RESEARCH PROPOSAL

Communication Research 2016, Vol. 43(5) 626 –646

© The Author(s) 2015 Reprints and permissions:

sagepub.com/journalsPermissions.nav DOI: 10.1177/0093650214565917

crx.sagepub.com

Article

Social Media Use During Disasters: How Information Form and Source Influence Intended Behavioral Responses

Brooke Fisher Liu1, Julia Daisy Fraustino1, and Yan Jin2

Abstract This study provides insights that can inform disaster communication management, policymaking, and theory building through a nationally representative field experiment (N = 2,015 U.S. adults) grounded in media richness theory, information and communication technologies (ICTs) succession theory, and the social-mediated crisis communication (SMCC) model. Key findings include the following: (1) Significant main effects of disaster information source were detected on how likely participants were to seek further disaster information from TV, local government websites, and federal government websites; (2) regardless of information form and source, participants reported strongest intentions to immediately communicate about the disaster predominately via offline interpersonal forms rather than through online organizational and personal forms; and (3) regardless of information source, participants reported strong intentions to evacuate if instructed to do so by the government. These findings call for developing crisis communication theory that is more focused on how publics communicate with each other rather than with organizations about disasters and predict a wider variety of crisis communication outcomes.

Keywords crisis communication, disaster communication, social media

1University of Maryland, College Park, MD, USA 2University of Georgia, Athens, GA, USA

Corresponding Author: Brooke Fisher Liu, Associate Professor, Department of Communication, University of Maryland, 2130 Skinner Building, College Park, MD 20742, USA. Email: [email protected]

565917CRXXXX10.1177/0093650214565917Communication ResearchLiu et al. research-article2015

Liu et al. 627

Crises create information holes that organizations and publics work together to fill through creating and sharing information, often in real time. For example, after the bombings at the 2013 Boston Marathon, The Boston Globe temporarily converted its homepage to a live blog that displayed tweets from government authorities, news outlets, and citizens (Gilgoff & Lee, 2013). Accordingly, scholars have increasingly looked to explain and predict how various crisis information forms and sources affect how publics respond to crisis information via traditional and social media (e.g., Jin, Liu, & Austin, 2014; Liu, Jin, & Austin, 2013; Schultz, Utz, & Göritz, 2011; Utz, Schultz, & Glocka, 2013). This bur- geoning scholarship points to key differences in how publics may use social versus tradi- tional media to obtain and share crisis information, and the impact of these choices on publics’ perceptions of and responses to organizations’ crisis communication. Yet, with only a few years of research on this topic, much remains to be understood.

While researchers work to construct more comprehensive knowledge of effective crisis communication in a rapidly changing media landscape, practitioners are altering how they approach crisis communication management and policymaking, largely with- out adequate empirical evidence to guide their decisions. For example, in August 2013, the Federal Emergency Management Agency (FEMA) launched Disaster Reporter with the purpose of crowdsourcing and sharing “disaster-related information for events occurring within the United States, allowing citizens, first responders, emergency man- agers, community response & recovery teams, and others to view and contribute infor- mation on a publicly accessible map” (FEMA, 2013, para. 1). Such disaster resilience initiatives could benefit from being informed by a stronger set of research findings to explain and predict communication needs, preferences, and outcomes surrounding disasters. Consequently, the purpose of this study was to provide additional insights into how disaster information form and source affect the publics’ disaster information seeking and sharing as well as likelihood to take protective actions. In doing so, we conducted a nationally representative field experiment of 2,015 U.S. adults grounded in media richness theory, information and communication technologies (ICTs) succession theory, and the social-mediated crisis communication (SMCC) model.

The study’s key findings include the following: (1) significant main effects of disaster information source on how likely participants were to seek further disaster information from TV, local government websites, and federal government websites; (2) participants reported strongest intentions to immediately communicate about the disaster predominately via offline interpersonal forms rather than through online orga- nizational and personal forms; and (3) participants reported strong intentions to evacu- ate if instructed to do so by the government. These findings call for developing crisis communication theory that is more focused on how publics communicate with each other (rather than with organizations during disasters), echoing calls for more audience- centric research (e.g., Kim & Dutta, 2009), as well as theory that predicts a wider variety of crisis communication outcomes.

Literature Review

The following section reviews literature regarding crisis and disaster communication, particularly related to social media.

628 Communication Research 43(5)

Crisis and Disaster Communication

A variety of definitions of what constitutes a crisis are proposed across crisis commu- nication/management scholarship. Yet as Heath (2010) summarized, most crisis defi- nitions include some amount of “whether the organization knew, appreciated, planned, and appropriately enacted sufficient control over operations to prevent, mitigate, respond, and learn from a crisis” (p. 3). Organizational reputation is often a primary concern of crisis communication research, as many have noted (e.g., Coombs, 2012; Heath, 2010; Heath & O’Hair, 2010; Palenchar, 2010). Thus, crises in particular tend to be viewed by the scholarly community as organization centered.

Associated yet conceptually distinct, disasters, on the other hand, are community- or society-based. Correspondingly, a National Science and Technology Council (2005) report defined a disaster as a “serious disruption of the functioning of a community or a society causing widespread human, material, economic or environmental losses which [sic] exceed the ability of the affected community or society to cope using its own resources” (p. 21). However, disasters can prompt organizational crises, such as when publics perceive an organization’s lack of preparation or other related mishandling sur- rounding a disaster event. Building on this understanding of crises and disasters and expanding others’ conceptions, we define disaster communication as information cre- ation, seeking, and/or sharing among individuals, organizations, and the media sur- rounding an event involving largely damaging violations of publics’ expectations.

Social Media

Social media can play prominent roles in how publics learn and communicate about crises and disasters (Schultz et al., 2011). Social media are digital or mobile tools that are interactive, allowing users not only to access but also to create or influence content (Wright & Hinson, 2009). Thus, social media include multiple platform types such as social bookmarking, social couponing, virtual worlds, video/picture sharing sites, and myriad others. However, social networking sites constitute the most popular social media category (Pew Internet, 2013). As of 2013, 72% of online Americans reported logging on to social networking sites such as Facebook and MySpace, a rate that has continually climbed from the mere 9% who reported doing so in 2005 (Brenner, 2013). With such a large and growing user base as well as two-way communication capabili- ties, social media may be prime venues for businesses, organizations, and individuals to engage in dialogue and content exchange (Bortree & Seltzer, 2009; Taylor & Perry, 2005), assuming that organizations have the resources to do so, at-risk publics are on social media, and social media are available during disasters.

Accordingly, researchers have begun to ask and answer questions to explain and predict publics’ responses to disaster information communicated via social media, finding some similarities and differences with traditional media. Although parsing out the factors contributing to differences in processing and responding to crisis informa- tion via social and traditional media is work in progress, some explanatory and predic- tive patterns have emerged. Such patterns have come to light, perhaps, particularly as a function of research based on the SMCC model (e.g., Jin & Liu, 2010; Liu, Jin, &

Liu et al. 629

Austin, 2013) and the networked crisis communication model (e.g., Schultz et al., 2011). More specifically, two of the factors found to influence publics’ communicative and behavioral responses to disaster information are (1) the mediated form (cf. chan- nel) through which crisis information is conveyed and (2) the source of the crisis information. Furthermore, media richness theory (Daft & Lengel, 1984, 1986) and ICTs succession theory (Stephens, 2007; Westerman, Van Der Heide, Klein, & Walther, 2008) provide insights into how disaster information form(s) independently and/or jointly affect publics’ responses to disasters. Applied research on disaster information source provides additional insights.

Impacts of Disaster Information Form on Disaster Behaviors

Disaster sociologists posit that publics engage in a four-step evaluative process when they receive alert and warning messages: understanding (attaching a personal mean- ing to the message), believing (determining if the risk/disaster, warning, and message contents are accurate), personalizing (understanding that the message is aimed at the recipient), and deciding (determining appropriate action; Mileti & Sorenson, 1990). Through a process called milling, at-risk publics seek information about what hap- pened and what to do, and they share information about the need to take protective action(s) (Mileti & Sorenson, 1990).

Disaster sociologists and communication researchers have posited that social media, particularly Twitter, are effective tools for rapid information seeking and shar- ing and may even generate positive outcomes such as stronger support for organiza- tions in crisis (Schultz et al., 2011; Sutton et al., 2013; Utz et al., 2013). Yet, social media might not be effective information forms to facilitate recommended protective action taking as they often provide content that is quite terse, which may prolong milling as at-risk publics seek additional information (Sutton et al., 2014). This terse- ness can be theoretically explicated through media richness theory (Daft & Lengel, 1984, 1986).

Media richness is a communication form’s “capacity to facilitate shared meaning,” and media richness theory advises communicators to use forms that are highly rich when communicating about uncertainty and equivocality (Trevino, Daft, & Lengel, 1990, p. 75). Forms that are highly rich include visual and social cues such as gestures, and for this reason, media richness research tends to rank face-to-face communication as richest and communication solely through documents such as memos as least rich, or leanest (D’Urso & Rains, 2008). Researchers applying media richness theory have advocated for also considering individual, situational, and organizational characteris- tics that can affect media use and perhaps be more important than media richness (e.g., Rice, D’Ambra, & More, 1998), and have noted that media richness may be fluid rather than fixed (Carlson & Zmud, 1999). Although media richness theory has not, to date, been directly applied to disaster communication, the theory indicates that using information rich forms could speed up milling and thus lead to quicker protective action taking. Furthermore, if publics receive disaster information initially from a source that is inadequately rich, they are likely to seek additional information from

630 Communication Research 43(5)

richer sources that may help them better understand, believe, and personalize disaster information—with the ultimate goal of deciding how to respond.

Building off of media richness theory is research on ICTs succession theory (Stephens, 2007), which suggests that “using complimentary ICTs creates a type of modality expansion, allowing additional modalities to be used to reach and influence message receivers” (p. 102). Modality expansion is linked to complimentary forms offering “expanded cues,” or higher richness, which, in turn, offer “error-reducing redundancy for equivocal and uncertain tasks” (Stephens, Sørnes, Rice, Browning, & Saetre, 2008, p. 197; Stephens & Rains, 2011). In other words, using two different information forms such as one that primarily relies on visuals (e.g., television) and the other that primarily relies on text (e.g., text messages) to repeat a message can be more persuasive than only using one information form or using the same information form twice. For example, Stephens, Barrett, and Mahometa (2013) found in a survey con- ducted after a campus shooting that faculty, students, and staff were more likely to understand the urgency of the emergency when they received information about the shooting via three redundant messages coming from at least one synchronous com- munication source. Research to date, however, has not tested how receiving informa- tion from sequential forms (whether they be the same forms, asynchronous, synchronous, and/or high or low in media richness) impacts disaster information seek- ing and sharing (i.e., milling) and subsequent protective action taking, nor has research tested how disaster information sources may impact publics’ communicative and behavioral responses to the same disaster information presented via a variety of forms.

Impacts of Disaster Information Source on Crisis Behaviors

While media richness theory and ICT succession theories have not tested how infor- mation source affects how publics consume and respond to information, applied research provides mixed results. Some research has found that publics perceive offi- cial sources, such as government agencies, as more credible (e.g., Wogalter, 2006) for disaster information than unofficial sources such as members of the public, both via traditional and social media. Other research, however, revealed that publics sometimes view official sources as slow or outdated, and thus less accurate than unofficial sources (Palen, Vieweg, Liu, & Hughes, 2009), and that publics typically use a combination of unofficial and official sources to make sense of disaster information (Palen, Starbird, Vieweg, & Hughes, 2010; Palen et al., 2009). Still other research points to journalists as the most credible sources of disaster information (Schultz et al., 2011) and indicates that online publics prefer to share news media coverage of disasters rather than gov- ernment coverage of disasters (Chew & Eysenbach, 2010). Yet, when it comes to com- plying with recommended protective actions, the public may be more likely to comply when messages come from a governmental source (e.g., Centers for Disease Control) than from a user-generated source (e.g., blog post shared on Facebook; Freberg, 2012). Finally, outside of disasters, research shows that publics may respond differently to local and national news sources (e.g., Moy, McCluskey, McCoy, & Spratt, 2004), but it is not known whether the same holds true in a disaster context. In sum, the research

Liu et al. 631

record shows that source matters in how publics respond to disaster messages, but it does not provide conclusive guidance on which source(s) might be most desirable for mitigating milling and increasing protective action taking.

Potential Joint Impacts of Disaster Information Form and Source

A limitation of media richness and ICT succession theories is that neither examines the potential joint impacts of information form and source on people’s reactions to mes- sages. Yet, in today’s vast media landscape, during disasters people can receive and consume information from a variety of sources (e.g., governments, media, citizens, etc.) and forms (e.g., social media, traditional media, offline word-of-mouth commu- nication) – and these sources and forms often draw from each other for information (Hughes & Palen, 2009; Jin & Liu, 2010; Reynolds & Seeger, 2012).

Scholars have begun to unveil the potential roles disaster information form and source may play in impacting publics’ information seeking and sharing as well as other thoughts, feelings, and behaviors surrounding crises and disasters. For example, through developing the SMCC model, researchers have attempted to hone understand- ing about how information form and source impact publics’ crisis communication activities (e.g., Jin & Liu, 2010; Liu, Jin, & Austin, 2013). The SMCC model explains and predicts the process of direct and indirect influence of social media in distributing crisis information via social media, traditional media, and word-of-mouth. Probing this process, scholars have looked to information form and source as key factors (e.g., Liu, Austin, & Jin, 2011; Liu, Jin, & Austin, 2013).

For example, Liu et al. (2011) conducted an experiment that manipulated crisis information form (social media vs. traditional media vs. word-of-mouth communica- tion) and crisis information source (a third party vs. the organization). They found a variety of significant differences among groups based on interaction effects of infor- mation form and source on participants’ acceptance of various types of organizational crisis response strategies as well as participants’ crisis-related emotions. Similarly, Austin, Liu, and Jin (2012) conducted mixed-methods research to examine publics’ use of social media, traditional media, and word-of-mouth communication for crisis information seeking and sharing. Experiment results uncovered several main and interaction effects of information form and source. For instance, when exposed to crisis information from social media, participants were more likely to seek additional information from friends’ Facebook pages than when they were exposed to the infor- mation from traditional media. Significant main effects also emerged from crisis infor- mation form and source on publics’ likelihood to engage in crisis information seeking via interpersonal communication, such that those in the social media condition were more likely to seek additional crisis information via word-of-mouth than were those in the traditional media condition; and the same was found for those exposed to crisis information via a third-party source as opposed to the organization. However, SMCC research to date has not yet fully addressed various crisis/disaster types, and it has not focused on prosocial outcomes such as publics’ likelihood to take protective actions that might foster resilience.

632 Communication Research 43(5)

Overall, although the literature reveals a variety of potential main and interaction effects among various disaster information forms/sources and various behavioral out- comes, the research record has (1) not comprehensively included local/national news media as source(s), (2) often blurred the differences between characteristics of media form and those of media content, and (3) rarely measured publics’ intentions to engage in protective actions as an outcome. Given these key research gaps coupled with the loose pattern of findings related to disaster information form and source effects on publics’ behavioral responses, the following research questions are posed:

Research Question 1: How, if at all, does disaster information form (social media vs. traditional media) and source (local government vs. national government vs. local news media vs. national news media) independently and/or jointly impact participants’ intentions to seek disaster information? Research Question 2: How, if at all, does disaster information form (social media vs. traditional media) and source (local government vs. national government vs. local news media vs. national news media) independently and/or jointly impact participants’ intentions to share disaster information? Research Question 3: How, if at all, does disaster information form (social media vs. traditional media) and source (local government vs. national government vs. local news media vs. national news media) independently and/or jointly impact participants’ behavioral intentions to take protective actions?

Method

We conducted a nationally representative field experiment to address the above research questions related to the impact of disaster information form and disaster information source on communicative behaviors (information seeking and sharing) and other behavioral (taking protective actions) outcomes. The effects of disaster information form (social media: tweet vs. Facebook post vs. traditional media: web- site article) and disaster information source (national government: FEMA vs. local government: San Francisco Mayor’s Office of Emergency Management vs. national news media: USA Today newspaper vs. local news media: San Francisco Chronicle newspaper) were examined via a 3 × 4 factorial design between-subjects experiment with a national random sample of 2,015 adult U.S. residents. The disaster type exam- ined was a hypothetical multiple coordinated terrorist attack.

Stimuli Development

Stimuli were modeled from government press releases, government social media posts, and media coverage of past similar terrorist attack disasters such as the 2008 Mumbai attacks. To capitalize on internal validity, for each stimulus the disaster infor- mation was held as constant as possible across conditions. This consistency ensured the manipulated form and source factors were not confounded by information content variance. External validity was also addressed by creating stimuli that appeared as if

Liu et al. 633

they were screenshots from the various forms and sources; and, where possible, stim- uli included exact language pulled from the reviewed articles, press releases, and news reports mentioned above.

Participants and Procedures

A total of 2,015 participants via a random national sample completed the study online in June 2013. The research team hired a leading market research firm, GfK Custom Research, to recruit participants and administer the experiment. GfK drew the sample from their web-enabled KnowledgePanel®, a probability-based (ran- dom, non-volunteer access) panel designed to be representative of the U.S. popula- tion.1 The firm holds a U.S. patent for its selection methodology that ensures reliable U.S. representativeness,2 and its researchers have published scholarly articles on the topic of online panel methodology and metrics (e.g., Callegaro & DiSogra, 2008). GfK engages in random address sampling, and KnowledgePanel® includes mem- bers from cell-phone-only households as well as non-Internet households. Of the 3,077 panel members who were sent the invitation to participate, a response rate of 65.5% (N = 2,015) was obtained. Participants were randomly assigned to one of the experiment’s 12 conditions.3

Participant characteristics. Participants’ mean age was 49 years old. Age cohorts were determined based on U.S. Census age brackets, as has been done by others conducting similar research (e.g., Freberg, 2012). Specifically, 335 (16.6%) were in the 18-29 age group, 461 (22.9%) were in the 30-44 age group, 588 (29.2%) were in the 45-59 age group, and 631 (31.3%) were older than 60. There were 1,023 females (50.8%) and 992 males (49.2%). In terms of the race/ethic group composition, the majority were White (73.7%), with the remaining being Black (9.5%), Hispanic (9.8%), Other (non- Hispanic; 3.7%), and more than two races (non-Hispanic; 3.3%). With regard to states of residence, 19.8% of participants resided in Northeastern states, 22.4% in Midwest- ern states, 36.1% in Southern states, and 21.7% in Western states.

Measures

The dependent variables were behavioral intentions in response to disaster information if participants were experiencing the disaster: intended information-seeking behavior, intended communicative behavior, and intended protective-action-taking behavior.

Information-seeking behavior. A total of 13 disaster information-seeking behavior items, adapted from Austin, Liu, and Jin’s (2012) study, were presented for participants to respond to by indicating where they would look for information from or by if they were in the disaster situation, measured on a 7-point Likert-type scale where “1 = Strongly Disagree and 7 = Strongly Agree.” The 13 items were as follows: a local newspaper (M = 4.41, SD = 2.07), a national newspaper (M = 4.10, SD = 2.02), televi- sion (M = 6.10, SD = 1.46), local government websites (M = 4.83, SD = 1.94), federal government websites (M = 4.76, SD = 1.94), online videos (M = 3.55, SD = 1.98),

634 Communication Research 43(5)

Facebook page updates (M = 3.40, SD = 2.01), Twitter (M = 2.64, SD = 1.92), others’ blogs (M = 2.48, SD = 1.73), interpersonal communication such as talking to people they know via face-to-face and/or phone conversations (M = 5.24, SD = 1.80), email- ing people they know (M = 4.28, SD = 2.06), texting people they know (M = 4.44, SD = 2.14), and viewing pictures related to the disaster on a site dedicated to photo sharing (M = 3.56, SD = 2.13).

Information-sharing behavior. A total of 15 disaster information communicative behav- ior items, modified from Liu, Jin, and Austin’s (2013) study, were presented for par- ticipants to respond to by indicating what they would do with the disaster information if they were in the disaster situation, measured on a 7-point Likert-type scale where “1 = Strongly Disagree and 7 = Strongly Agree.” The 13 items included: “Like” the Face- book post (M = 2.54, SD = 1.84), “Retweet” at least one of the tweets (M = 2.58, SD = 1.92), “Email” the website link (M = 3.74, SD = 2.05) they just read to other people, tell people they know via face-to-face conversations about the disaster (M = 5.14, SD = 1.90), tell people they know by emailing them about the disaster (M = 34.43, SD = 2.05), call people they know by phone to talk about the disaster (M = 5.71, SD = 1.63), text people they now about the disaster (M = 4.78, SD = 2.13), “like” a govern- ment Facebook post about the disaster (M = 2.54, SD = 1.86), “share” a government Facebook post about the disaster on their own Facebook page (M = 3.00, SD = 2.05), comment on a government Facebook page about the disaster (M = 2.47, SD = 1.78), post information on their friends’ Facebook pages or groups about the disaster (M = 2.83, SD = 2.00), retweet a Twitter post about the disaster (M = 2.26, SD = 1.79), tweet about the disaster (M = 2.13, SD = 1.72), write a blog post on their own blog about the disaster (M = 1.91, SD = 1.51), post a comment on someone else’s blog about the disaster (M = 2.14, SD = 1.65), make a comment on someone else’s online video about the disaster (M = 2.16, SD = 1.68), and upload a picture related to the disaster on a dedicated photo sharing site (M = 2.48, SD = 1.88).

Taking protective actions. Five protective actions were presented for participants to respond to by indicating how likely or unlikely they would be to engage in each of five actions, measured on a 7-point Likert-type scale where “1 = Very unlikely and 7 = Very likely.” The five action items were as follows: (1) “I would evacuate from the area affected by the disaster no matter what” (M = 4.69, SD = 1.77); (2) “I would evacuate from the area affected by the disaster if instructed to evacuate by government officials” (M = 6.13, SD = 1.39); (3) “I would follow government instructions step by step” (M = 5.62, SD = 1.46); (4) “I would tell others to follow government instruc- tions” (M = 5.29, SD = 1.68); and (5) “I would listen for more information from gov- ernment sources” (M = 5.83, SD = 1.46).

Manipulation Checks and Data Analyses

The experiment included two sets of manipulation check items to determine whether the participants perceived the disaster information form and source as manipulated in the stimuli.

Liu et al. 635

Disaster information form. Participants were asked to respond to, “The disaster infor- mation was initially brought to my attention through a . . .,” selecting the number that best indicated their agreement with each of the three listed items (i.e., tweet, Facebook post, and website page), with 1 = Strongly Disagree and 7 = Strongly Agree. MANOVA found significant differences among all three crisis information forms (tweet condi- tion: F(2, 1984) = 260.87, p < .001, partial η2 = .21; Facebook post condition: F(2, 1984) = 147.80, p < .001, partial η2 = .13; website page condition: F(2, 1984) = 31.12, p < .001, partial η2 = .03). Therefore, the manipulation of disaster information form was successful.

Disaster information source. Participants were asked to respond to, “In this scenario, the detailed information about the disaster was provided by . . .,” selecting the number that best indicated their agreement with each of the four listed items (i.e., the USA Today online newspaper, FEMA, the San Francisco Chronicle online newspaper, and the San Francisco Mayor’s Office of Emergency Management), with 1 = Strongly Disagree and 7 = Strongly Agree. MANOVA found significant differences among all three crisis information forms (USA Today condition: F(3, 1981) = 240.88, p < .001, partial η2 = .27; FEMA condition: F(3, 1981) = 161.55, p < .001, partial η2 = .20; San Francisco Chronicle condition: F(3, 1981) = 183.36, p < .001, partial η2 = .22; Mayor’s Office condition: F(3, 1981) = 236.23, p < .001, partial η2 = .26). Therefore, the manipulation of disaster information source was successful. MANOVAs examined main and inter- action effects of disaster information form and disaster information source on several dependent measures.

Results

Before turning to analyses pertaining to the research questions, this section reports results of analyses examining baseline media consumption behaviors, which provide context. Participants rated how often they typically use traditional and social media communication tools, measured on a 7-point Likert-type scale (1 = not at all, 7 = all the time). As displayed in Table 1, participants most frequently used TV (M = 5.86, SD = 1.64), followed by face-to-face conversations (M = 5.63, SD = 1.66) and phone calls (M = 5.51, SD = 1.66). Participants least frequently used Twitter (M = 1.73, SD = 1.51), blogs (M = 1.97, SD = 1.50), and sites dedicated to photo sharing (M = 2.04, SD = 1.59).

Disaster Information Form and Source

Intentions to engage in information seeking. Research question 1 asked how disaster information form and source independently and/or jointly impacted participants’ com- municative behavioral intentions of information seeking via an array of communica- tion forms. Table 2, below, summarizes participants’ level of agreement with the statement that they would seek additional information using each respective tradi- tional or social media outlet. Television (M = 6.10, SD = 1.46), interpersonal

636 Communication Research 43(5)

communication forms (i.e., “Talking to people I know via face-to-face and/or phone conversations”; M = 5.24, SD = 1.80), and government websites (local: M = 4.83, SD = 1.94; federal: M = 4.76, SD = 1.94) received the highest agreement scores among participants.

Significant main effects of disaster information source were detected on partici- pants’ likelihood to seek further disaster information from TV, local government web- sites, and federal government websites. First, participants were more likely to seek

Table 1. Participants’ Typical Use of Media Communication Tools.

Media/communication type M SD

Television 5.86 1.64 Face-to-face conversations 5.63 1.66 Phone calls 5.51 1.66 Local newspapers 4.16 2.25 Text messages 4.07 2.38 Facebook 3.52 2.36 National newspapers 2.74 1.87 Online videos 2.77 1.78 Local government sites 2.44 1.58 Federal government sites 2.40 1.55 Sites dedicated to photo sharing 2.04 1.59 Blogs 1.97 1.50 Twitter 1.73 1.51

Table 2. Participants’ Disaster Information-Seeking Behavioral Intentions.

Likelihood of seeking additional information from communication outlets M SD

Television 6.10 1.46 Talking to people I know via face-to-face and/or phone conversations 5.24 1.80 Local government websites 4.83 1.94 Federal government websites 4.76 1.94 Texting people I know 4.44 2.14 A local newspaper 4.42 2.07 Emailing people I know 4.28 2.06 A national newspaper 4.10 2.02 Viewing pictures related to the disaster on a site dedicated to photo

sharing 3.56 2.13

Online videos 3.55 1.98 Facebook page updates 3.40 2.01 Twitter 2.64 1.92 Others’ blogs 2.48 1.73

Liu et al. 637

further information from TV when they received disaster information from national government (M = 6.19, SE = 0.08) than from local news media (M = 5.89, SE = 0.08; F(3, 1452) = 2.88, p = .035, partial η2 = .006). Second, participants were more likely to seek further information from local government websites when the initial disaster information came from local government (M = 4.93, SE = 0.10) rather than from national news media (M = 4.51, SE = 0.10; F(3, 1452) = 2.98, p = .031, partial η2 = .006). Third, participants were more likely to seek further information from federal government websites when the initial disaster information came from local govern- ment (M = 4.86, SE = 0.10) or FEMA (M = 4.83, SE = 0.10) rather than national news media (M = 4.42, SE = 0.10; F(3, 1452) = 4.27, p = .005, partial η2 = .009).

Although no significant main effects of information form were detected, informa- tion form was found to interact with information source to exert joint effects: Participants were most likely to seek further information from TV when the disaster information was from the national government source in the form of a web post (M = 6.27, SE = 0.13; F(6, 1452) = 3.24, p = .004, partial η2 = .013).

Intentions to share information. Research question 2 asked how disaster information form (social media vs. traditional media) and/or source (local government vs. national government vs. local news media vs. national news media) independently and jointly impacted participants’ communicative behavioral intentions of information sharing via an array of communication forms. Table 3, below, summarizes participants’ level of agreement with the statement that they would communicate in each manner if they were in the disaster scenario.

Overall, participants reported strong intentions to take interpersonal communica- tion approaches, such as engaging in phone conversations (M = 5.71, SD = 1.63), face- to-face conversations (M = 5.14, SD = 1.90), text messaging (M = 4.78, SD = 2.13), and emailing (M = 4.43, SD = 2.05). Neither the information form nor the source impacted participants’ reported intentions to engage in these or a host of other com- munication behaviors. Also, no significant effects of form or source were detected related to the communicative behavioral intention of information sharing.

Intentions to take protective actions. Research question 3 asked how disaster informa- tion form and/or source independently and jointly impacted participants’ behavioral intentions to take protective actions. Table 4, below, summarizes participants’ assess- ment of how likely they would be to take each protective action if they were in the disaster scenario presented in the experiment.

Participants reported the highest likelihood of evacuating from the area affected by the hypothetical disaster if instructed to evacuate by government officials (M = 6.13, SD = 1.39), followed by listening for more information from government sources (M = 5.83, SD = 1.47), and following government instructions step by step (M = 5.62, SD = 1.46).

The form through which the disaster information was initially communicated sig- nificantly influenced participants’ reported likelihood of evacuating from the disaster area “no matter what” (F(2, 1977) = 4.22, p = .015, partial η2 = .004) but not their

638 Communication Research 43(5)

reported likelihoods of taking any other protective actions. Specifically, when the hypothetical terrorist attack information came from a tweet (M = 4.79, SE = 0.07), participants reported significantly higher likelihoods of evacuating “no matter what” than when the information came from a website post (M = 4.53, SE = 0.07). No signifi- cant differences were detected in terms of the effects of a Facebook post (M = 4.76, SE = 0.07). The sources of disaster information did not differentially influence partici- pants’ reported likelihoods of evacuating “no matter what” or their reported likelihood

Table 3. Participants’ Intentions to Share Disaster Information.

Communication behavior M SD

Call people I know via telephone to talk about the disaster. 5.71 1.63 Tell people I know via face-to-face conversations about the

disaster. 5.14 1.90

Text people I know the disaster. 4.78 2.13 Tell people I know by emailing them about the disaster. 4.43 2.05 “Email” the website link I just read to other people.a 3.74 2.05 “Share” a government Facebook post about the disaster on

my Facebook page. 3.00 2.05

Post information on my friends’ Facebook pages or groups about the disaster.

2.83 2.00

“Retweet” at least one of the tweets I just read.b 2.58 1.92 “Like” the Facebook post I just read.c 2.54 1.84 “Like” a government Facebook post about the disaster. 2.54 1.86 Comment on a government Facebook page about the disaster. 2.47 1.78 Retweet a Twitter post about the disaster. 2.26 1.79 Make a comment on someone else’s online video about the

disaster. 2.16 1.68

Post a comment on someone else’s blog about the disaster. 2.14 1.65 Tweet about the disaster. 2.13 1.72 Write a blog post on my own blog about the disaster. 1.91 1.51

aFor the website post conditions only. bFor the tweet conditions only. cFor the Facebook post conditions only.

Table 4. Participants’ Likelihood to Take Protective Actions.

Protective actions M SD

I would evacuate from the area affected by the disaster if instructed to evacuate by government officials.

6.13 1.39

I would listen for more information from government sources. 5.83 1.47 I would follow government instructions step by step. 5.62 1.46 I would tell others to follow government instructions. 5.29 1.68 I would evacuate from the area affected by the disaster no matter what. 4.69 1.77

Liu et al. 639

of taking any other protective actions. No interaction effects between form and source were uncovered related to intentions to take any protective action.

Discussion

For decades, disaster sociologists have empirically validated that before at-risk publics decide whether to take protective actions, they engage in disaster information seeking and sharing, a process some call milling (Mileti & Beck, 1975; Mileti & Sorenson, 1990). Yet, scholars to date have neither sufficiently investigated publics’ preferences for disaster information seeking and sharing nor fully uncovered how communication via different information forms and sources might impact milling behaviors and moti- vations to take protective actions.

A noteworthy finding of this study is that upon initial exposure to information about the disaster, participants reported intentions to communicate about it predominately via interpersonal forms such as telephone calls, face-to-face conversations, texting, and emailing rather than through social media forms such as “liking,” “sharing,” or “commenting” on a government Facebook post. This showcases the importance of publics’ interpersonal communication with each other in making decisions about how to respond to disaster information rather than publics’ communication with organiza- tions through organizations’ online forms. Consequently, this finding supports calls for additional crisis communication research from a publics perspective to better under- stand how publics influence each other during disasters, rather than research on how organizations’ crisis communication influences publics, as is represented in the major- ity of research (Kim & Dutta, 2009). This audience-centric view of crisis communica- tion echoes Botan and Taylor’s (2004) call for a paradigm shift in strategic communication research, moving from functional approaches to more cocreational approaches with genuine interest and concrete emphasis on how publics create and share meaningful messages among themselves and with organizations. This finding also adds to the body of knowledge distinguishing the value of offline interpersonal communication (e.g., texting, calling, and talking face-to-face) from online interper- sonal interactions such as via social media. Prior research concluded that people tend to turn to information sources with which they have stronger prior relationships (Dellarocas, 2003; Goldsmith & Horowitz, 2006; Sen & Lerman, 2007), which is sup- ported by this study’s findings in a disaster context.

Overall, these findings provide additional avenues for future qualitative research to better understand why publics seem to prefer offline interpersonal communication forms (with the exclusion of email) to seek additional disaster information, perhaps also in part because of their media richness. In fact, these findings may be particularly enlightening when interpreted through the lens of media richness theory. Although the theory is usually examined in terms of explaining or predicting which particular media forms are chosen for information dissemination, nonetheless tenets of media richness can offer insights here. Namely, the theory asserts that richer communication forms are better, relative to leaner forms, at helping the information receiver overcome uncer- tainty or ambiguity (Daft & Lengel, 1984, 1986). Our data revealed that in the context

640 Communication Research 43(5)

of a highly uncertain disaster situation, participants favored sharing information via forms that could display greater richness, such as through multiple information cues and rapidity of exchanges, in comparison with their least favored forms. Although this study is not a formal test of media richness theory, the findings provide meaningful implications about how perceived richness of media might influence individuals’ com- munication form selections and source preferences in their further disaster information- seeking and sharing activities. As such, future formal tests of media richness theory would be valuable to further explore and quantify “richness” of social media platforms as compared with traditional media and offline communications during disasters. A refined definition and operationalization in the context of disaster communication will help researchers and practitioners more precisely evaluate the “richness” of not only information form as media forms but also information source as influential content creators.

In addition, the current data also hint at possible fit with emerging ICT succession theory (Stephens, 2007). Specifically, this theory might predict that following expo- sure to initial disaster information, people would respond better to additional informa- tion that engages different senses through varied modalities in order to mitigate the “overload” that can occur with multiple identical messages via identical communica- tion forms. Correspondingly, our experiment found that participants were significantly more likely to seek further disaster information via TV (a primarily audio-visual form) after the information initially came from a national government web post (a primarily text-based form). This is in keeping with prior research that showed that experiment participants who received a sequence of two messages through modality-expanding ICTs (cf. text-based then audio-visual based in the current research) exhibited higher evaluations of message effectiveness and increased behavioral intentions (Stephens & Rains, 2011).

Furthermore, the study’s findings caution crisis communication researchers against studying social media in isolation of other communicative behaviors during disasters as the study’s findings clearly indicate that social media were not the pre- ferred sources for information seeking and sharing in this nationally representative field experiment. Consequently, approaches such as the SMCC model that consider how multiple information forms and sources interact to influence how publics respond to disasters are promising explanatory models to guide future research and practice (e.g., Liu et al., 2011). Furthermore, future research may consider integrating the explanatory benefits of the SMCC model with the predictive benefits of media rich- ness and ICT succession theories. Such an integrated model could yield comprehen- sive explanations and predictions of how information form and source sequences independently and/or jointly influence publics’ communication and other behaviors during disasters. Future research could also integrate aspects of the SMCC model not tested here including (1) how disaster type and origin affect how publics respond to disaster information form and/or information source and (2) whether there are differ- ences by groups of publics such as those who purely consume disaster information produced by others versus those who create their own disaster information (Liu, Jin, Briones, & Kuch, 2012).

Liu et al. 641

Turning to our findings about source, the data suggest that researchers may want to broaden their scope to better understand how multiple organizations, publics, and groups of publics interpersonally communicate with each other and how that, in turn, may affect a variety of outcomes, including likelihood of taking protective actions, sharing disaster information, emotional responses, and attribution of responsibility. As Heath (2010) noted, the majority of research has focused on how a single organiza- tion’s crisis communication affects publics, which does not reflect the reality of the multiple information sources (and forms) available to the public during disasters. Here again is where expanded research on the SMCC model to move it from an explanatory model to a predictive model would be beneficial.

Furthermore, the present data provide a cautionary note against overrelying on a single source or a handful of sources to share disaster information, as well as point to the importance of having multiple forms and sources provide similar disaster informa- tion to publics. The data also contradict conventional wisdom among many emergency managers that local sources are more effective than national sources for motivating the public to respond to disaster information (e.g., Col, 2007; Perry, 2003), suggesting the need for experimental research to substantiate claims about cause and effect relation- ships between information source and communication effectiveness during disasters.

Finally, for theory development, researchers may want to build on the protective actions tested here as behavioral outcomes to expand what we know about crisis com- munication outcomes. As scholars have noted (e.g., Fediuk, Pace, & Botero, 2010), a primary limitation of crisis communication scholarship is the lack of well-developed and well-tested outcomes beyond attribution of responsibility. In this study, we found that participants were highly likely to report they would evacuate from the area affected by the hypothetical disaster if instructed by government officials to evacuate, which is a promising finding for emergency managers. In the future, researchers could test how likely publics are to comply with recommended protective actions that come from other sources such as non-profits active in disasters, corporations responding to disasters, and friends/family—and specifically whether government sources perform better than other sources in terms of motivating protective action taking. This may be especially important to empirically test, given that prior research suggested that pub- lics may view government sources as slower and less timely during disasters com- pared to other sources (Palen et al., 2010).

Limitations and Conclusion

Like all research, this study is limited by several factors. First, the experiment tested only one disaster type and thus the findings may not generalize to other disaster types. Also, the experiment tested behavioral intentions rather than actual behaviors, and behavioral intentions may not be accurate proxies for actual behavior. Furthermore, the experiment only tested three information forms and future research is needed to see if the findings hold up for additional forms. In addition, the linkage and difference between baseline media consumption and the use of different media forms at times of a disaster are yet to be further examined by future studies. Finally, the findings only

642 Communication Research 43(5)

apply to the United States, and crisis communication theory development needs to be tested internationally and cross-culturally.

In sum, while no single study can provide all of the information necessary to inform disaster communication management, building a chain of evidence can and should inform ongoing disaster communication efforts. Findings reported here provide com- pelling evidence for the importance of using multiple forms and sources of informa- tion to communicate with publics during disasters, as well as the importance of the publics’ communication with each other (rather than with organizations) in deciding how to respond to initial disaster information.

Authors’ Note

The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Department of Homeland Security or Study of Terrorism and Responses to Terrorism (START).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the U.S. Department of Homeland Security Science and Technology Directorate’s Homeland Security Advanced Research Projects Agency through Grant Award Number 2012ST061CS0001 made to the National Consortium for the Study of Terrorism and Responses to Terrorism (START).

Notes

1. Initially, participants are chosen scientifically by a random selection of telephone numbers and residential addresses. Persons in selected households are then invited by telephone or mail to participate in KnowledgePanel®. For those who agree to participate but do not already have Internet access, GfK provides at no cost a laptop and Internet service provider (ISP) connection. People who already have computers and Internet service are permitted to participate using their own equipment. Panelists then receive unique login information for accessing surveys online, and they are sent a limited number of emails throughout each month inviting them to participate in research. The experiment herein was one such research study that randomly selected panel members received a request to complete.

2. Knowledge Networks, a GfK company obtained U.S. Patent No. 7,269,570 in 2007 for its methodology for selecting multiple online survey samples from a research participant panel. Their selection methodology assures that multiple sequential KnowledgePanel® samples drawn from the finite panel membership will each reliably represent the U.S. population.

3. The breakdown of participants by condition is as follows: (1) Condition 1 had 157 par- ticipants, (2) Condition 2 had 174 participants, (3) Condition 3 had 165 participants, (4) Condition 4 had 157 participants, (5) Condition 5 had 164 participants, (6) Condition 6

Liu et al. 643

had 173 participants, (7) Condition 7 had 179 participants, (8) Condition 8 had 172 partici- pants, (9) Condition 9 had 159 participants, (10) Condition 10 had 166 participants, (11) Condition 11 had 170 participants, and (12) Condition 12 had 179 participants.

References

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

Bortree, D. S., & Seltzer, T. (2009). Dialogic strategies and outcomes: An analysis of envi- ronmental advocacy groups’ Facebook profiles. Public Relations Review, 35, 317-319. doi:10.1016/j.pubrev.2009.05.002

Botan, C. H., & Taylor, M. (2004). Public relations: State of the field. Journal of Communication, 54, 645-661. doi:10.1111/j.1460-2466.2004.tb02649.x

Brenner, J. (2013). Pew Internet: Social networking (full detail). Retrieved from http://pewin- ternet.org/Commentary/2012/March/Pew-Internet-Social-Networking-full-detail.aspx

Callegaro, M., & DiSogra, C. (2008). Computing response metrics for online panels. Public Opinion Quarterly, 72, 1008-1031. doi:10.1093/poq/nfn065

Carlson, J. R., & Zmud, R. W. (1999). Channel expansion theory and the experimental nature of media richness perceptions. Academy of Management Journal, 42, 153-170. doi:10.2307/257090

Chew, C., & Eysenbach, G. (2010). Pandemics in the age of Twitter: Content analysis of tweets during the 2009 H1N1 outbreak. PLoS ONE, 5(11), 1-13. doi:10.1371/journal. pone.0014118

Col, J.-M. (2007). Managing disasters: The role of local government. Public Administration Review, 67, 114-124. doi:10.1111/j.1540-6210.2007.00820.x

Coombs, W. T. (2012). Ongoing crisis communication: Planning, managing, and responding (3rd ed.). Thousand Oaks, CA: Sage.

Daft, R. L., & Lengel, R. H. (1984). Information richness: A new approach to managerial behav- ior and organization design. Research in Organizational Behavior, 6, 191-233.

Daft, R. L., & Lengel, R. H. (1986). Organisational design. Management Science, 32, 554- 571.

Dellarocas, C. (2003). The digitization of word of mouth: Promise and challenges of online feedback mechanisms. Management Science, 49, 1407-1424.

D’Urso, S. C., & Rains, S. (2008). Examining the source of channel expansion: A test of channel expansion theory with new and traditional communication media. Management Communication Quarterly, 21, 486-507. doi:10.1177/0893318907313712

Federal Emergency Management Agency. (2013, August 23). Disaster Reporter. Retrieved from http://www.fema.gov/disaster-reporter

Fediuk, T. A., Pace, K. M., & Botero, I. C. (2010). Exploring crisis from a receiver perspective: Understanding stakeholder reactions during crisis events. In T. Coombs & S. J. Holladay (Eds.), The handbook of crisis communication (pp. 635-656). New York, NY: Wiley- Blackwell.

Freberg, K. (2012). Intention to comply with crisis messages communicated via social media. Public Relations Review, 38, 416-421. doi:10.1016/j.pubrev.2012.01.008

Gilgoff, D., & Lee, J. J. (2013, April 15). Social media shapes Boston bombings response: Twitter and Facebook created national response, may help authorities. Retrieved from http://news.nationalgeographic.com/news/2013/13/130415-boston-marathon-bombings- terrorism-social-media-twitter-facebook/

644 Communication Research 43(5)

Goldsmith, R. E., & Horowitz, D. (2006). Measuring motivations for online opinion seeking. Journal of Interactive Advertising, 6(2), 3-14. doi:10.1080/15252019.2006.10722114

Heath, R. L. (2010). Crisis communication: Defining the beast and de-marginalizing key publics. In W. T. Coombs & S. Holladay (Eds.), The handbook of crisis communication (pp. 1-13). Malden, MA: Wiley-Blackwell.

Heath, R. L., & O’Hair, H. D. (2010). The significance of crisis and risk communication. In R. L. Heath & H. D. O’Hair (Eds.), Handbook of risk and crisis communication (pp. 5-30). New York, NY: Routledge.

Hughes, A. L., & Palen, L. (2009). Twitter adoption and use in mass convergence and emer- gency events. International Journal of Emergency Management, 6, 248-260. doi:10.1504/ IJEM.2009.031564

Jin, Y., & Liu, B. F. (2010). The blog-mediated crisis communication model: Recommendations for responding to influential external blogs. Journal of Public Relations Research, 22, 429- 455. doi:10.1080/10627261003801420

Jin, Y., Liu, B. F., & Austin, L. (2014). Examining the role of social media in effective crisis management: The effects of crisis origin, information form, and source on publics’ crisis responses. Communication Research, 41, 74-94. doi:10.1177/0093650211423918

Kim, I., & Dutta, M. J. (2009). Studying crisis communication from the subaltern studies frame- work: Grassroots activism in the wake of Hurricane Katrina. Journal of Public Relations Research, 21, 142-164. doi:10.1080/10627260802557423

Liu, B. F., Austin, L., & Jin, Y. (2011). How publics respond to crisis communication strate- gies: The interplay of information form and source. Public Relations Review, 37, 345-353. doi:10.1016/j.pubrev.2011.08.004

Liu, B. F., Jin, Y., & Austin, L. L. (2013). The tendency to tell: Understanding publics’ com- municative responses to crisis information form and source. Journal of Public Relations Research, 25, 51-67. doi:10.1080/1062726X.2013.739101

Liu, B. F., Jin, Y., Briones, R., & Kuch, B. (2012). Managing turbulence in the blogo- sphere: Evaluating the blog-mediated crisis communication model with the American Red Cross. Journal of Public Relations Research, 24, 353-370. doi:10.1080/10627 26X.2012.689901

Mileti, D. S., & Beck, E. M. (1975). Communication in crisis: Explaining evacuation symboli- cally. Communication Research, 2, 24-29. doi:10.1177/009365027500200102

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, U.S. Department of Energy.

Moy, P., McCluskey, M. R., McCoy, K., & Spratt, M. A. (2004). Political correlates of local news media use. Journal of Communication, 54, 532-536. doi:10.1111/j.1460-2466.2004. tb02643.x

National Science and Technology Council. (2005). Grand challenges for disaster reduction: A report of the subcommittee on disaster reduction. Washington, DC: National Science and Technology Council, Executive Office of the President. Retrieved from http://www.sdr. gov/docs/SDRGrandChallengesforDisasterReduction.pdf

Palen, L., Starbird, K., Vieweg, S., & Hughes, A. (2010). Twitter-based information distribu- tion during the 2009 Red River Valley Flood Threat. Bulletin of the American Society for Information Science and Technology, 36(5), 13-17. Retrieved from http://www.asis.org/ Bulletin/Jun-10/JunJul10_Palen_Starbird_Vieweg_Hughes.pdf

Palen, L., Vieweg, S., Liu, S., & 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, 467-480. doi:10.1177/0894439309332302

Liu et al. 645

Palenchar, M. (2010). Historical trends of risk and crisis communication. In R. L. Heath & D. O’Hair (Eds.), Handbook of risk and crisis communication (pp. 31-52). New York, NY: Routledge.

Perry, R. W. (2003). Incident management systems in disaster management. Disaster Prevention and Management, 12, 405-412. doi:10.1108/09653560310507226

Pew Internet. (2013). Social media update 2013. Retrieved from http://www.pewinternet.org/ Reports/2013/Social-Media-Update.aspx

Reynolds, B., & Seeger, M. (2012). Crisis and emergency risk communication. U.S. Department of Health and Human Services Centers for Disease Control. Retrieved from http://emer- gency.cdc.gov/cerc/

Rice, R. E., D’Ambra, J., & More, E. (1998). Cross-cultural comparison of organizational media evaluation and choice. Journal of Communication, 48(3), 3-26. doi:10.1111/ j.1460-2466.1998.tb02757.x

Schultz, F., Utz, S., & Göritz, A. (2011). Is the medium the message? Perceptions of and reac- tions to crisis communication via twitter, blogs and traditional media. Public Relations Review, 37, 20-27. doi:10.1016/j.pubrev.2010.12.001

Sen, S., & Lerman, D. (2007). Why are you telling me this? An examination into negative con- sumer reviews on the Web. Journal of Interactive Marketing, 21(4), 76-94. doi:10.1002/ dir.20090

Stephens, K. K. (2007). The successive use of information and communication technologies at work. Communication Theory, 17, 486-509. doi:10.1111/j.1468-2885.2007.00308.x

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, 230-251. doi:10.1111/hcre.12002

Stephens, K. K., & Rains, S. A. (2011). Information and communication technology sequences and message repetition in interpersonal interaction. Communication Research, 38(1), 101- 122. doi:10.1177/0093650210362679

Stephens, K. K., Sørnes, J. O., Rice, R. E., Browning, L. D., & Saetre, A. S. (2008). Discrete, sequential, and follow-up use of information and communication technology by experienced ICT users. Management Communication Quarterly, 22, 197-231. doi:10.1177/08933189- 08323149

Sutton, J., Spiro, E. S., Johnson, B., Fitzhugh, S., Gibson, B., & Butts, C. T. (2013). Warning tweets: Serial transmission of messages during the warning phase of a disas- ter event. Information, Communication & Society, 17, 765-787. doi:10.1080/13691 18X.2013.862561

Sutton, J., Spiro, E. S., Johnson, B., Fitzhugh, S., Gibson, B., & Butts, C. T. (2014, May). Terse message amplification in the Boston bombing response. Paper presented at Proceedings of the 11th International ISCRAM Conference, University Park, Pennsylvania, USA.

Taylor, M., & Perry, D. (2005). Diffusion of traditional and new media tactics in crisis commu- nication. Public Relations Review, 31, 209-217. doi:10.1016/j.pubrev.2005.02.018

Trevino, L. K., Daft, R. L., & Lengel, R. H. (1990). Understanding managers’ media choices: A symbolic interactionist perspective. In J. Fulk & C. W. Steinfield (Eds.), Organizations and communication technology (pp. 71-94). Newbury Park, CA: Sage.

Utz, S., Schultz, F., & Glocka, S. (2013). Crisis communication online: How medium, cri- sis type and emotions affected public reactions in the Fukushima Daiichi nuclear disaster. Public Relations Review, 39, 40-46. doi:10.1016/j.pubrev.2012.09.010

Westerman, D., Van Der Heide, B., Klein, K. A., & Walther, J. B. (2008). How do people really seek information about others?: Information seeking across Internet and traditional

646 Communication Research 43(5)

communication channels. Journal of Computer-Mediated Communication, 13, 751-767. doi:10.1111/j.1083-6101.2008.00418.x

Wogalter, M. S. (2006). Communication-human information processing (C-HIP) Model. In M. S. Wogalter (Ed.), Handbook of warnings (pp. 51-61). Mahwah, NJ: Lawrence Erlbaum.

Wright, D. K., & Hinson, M. (2009). An updated look at the impact of social media on public relations practices. Public Relations Journal, 3(2), Article 2.

Author Biographies

Brooke Fisher Liu is an associate professor in the Department of Communication, University of Maryland and director of the Risk Communication and Resilience Program at the National Consortium for the Study of Terrorism and Responses to Terrorism (START), specialized in risk and crisis communication research.

Julia Daisy Fraustino is a PhD student in the Department of Communication, University of Maryland and affiliated with START, specialized in risk and crisis communication research. In August 2015, she joined the Reed College of Media, West Virginia University as an assistant professor.

Yan Jin is an associate professor and associate director for the Center for Health and Risk Communication in the Grady College of Journalsim and Mass Communication, University of Georgia and affiliated with START, specialized in risk and crisis communication research.