Febuary 2018, pp. 100–116

ISSN 1057-0314 (print)/ISSN 1745-1027 (online) © 2017 Western States Communication Association DOI: 10.1080/10570314.2017.1308005

public crisis (Kopel, Sims, & Chin, 2014). Similarly, many universities have their own departments of emergency preparedness that help students prepare for an on- campus incident (Thompson & Schlehofer, 2014). They use warning channels such as voice calls, text messages, social media posts, and announcements posted on the schools’ official websites (Horton, 2012; Romano, 2013). However, there is little research about how students respond to those warning messages. Most previous studies have focused on how the public uses media to communicate interperson- ally with each other during a natural or a technological disaster (e.g., Palen, Vieweg, Hughes, & Hughes, 2009; Procopio & Procopio, 2007).

To our knowledge, there are no studies focusing on how students use social media during a natural disaster or a crisis. There are two studies investigating text messages. McGee and Gow (2012) from the University of Alberta in Canada examined how undergraduate students living on campus may respond to an emergency alert warning sent by the university through their cell phone. They found that students would believe an emergency alert warning sent by the university as long as the phone number that appeared on students’ phones was a number that they associated with the university’s emergency alert system. However, due to the length of the text messages, students complained that the lack of details created uncertainty about the threats so they did not know how to protect themselves. Another study (Lee, Chung, & Kim, 2013) examined factors that motivate individuals’ adoption or rejection of mobile SMS text alerts. Lee et al. (2013) found that individuals’ were more likely to adopt the mobile-based text alert SMS service when their friends and family urged them. However, users rejected messages that they perceived to be intrusive. Lee et al., therefore, suggested ways to reduce the perceived intrusiveness by promoting message receivers to pass along emergency messages through social media such as Face- book and Twitter.

The role of social media has been largely unexplored in emergency communica- tions, although they are the most popular media choice among college students and young adults (Pew Research Center, 2011; Sheldon, 2015). Social media is a group of Internet-based applications that build on the technological foundations of Web 2.0, which allows the creation and exchange of user-generated content (UGC) (O’Reilly, 2005). User-generated content refers to any form of content that is created by consumers. Social media technologies take many different forms, including social network sites (Facebook, LinkedIn, Google+), blogs, wikis, online video and photo- sharing sites (YouTube, Pinterest), and video/text chatting sites (Skype). Public officials have effectively used social networking services, including Facebook, as a warning tool during natural disasters, especially when trying to broadcast a message to those at risk (Sutton et al., 2014). Social network sites could also serve as a secondary communication tool, encouraging students to share an alert message with another person (Austin, Fisher Liu, & Jin, 2012). However, the primary function of social network sites is to connect with friends and family, as users have a list of people that they follow (Sheldon, 2008).

Western Journal of Communication 101

Social Media’s Role During a Crisis

Natural disasters and crises are not a new thing; however, with the emergence of social media, things have changed as the entire world can learn about them in a matter of seconds. Ordinary people use social media to spread news and raise money; rescue and relief workers communicate with one another and with citizens who need help; the displaced reach out to their family and friends. For many, social media provide the latest developments during and after a disaster. Social media have the potential to get a message out quickly to anywhere in the world. Social media is so important that the Federal Emergency Management Agency has urged citizens in a crisis to let their loved ones know that they are okay by updating their Facebook status (Sheldon, 2015). They also have a viral potential as people share videos taken by smartphones on social network sites.

The terrorist attacks on 9/11 showed the need for advanced technology in com- municating in a disaster situation (Freberg, Saling, Vidoloff, & Eosco, 2013). In contrast to 9/11, when US Airways Flight 1549 crashed into the Hudson River in 2009, social media got the news out immediately. Twitter was the first medium to publish the photo (Baron & Philbin, 2009). 2011 Hurricane Irene was the first U.S. weather event where official sources used social media to disseminate information about the disaster and preparation (Abbasi, Kumar, Filho, & Liu, 2012). 1.1 million tweets mentioned the word “hurricane” during Hurricane Sandy in 2012. Sandy became the second most talked about topic on Facebook in 2012 (Fraustino, Liu, & Jin, 2012). Even before the popularity of social network sites, people were using other types of social media to either get information or to discuss a disaster. Twenty minutes after the Virginia Tech shooting in 2007, a Facebook page appeared: “I’m OK at VT.” Only 90 minutes after the event, a Wikipedia webpage was created that accurately described the massacre (Sheldon, 2015).

Social media have a dual role during emergency and crisis situations. They can be used as a) a warning tool; or as b) a secondary crisis communication tool. Most studies focus on using media as a secondary crisis communication tool, or as a way to share the message with other people. For example, Cho and Park (2013) explored social media use during Japan’s 2011 catastrophic earthquake. They found that immediately after the earthquake landline and mobile phones were unavailable, and people had to rely on social media. Austin et al. (2012) also found that audiences use social media during crises for insider information and to check in with family and friends, while they use traditional media for educational purposes.

In addition, social media can be used as an alternative to telephones when calling for help (Armour, 2010). They can function as a technology that alerts others of the safety of those affected by the crisis (Palen et al., 2009). They can help others learn what type of emergency has occurred and how serious the event was (Palen et al., 2009). In addition, they can encourage others to donate to disaster response efforts. One does not have to live locally to get immediate news about what happened (Tinker, Dumlao, & McLaughlin, 2009).

102 P. Sheldon

Cell phones and text messages can serve both functions—to alert the public about a crisis or as secondary crisis communication tools to inform others about an ongoing emergency. In 2001, during the September 11 terrorist attacks in the United States, cell phones were used to communicate during the emergency for the few minutes that they worked inside the buildings after the planes had crashed (Nelson, Sigal, & Zambrano, 2010). When cell phone services were down during Hurricane Katrina in 2005, people continued using text messaging (Juric & Sylvester, 2007). However, due to the length of the text messages, some people complained that the lack of details created uncertainty about the threats so they did not know how to protect themselves (McGee & Gow, 2012). Media richness theory (Daft & Lengel, 1984, 1986) can explain some of their findings. Face-to-face is the richest medium due to its visual and social cues, including body language and tone of voice. Telephones are the second richest medium (Daft & Lengel, 1986). Media of low richness present fewer cues and restrict feedback. In times of uncertainty, communicators should use forms that are highly rich (Trevino, Daft, & Lengel, 1990, p. 75).

To this date, researchers have not tested how audiences react to a message that is sent through different forms of media. Social media did not exist when Daft and Lengel (1984) proposed media richness theory. This study compares rich, but asyn- chronous, social media to a narrowly focused and more direct text messaging. It further explores media as a secondary crisis communication tool to find out if students use face-to-face communication, text messaging, phone calls, or social media to share the news about the crisis. According to media systems dependency theory (Ball-Rokeach & DeFleur, 1976; Lin & Lagoe, 2013), the individuals most highly dependent on media are the most affected by media content. This is especially true in periods of social disruption (Lowrey, 2004). For example, in the aftermath of the September 11, 2001, terrorist attacks, media dependency theory significantly predicted audience behavior (Lowrey, 2004). People were highly dependent on tele- vision and print, and less dependent on interpersonal communication, radio, and the Web.

The National Research Council of the National Academies published a workshop report (National Academies Press, 2011) outlining six research gaps when it comes to the public response to alerts and warnings concerning a crisis. One of the gaps included message dissemination and information-seeking behavior: What do people do in response to alerts? For example, will people place telephone calls to friends or relatives, search for Internet information, or browse news and information? The council emphasized the importance of social media, especially when used as a secondary source during an alert or warning.

Therefore, the purpose of the current research was to experimentally test how emergency alerts sent through a text message vs. social media impact students’ perception of the seriousness of a crisis, as well as how they share those messages with others, both in terms of the frequency of secondary crisis communication and a preferred channel. Crises include an active shooting on campus and a tornado warning.

Western Journal of Communication 103

A disaster is a “serious disruption of the functioning of a community or a society causing widespread human, material, economic, or environmental losses which exceed the ability of the affected community or society to cope using its own resources” (National Science and Technology Council, 2005, p. 21). Unlike disasters that are community-based and often uncontrollable, crises are organization-based and often man-made (Seeger, Sellnow, & Ulmer, 1998). A crisis is defined as “an event, which is often sudden or unexpected, that disrupts the normal operations of the institution or its educational mission and threatens the well-being of personnel, property, financial resources, and/or reputation of the institution” (Zdziarski, 2006, p. 5). While shooting on a campus would be classified as a crisis (man-made), a tornado would be classified as a natural disaster (uncontrollable) (Seeger et al., 1998). In this manuscript, we use the terms interchangeably as a natural disaster can lead to a crisis.

Threat Assessment and Response to a Disaster

The Protective Action Decision Model (PADM), developed by Lindell and Perry (2004), explains the process of how people “typically” make decisions about adopting actions to protect against environmental hazards. This model can be applied to any type of crisis. The response to the threat depends on evaluations of both the threat and the available protective actions. It may include either resuming normal activities, seeking more infor- mation, pursuing an action to protect themselves and other people, or engaging in emotion-focused actions to reduce their immediate psychological distress.

However, not all warnings will lead to behavioral compliance. For example, Sher- man-Morris (2010) examined student responses following a tornado warning on a university campus in Mississippi. The results showed that student behavior was influenced by belief about whether the tornado was a serious threat. Furthermore, Kopel et al. (2014) discovered that students at the University of Central Florida did not take their emergency alert system seriously. McGee and Gow (2012) studied three types of risk messages sent to university students in Canada: threatening warnings, which they perceived as indicating the greatest risk, tornado warnings, which they perceived to be high in risk, and fire warnings, which were perceived to be a low risk. Students said that instead of confirming a warning message with family members, they would confirm it with fellow students and friends, as well as faculty and staff.

Schultz, Utz, and Goritz (2011) experimentally tested the effects of an apology, sympathy, and information as response strategies via traditional and social media on crisis perceptions and secondary crisis communication. They found that media use had a strong effect on secondary crisis communication. One examined crisis involved Mercedes-Benz. The case described that there have been thousands of accidents and 10 fatalities due to problems with spark plugs. Depending on the experimental condition, respondents were presented with one of the three reactions via one of the three media (newspapers, Twitter, blogs). Schultz et al. found that people talk more about newspaper articles than about blogs or tweets. Twitter users were more likely to share the message than blog users and nonusers of social media. Interestingly,

104 P. Sheldon

they were also more likely to share the newspaper article than the blog post or tweet. The authors argued that future research should focus not only on the content of the communications but also on different media and their effects on the recipients’ communications and reactions, in order to evaluate efficient ways of responding to the crisis. According to Lasswell’s model (Lasswell & Bryson, 1948), all communica- tion should be analyzed in terms of “who” says “what,” via what “medium,” to “whom,” and directed at what kind of change.

Therefore, the goal of the present study was to answer the following questions:

Research Question 1: How do university alert messages sent through a text message vs. social media influence students’ perceptions of the seriousness of a crisis? Research Question 2a: How do students share university alert messages with other people (in terms of the frequency of secondary crisis communication and a pre- ferred channel)? Research Question 2b: Who do students share the alert messages with? Research Question 2c: Which media do students use to share the alert messages?

Based on reviewed studies (Kopel et al., 2014; McGee & Gow, 2012; Sherman-Morris, 2010), we hypothesize that the perceived severity of a warning message will influence students’ likelihood to share the message with their significant others (Hypothesis 1). In other words, the more serious that they perceived the alert to be, the more likely they were to share it with other people.

METHODS

Participants

One hundred seventy-two students (89 men and 83 women) participated in a 2 by 2 within- subject experiment involving technology type (social media vs. text message) and crisis type (shooting vs. tornado). Respondents ranged in age from 19 to 55 (mean age = 24 years; SD = 5.64). Following Institutional Review Board approval, principal investigator and a student assistant recruited the participants through various classes offered at a southern research university. Instructors were contacted and asked to have their students fill out an online survey. A variety of majors were represented: 17.6% engineering majors, 17% business, 9% nursing majors, 9% biology, 6% psychology. The rest of the students were majors in departments in the College of Liberal Arts, as well as College of Science. In terms of race/ethnicity, 56.3% of students identified as Caucasian, 15.3% as African Americans, 4.5% as Asian Americans, 0.6% as Hispanics, 1.7% as American Indians, and 16.5% of the participants did not fall into any of the aforementioned ethnicities. On average, participants spent 1.38 hours on Facebook per day (SD = 1.64), .57 hours on Twitter (SD = 1.19), and sent 55 text messages per day (SD = 75.59).

Stimulus Material

Due to the time-sensitive topic of this study, it was not possible to examine students’ actual reactions to an emergency alert message issued by the university. However,

Western Journal of Communication 105

hypothetical scenarios were created and presented as mocked screenshots that could resemble an actual emergency alert (see Appendix for the mock scenarios). The text for each of the warning messages was taken from the university’s previous text messages sent to students about crises.

After reading four different scenarios (social media shooting alert, text mes- sage shooting alert, social media tornado alert, text message tornado alert), participants filled out a questionnaire measuring their secondary crisis commu- nication, their perception of the severity of the event, and the media that they would use to share this message with their immediate family. Each participant received all of the four conditions. Stimuli were presented in four different orders as a counterbalance to randomly distribute variables. Surveys were posted on Qualtrics.

Measures

The perceived severity of an alert message was measured with a Likert-scale item asking, “How serious would you consider this message to be?” Responses range from “extremely serious” (5) to “not at all serious (1)”. See Table 1 for the means and standard deviations across all four scenarios.

To measure secondary crisis communication, participants first indicated how likely they are to share the message with the following people: a) anyone sitting around you; b) parents; and c) friends. Answers were given on a 5-point scale ranging from “very unlikely” to “very likely.” In order to determine the overall likelihood of sharing a certain message, four composite variables were created from averages for each med- ium (social media and text message) and crisis type (shooting and tornado) (see Table 2). After each scenario, an open-ended question was included to allow subjects

Table 1 Perceived Seriousness of a Crisis Alert

Crisis Type Medium M SD

Shooting Social Media 3.98 1.63

Shooting Text Message 4.74#1 .70

Tornado Social Media 3.22 1.44

Tornado Text Message 3.73 .92

Table 2 Likelihood to Share the Message With Other People

Crisis Type Medium M SD

Shooting Social Media 3.36 1.64

Shooting Text Message 3.81 1.09

Tornado Social Media 4.49#1 .91

Tornado Text Message 3.94 1.54

106 P. Sheldon

to explain their answers. The second question asked participants to indicate the first people who they would likely share the message with. Response options included parents, friends, anyone sitting around you, and other. Participants were also asked to indicate which medium they would use to share the message with their immediate family (parents, brothers, and sisters). Response options consisted of a phone call, text message, Facebook post, and e-mail.

After finishing all of the scenarios, participants answered basic demographics questions (age, gender, and race), as well as two questions concerning their previous experience with the tornado and shooting alerts.

Statistical Analysis

Repeated measures multivariate analysis of variance (MANOVA) was used to test for the effect of experimental stimuli on two dependent variables: secondary crisis com- munication and the perceived severity of an alert message. Frequencies were counted to further explain some of the findings. Open-ended questions were analyzed by coding answers into different categories. Hypothesis 1 was tested using linear regression.

RESULTS

RQ1 and RQ2a asked how alert messages about a tornado and a shooting sent through a text message vs. social media influence recipients’ perceptions of the seriousness of a crisis and a secondary crisis communication. Repeated measures MANOVA was used to answer the questions. Results indicated a significant interac- tion effect (crisis * medium) for the experimental condition: Wilks’s λ = .86, F (2,170) = 14.38, p < .001, partial η2 = .15, with both significant main effects. Mauchly’s Test of Sphericity could not be performed because there are only two levels of repeated measures (two for each crisis). Therefore, Greenhouse-Geisser was used to show the effects of crisis and medium type on the threat assessment and secondary crisis communication. Results revealed that the channel used to notify students had a significant effect on the perceived threat, F(1,171) = 39.45, p < .0001, but not secondary crisis communication, p = .37. Thus, warning messages sent through text messages were perceived to be more serious than warning messages sent through social media (Table 1).

In contrast, a crisis type had a significant effect on both the perceived threat F (1,171) = 114.18, p < .0001, and a secondary crisis communication, F(1,171) = 61.83, p < .0001. When it comes to the influence of the crisis type, overall, a shooting was perceived to be a more serious crisis (M = 4.36, SE = .07) than a tornado (M = 3.47, SE = .08). That difference was statistically significant (p = .000). As evident from Table 1, the alert message about the shooting sent via text message was perceived to be the most serious, followed by the social media alert about the shooting. The least serious was the social media alert containing the tornado warning. Interestingly, more

Western Journal of Communication 107

secondary crisis communication for tornado alerts (M = 4.22, SE = .08) was produced as opposed to shooting alerts (M = 3.58, SE = .09) (Table 2).

RQ2b asked about people with whom students would share the warning message. Regardless of the channel and crisis type, most participants would first share the crisis news with anyone sitting around them. This is followed by friends and then parents (Table 3). These results indicate the importance of word-of-mouth communication.

Furthermore, Table 4 shows the percentage of people whom they would contact first during each scenario. As shown in the table, most people said that they would first notify someone sitting nearby. This was true for all four scenarios, although in the case of a shooting that number was higher than in the case of a tornado warning. Among those who would contact parents first, more would do it after receiving the warning about the tornado.

RQ2c asked about the medium that students would use to share the alert messages. As shown in Table 5, in more than 90% of cases, students would either use a phone call or a text message to notify their immediate family (parents and siblings) about the crisis. When it comes to a shooting, more students would choose a phone call over a text message. However, text messages were the first choice for tornado alerts. The type of crisis, therefore, influenced whether a student would choose a phone call or a text message. Very few students would choose Facebook or an e-mail.

H1 predicted that the perceived severity of a crisis would be positively associated with sharing the message with other people. The overall regression model was significant only for the shooting message sent through social media, F (1,171) = 94.56, p = .000, and results indicated that the more serious the warning message about the shooting was perceived, the higher the likelihood that it would be

Table 3 Likelihood to Share Crises Alert With Different People (Mean Values)

Person

Social Media

shooting alert

Text Message

shooting alert

Social Media

tornado alert

Text Message

tornado alert

Someone Nearby 4.32 4.84 3.69 4.17

Parents 3.55 4.17 3.12 3.56

Friends 3.97 4.45 3.26 3.70

Table 4 First Person Being Alerted About the Crisis (% per Scenario)

First Person

Social Media

shooting alert

Text Message

shooting alert

Social Media

tornado alert

Text Message

tornado alert

Someone Nearby 73.2 68.9 54.5 59.4

Parents 11.4 17.4 25.2 20

Friends 15.4 13.2 18.9 20.6

Other 0 .5 1.4 0

108 P. Sheldon

shared with others (β = .60, SE = .06, p = .000). Considering that the previous experience with a disaster might influence the perceived severity of it and the like- lihood to share the message about it with other people, we ran a t-test to compare those who have experienced shooting/tornado warnings before, and those that did not. Our results revealed significant differences when it comes to the perceptions of the level of seriousness of a tornado warning, but only for those sent through the text message. First, those who have experienced tornado warnings before were more likely to perceive the text message to be less serious (M = 3.67; SD = .92), compared to those who have never experienced tornado warnings (M = 4.10; SD = .87); t(170) = −2.00; p = .048. Those who have experienced tornado warnings (N = 150) were also less likely to share the text message alert with people sitting around them (M = 4.10, SD = 1.21), compared to those that have not experienced it before (M = 4.73, SD = .46), t(170) = −2.42; p = .016. The likely explanation might be that most tornado warnings do not end in an actual tornado. When it comes to the shooting, the only statistically significant difference was in sharing the text message about the shooting with parents. Those who had experienced warnings about the shooting before (N = 24) were less likely to share the text message with their parents (M = 3.58, SD = 1.53), compared to those that had not experienced it before (M = 4.27, SD = .46), t(170) = −2.42; p = .016.

DISCUSSION

The purpose of this study was to experimentally explore how channels used to alert students about a campus crisis affect students’ perceptions of the seriousness of the event, their intentions to share the received information with parents and friends, and the preferred medium to do so. First, the results show that warning messages about the crisis that the university sends through text messages are perceived to be more serious than warning messages sent through social media. This was the case in both the shooting and tornado scenarios. These results could be explained if we consider the nature of texting and text messages. Texting is not only faster than logging into social media, but it is near-synchronous (Mentor, 2010). Text messages are also more personal and could be perceived as sent only to us. Moreover, most universities use social media such as Facebook and Twitter to advertise their events, often posting multiple status updates per day (Sheldon, 2015). However, rarely do universities send

Table 5 Medium Used to Alert the Immediate Family About the Crisis (% per Scenario)

Medium

Social Media

shooting alert

Text Message

shooting alert

Social Media

tornado alert

Text Message

tornado alert

Phone Call 51.6 55.5 37.8 39.4

Text Message 46.4 43.9 60.1 60

Facebook Post 1 .6 .7 0

E-Mail 1 0 1.4 .6

Western Journal of Communication 109

text messages to students. It is likely that students will give more attention to a text message alert than to a social media alert, as they are not used to receiving that many text messages from that particular source. Therefore, it appears that messages sent through a medium that is atypical for the organization are likely to be perceived as more serious or important than messages sent through more common means.

In addition, an alert about a shooting is perceived to be more serious than a tornado alert. These results might be due to the location where the study took place. Most participants had experienced a series of tornado warnings in the past (150 out of 172) but did not have experience with a campus shooting (24 out of 172). We also found that the few participants who did not experience tornado warnings before perceived it to be more serious compared to everyone else who experienced them. Research about how prior experiences with specific hazards influence the likelihood to take protective action has been inconsistent. According to some researchers (e.g., Choi & Lin, 2009; Claeys & Cauberghe, 2013), individuals with a high crisis involvement scrutinize crisis information more in depth than those who are low in involvement. However, another group of researchers (for reviews, see Lindell & Perry, 2004) did not find a relationship between personal experience and compliance. According to Mileti and Sorensen (1990), in order to take it seriously, the person must hear and personally attach meaning to the warning by fully understanding it. They must believe it to be true. Unfortunately, people have a tendency to dismiss warnings, not seeing themselves as risks. Optimism bias refers to the mistaken belief that one is less at risk for negative events than one’s peers (Weinsein, 1980). This optimism exists in response to tornado warnings as well (Klockow, Peppler, & McPherson, 2014; Paul, Stimers, & Caldas, 2014). In the Lovekamp and McMahon (2011) study, many students reported a fear of tornadoes, but perceived their personal risk to be low. These findings indicate the importance of taking past experiences into account when trying to understand students’ perceptions, and also when trying to design a crisis plan on the university level.

As findings of this study show, regardless of the channel and crisis type, the first person notified is likely to be someone sitting nearby. The explanation again might be in the nature of the media available to students. To alert parents, students need to use their mobile device. Word-of-mouth is the simplest way to notify anyone sitting around us. McGee and Gow (2012) also found that when receiving a warning message from the university, students would first want to confirm it with fellow students and friends. In their experiment about media use during crises, Austin et al. (2012) found that face-to-face communication was the most important form of crisis communica- tion. Face-to-face is also the richest medium, according to Daft and Lengel (1986), and is preferred in times of uncertainty (Trevino et al., 1990). The telephone is the second richest medium and is clearly preferred as a secondary communication tool when there is an alert about a shooting.

In this study, among students who said that they would contact their parents first, more would do it after receiving the alert about the tornado. This finding is surprising considering that students perceive the shooting alert to be more serious. However, a killer on campus does not pose an immediate danger to the students’ family members.

110 P. Sheldon

Overall, students are more likely to share the alert about a tornado than an alert about a shooter on campus.

In terms of the channels used to contact their immediate family, more than 90% of students would choose either to make a phone call or send a text message to the loved ones. Very few students would send an e-mail or make a post on social media. Text messages are the first choice for tornado alerts while phone calls are the first choice for a shooting. This difference in the medium choice might have to do with the nature of each crisis. A tornado warning alerts students about the possibility of severe weather. In this stage, participants would text-message their loved ones. An alert about a shooting indicates a more serious and acute situation where participants would make a phone call. As a richer medium, allowing audio cues, personal phone calls provide more emotional support than a simple text message (Juric & Sylvester, 2007). However, more secondary crisis communication occurs after students receive an alert about a tornado. Apparently, students are more concerned about their own safety when they hear about a shooting on campus. But when receiving info about a tornado that might also impact their family–especially if they live in the area–students also worry about their significant others’ safety.

When it comes to the preferred channel, two interesting findings emerged. First, students rely on one of the oldest channels—phone calls—to communicate the warn- ing message to their closest family members. Phone calls are fast, immediate, and personal. Although they did not use social media much, social media alerts about a shooting made students share the message more often if they perceived it to be serious. Of all possible scenarios, the relationship between the perceived seriousness of a warning message and secondary crisis communication was only positive when the alert was about a shooting and when it was posted on social media. Unlike text messages, social media messages are usually seen by a larger number of people. The results suggest the importance of social media alerts when it comes to potentially saving the lives of students, faculty raises concern for the family, and staff.

Although this study focused on the emergency alerts that the university sends to students, its findings could be further related to a previous study of Stephens, Barrett, and Mahometa (2013) who found that when organizational members received multi- ple notifications about a large-scale emergency, those who received them through at least one synchronous communication source (e.g., phone or face-to-face commu- nication) perceived the urgency of the situation the most quickly. Those receiving official messages through asynchronous channels (text message) had the lowest sense of urgency. It is possible that the threat assessment and urgency or willingness to take an action are two different things that future research should further investigate.

This study extends prior work on secondary crisis communication by focusing not only on the likelihood to share the crisis alert with other people, but also on the channels used to further share the message, and the first person to be contacted. Furthermore, it indicates the importance of studying the warning channels and their influence on taking protective actions. As described in the protective action decision model (Lindell & Perry, 2004), the channel through which warning information is sent does matter. In this study, the channel influences threat perception. The

Western Journal of Communication 111

perception of the threat, however, mostly relates to the perception of the seriousness of the crisis and less for the secondary crisis communication. When it comes to secondary crisis communication and the likelihood to communicate the alert with others, the choice of the medium used to share the news matters. Cell phones are the most frequently used tool for secondary crisis communication. Although a shooting is perceived as a more serious event, students are more likely to share a tornado alert with others. Thus, the audience chooses media channels to transmit warning messages depending on the functions those media can perform.

This study also has practical implications. Despite previous research (Howell & Taylor, 2012) suggesting that social media might have become the most popular tool of disaster and crisis communications, text messages have a more important role in both alerting people about the crises and also as a way to contact close family members and friends. The findings are in line with emergency preparedness experts’ suggestions to communicate via SMS text messaging when a disaster strikes. Prepa- redness experts warn against making nonessential phone calls, as they can prevent 911 calls from getting through (Text First Talk Second, 2014). As students seem to underestimate the severity of tornadoes, another consideration might be educating them about the deadliness of tornadoes as one of the most dangerous storms on Earth. In addition, a statement that emphasizes the seriousness of a threat should be included in an emergency text alert. Considering that cell phone towers might be down during natural disasters, emergency professionals should develop an alternative plan of contacting students and family when a crisis occurs. One possibility would be to ask students and parents to “like” the university’s official Facebook or Twitter page. If a disaster happens, and phone calls and text messages are disabled, social media could be a way to reach out to the public. Collaborative knowledge can also be more accurate than that from a single source. People talk about the event with or without the official sources (Orlando, 2010). Based on their results Page, Freberg, and Saling (2013) recommend the use of visual social platforms such as Instagram, Pinterest, and Tumblr to inform and teach students through videos and photographs about the importance of taking safety measures before anything occurs. Visual platforms are becoming more popular than text-based forms (Page et al., 2013) and could be effectively used during preparedness stages as well.

Limitations and Future Research

This study was based on four hypothetical scenarios. To improve ecological or external validity, future studies should focus on the perceptions and reactions to a real crisis. The advantage of fake scenarios, however, was simultaneously studying participants’ reactions to different crisis scenarios. Considering that experience emerged as an important predictor of how students respond to an alert message, future studies should test how students react to an alert that contains a statement that emphasizes the seriousness of a threat. In addition, previous experience should be considered when trying to design a crisis plan on the university level. Future studies

112 P. Sheldon

could further test the protective action decision model (Lindell & Perry, 2004), by studying students’ personality differences and social cues, and behavioral response such as information search and situational demands and constraints.

Conclusion

In conclusion, campuses should continue using text messages to alert students about critical events happening on and around campus. However, they should also consider using social media, especially in the preparedness stages of a disaster, by showing students videos of how to properly respond to a crisis. As our study shows, natural disasters are often perceived as being less dangerous than a shooting on campus, despite the fact that they kill more people than a man-made crises.

References

Abbasi, M. A., Kumar, S., Filho, J. A. A., & Liu, H. (2012). Lessons learned in using social media for disaster relief - ASU crisis response game. In SBP’s 12 Proceedings of the 5th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction (pp. 282–289), College Park, MD.

Armour, G. (2010). Communities communicating with formal and informal systems: Being more resilient in times of need. Bulletin of the American Society for Information Science & Technology, 36(5), 34–38.

Austin, L., Fisher Liu, B., & Jin, Y. (2012). How audiences seek out crisis information: Exploring the social-mediated crisis communication model. Journal of Applied Communication Research, 40 (2), 188–207. doi:10.1080/00909882.2012.654498

Ball-Rokeach, S. J., & DeFleur, M. L. (1976). A dependency model of mass-media effects. Commu- nication Research, 3(1), 3–21. doi:10.1177/009365027600300101

Baron, G., & Philbin, J. (2009). Social media in crisis communication: Start with a drill. Public Relations Tactics, 16(4), 12.

Cho, S. E., & Park, H. W. (2013). Social media use during Japan’s 2011 earthquake: How Twitter transforms the locus of crisis communication. Media International Australia, 149, 28–40.

Choi, Y., & Lin, Y. H. (2009). Consumer response to crisis: Exploring the concept of involvement in Mattel product recalls. Public Relations Review, 35(1), 18–22. doi:10.1016/j. pubrev.2008.09.009

Claeys, A. S., & Cauberghe, V. (2013). What makes crisis response strategies work? The impact of crisis involvement and message framing. Journal of Business Research, 67(2), 182–189. doi:10.1016/j.jbusres.2012.10.005

Daft, R. L., & Lengel, R. H. (1984). Information richness: A new approach to manager information processing and organization design. In B. Staw & L. L. Cummings (Eds.), Research in organizational behavior (pp. 191–233). Homewood, IL: JAI Press.

Daft, R. L., & Lengel, R. H. (1986). Organizational information requirements, media richness and structural design. Management Science, 32(5), 554–571.

Fraustino, J. D., Liu, B., & Jin, Y. (2012). Social media use during disasters: A review of the knowledge base and gaps. Final report to Human Factors/Behavioral Sciences Division. Science and Technology Directorate. U.S. Department of Homeland Security. College Park, MD: START. Retrieved from http://www.start.umd.edu/sites/default/files/files/publications/START_Social- MediaUseduringDisasters_LitReview.pdf

Western Journal of Communication 113

Freberg, K., Saling, K., Vidoloff, K. G., & Eosco, G. (2013). Using value modeling to evaluate social media messages: The case of Hurricane Irene. Public Relations Review, 39(3), 185–192. doi:10.1016/j.pubrev.2013.02.010

Horton, J. (2012). Chardon high school shooting hits close to home. Retrieved from http://www. cleveland.com/chardon-shooting/index.ssf/2012/02/chardon_school_shooting_hits_c.html

Howell, G. V. J., & Taylor, M. (2012). When a crisis happens, who turns to Facebook and why? Retrieved from https://blogs.deakin.edu.au/apprj/wp-content/uploads/sites/33/2014/02/12- howell-taylor.pdf

Juric, P., & Sylvester, J. (2007). Mass media use during a natural disaster: Louisiana State University students and Hurricane Katrina. Southwestern Mass Communication Journal, 22, 85–96.

Klockow, K. E., Peppler, R. A., & McPherson, R. A. (2014). Tornado folk science in Alabama and Mississippi in the 27 April 2011 tornado outbreak. GeoJournal, 79(6), 791–804. doi:10.1007/ s10708-013-9518-6

Kopel, D. E., Sims, V. K., & Chin, M. G. (2014). Taking emergency warnings seriously. Proceedings of the Human Factors and Ergonomics Society Annual Meeting September, 58, 1129–1133. doi:10.1177/1541931214581236

Lasswell, H., & Bryson, L. (Eds.). (1948). The structure and function of communication in society. The communication of ideas. New York, NY: Institute for Religious and Social Studies.

Lee, D., Chung, J. Y., & Kim, H. (2013). Text me when it becomes dangerous: Exploring the determinants of college students’ adoption of mobile-based text alerts short message service. Computers in Human Behavior, 29(3), 563–569. doi:10.1016/j.chb.2012.11.014

Lin, C. A., & Lagoe, C. (2013). Effects of news media and interpersonal interactions on H1N1 risk perception and vaccination intent. Communication Research Reports, 30(2), 127–136. doi:10.1080/08824096.2012.762907

Lindell, M. K., & Perry, R. W. (2004). Communicating environmental risk in multiethnic commu- nities. Thousand Oaks, CA: Sage Publications.

Lovekamp, W. E., & McMahon, S. K. (2011). I have a snickers bar in the trunk of my car: Student narratives of disaster risk, rear, preparedness, and reflections on Union University. Interna- tional Journal of Mass Emergencies and Disasters, 29(2), 132–148.

Lowrey, W. (2004). Media dependency during a large-scale social disruption: The case of September 11. Mass Communication & Society, 7(3), 339–357. doi:10.1207/s15327825mcs0703_5

McGee, T. K., & Gow, G. A. (2012). Potential responses by on-campus university students to a university emergency alert. Journal of Risk Research, 15(6), 693–710. doi:10.1080/ 13669877.2011.652653

Mentor, D. (2010). Social connectedness via SMS texting. Proceedings of ACM CHI 2010 Conference on Human Factors in Computing Systems, Atlanta, GA.

Mileti, D. S., & Sorensen, J. H. (1990). Communication of emergency public warnings: A social science perspective and state-of-the-art assessment. Oak Ridge, TN: Oak Ridge National Laboratory.

National Academies Press. (2011). Public response to alerts and warnings on mobile devices: Summary of a workshop on current knowledge and research gaps. Committee on public response to alerts and warnings on mobile devices: Current knowledge and research gaps. Retrieved from https://www.nap.edu/catalog/13076/public-response-to-alerts-and-warnings- on-mobile-devices-summary

National Science and Technology Council. (2005). Grand challenges for disaster reduction. Retrieved from http://www.preventionweb.net/publications/view/2538

Nelson, A., Sigal, I., & Zambrano, D. (2010). Media, information systems and communities: Lessons from Haiti. Retrieved from https://www.knightfoundation.org/reports/media-information- system-and-communities-lessons-h

O’Reilly, T. (2005). What is Web 2.0: Design patterns and business models for the next generation of software. Retrieved from http://oreilly.com/web2/archive/what-is-web-20.html

114 P. Sheldon

Orlando, J. (2010). Harnessing the power of social media in disaster response. Continuity Insights. Retrieved from http://www.continuityinsights.com/articles/2010/09/harnessing-power-social- media-disaster-response

Page, S., Freberg, K., & Saling, K. (2013). Emerging media crisis value model: A comparison of relevant, timely message strategies for emergency events. Journal of Strategic Security, 6(2), 20–31. doi:10.5038/1944-0472.6.2.2

Palen, L., Vieweg, S., Hughes, S. B., & Hughes, A. L. (2009). Crisis in a networked world: Features of computer-mediated communication in the April 16, 2007, Virginia Tech Event. Social Science Computer Review, 27(4), 467–480.

Paul, B. K., Stimers, M., & Caldas, M. (2014). Predictors of compliance with tornado warnings issued in Joplin, Missouri in 2011. Disasters, 39(1), 108–124. doi:10.1111/disa.12087

Pew Research Center. (2011). Twitter responds to Japan’s disaster. Retrieved from http://pewresearch. org/pubs/1933/twitter-response-japan-earthquake-breaking-news-eyewitness-bulletin-board- fund-raising.

Procopio, C., & Procopio, S. (2007). Do you know what it means to miss New Orleans? Internet communication, geographic community, and social capital in crisis. Journal of Applied Communication Research, 35(1), 67–87. doi:10.1080/00909880601065722

Romano, A. J. (2013). The use of social media during a crisis on campus (Master’s thesis). Retrieved from http://cardinalscholar.bsu.edu/handle/123456789/197836

Schultz, F., Utz, S., & Goritz, A. (2011). Is the medium the message? Perceptions of and reactions to crisis communication via Twitter, blogs and traditional media. Public Relations Review, 37(1), 20–27. doi:10.1016/j.pubrev.2010.12.001

Seeger, M. W., Sellnow, T. L., & Ulmer, R. R. (1998). Communication, organization, and crisis. Communication Yearbook, 21, 231–275.

Sheldon, P. (2008). The relationship between unwillingness to communicate and students’ Facebook use. Journal of Media Psychology, 20(2), 67–75. doi:10.1027/1864-1105.20.2.67

Sheldon, P. (2015). Social media: Principles and applications. Lanham, MD: Lexington Books. Sherman-Morris, K. (2010). Tornado warning dissemination and response at a university campus.

Natural Hazards, 52(3), 623–638. Stephens, K. K., Barrett, A. K., & Mahometa, M. J. (2013). Organizational communication in

emergencies: Using multiple channels and sources to combat noise and capture attention. Human Communication Research, 39(2), 230–251.

Sutton, J., Spiro, E. S., Johnson, B., Fitzhugh, S., Gibson, B., & Butts, C. T. (2014). Warning tweets: Serial transmission of messages during the warning phase of a disaster event. Information, Communication, & Society, 17(6), 765–787.

Text First Talk Second. (2014). Safe America Foundation. Retrieved from https://www.safeamerica. org/text-first-talk-second/

Thompson, S. C., & Schlehofer, M. M. (2014). Undermining optimistic denial reactions to domestic and campus emergency warning messages. Applied Psychology: Health and Well-Being, 6(2), 192–213.

Tinker, T. L., Dumlao, M., & McLaughlin, G. (2009). Effective social media strategies during times of crisis: Learning from the CDC, HHS, FEMA, the American Red Cross and NPR. Public Relations Strategist, 15, 25–29.

Trevino, L. K., Daft, R. L., & Lengel, R. H. (1990). Understanding managers’ media choices: A symbolic interactionist perspective. In J. Fulk & C. Steinfield (Eds.), Organizations and communications technology. Newbury Park, CA: Sage.

Weinsein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39(5), 806–820.

Zdziarski, E. L. (2006). Crisis in the context of higher education. In K. S. Harper, B. G. Paterson, & E. L. Zdziarski (Eds.), Crisis management: Responding from the heart. Washington, DC: National Association of Student Personnel Administrators.

Western Journal of Communication 115

Appendix

Mock up scenarios (modified to remove the name of the university)

Social media tornado alert

Social media shooting alert

Text message shooting alert

Text message tornado alert

116 P. Sheldon

Copyright of Western Journal of Communication is the property of Western States Communication Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.