MASTER’S THESIS
Dynamic framing in times of crisis
Investigating frame building and frame alignment of news media and the public in a violent crime crisis
University of Amsterdam Graduate School of Communication Master in Political Communication (MSc.)
Supervisor: Dr. Marjolein Moorman
Student: Thilo Schröder Student ID: 11190043 Date of completion: 28 June 2017
Abstract
In times of crisis, communications are an important means for news media and the public to forge an understanding of the crisis and to negotiate crisis meaning. However, so far, remarkably little is understood about how these actors (collectively) make sense of crisis situations. This study aims to advance the understanding of these communicative processes by focusing on the concepts of frame building and frame alignment. The 2016 Munich shooting crisis, one of the most severe violent crimes in recent German history, offers a suitable test case. A method innovation of semantic-network analysis is conducted to automatically identify frames present in German public social media manifestations (N = 3256) and news articles (N = 325) over different crisis stages. First, the study documents the dynamic character of crisis framing amongst news media and the public over time. Second, results reveal that the news media and the public align their frames after initially deviating from one another in the frame building process. The actors’ frames eventually de-align again, restoring actor-specific framing patterns. The results imply that the actors involved in collective sensemaking to drive the communicative resolution of the crisis. They do so by coming to a common understanding, thereby potentially signalling the communicative closure of the crisis. The study contributes to the growing body of literature in crisis communication that emphasizes the complex dynamics of crisis-meaning construction and framing between actors.
Introduction
Societies are frequently confronted with crisis situations that inflict major disruptions upon public life, increase uncertainty and endanger public safety. Major violent attacks, like mass shootings, belong to such events that hit societies unexpectedly. Violent crimes foremost affect the public, endangering citizens’ safety and emotionally touching them (Liu, Austin & Jin, 2011). At the same time, mass shootings are an attractive item for news media (Lerbinger, 2011), triggering multiple news values like unexpectedness, ambiguity and conflict (Galtung & Ruge, 1965; Harcup & O’Neill, 2001).
Facing a rapidly unfolding violent attack, news media and the public may therefore perceive the situation as a crisis and try to socially (co-)construct its meaning in their
communications (Schultz & Raupp, 2010; Schultz, Kleinnijenhuis, Oegema, Utz & Attenveld, 2012). In the context of today’s mediatized society, the actors’ communications are thus an important mean of influencing the magnitude, evolution and consequences of a crisis (e.g. Kleinnijenhuis, Schultz, Utz & Oegema, 2015). To understand their roles in times of crisis, it is thus crucial to develop inquiries about how news media and the public make sense of them.
To advance our understanding of these sensemaking processes, framing theory offers a powerful body of literature to study the actors’ crisis communication patterns over time (Entman, 1993; Hellsten, Dawson & Leydesdorff, 2010). Frames highlight certain aspects of the actors’ understanding of the crisis in their communications (Entman, 1993), implying that they can thus also direct attention to certain aspects and withhold others (McCombs, 2005).
Studying frame building, i.e. which frames the actors construct, can therefore help gain crucial insights into how both actors make sense of a crisis (Schultz et al., 2012). News media play a pivotal role in making information on a crisis visible to the public, not only by selecting what to include in a news story, but also by framing the crises under their
professional routines and context (Holland, Sweet, Blood & Fogarty, 2014). On the other side, the public has gained significant power in framing crises online through the advent of social
media in recent years (Neuman, Guggenheim, Mo Jang & Bae, 2014), most notably in close temporal proximity to quickly evolving events (Sung & Hwang, 2014). In this sense, studying public frames is relevant because they may contest media frames by introducing different attributes of the crisis based on people’s personal values (Liu, 2010). Given these different social roles, the relevant question arises how, if at all, news media and the public distinctly make sense of crisis situations. Beyond that, the question gains in relevance considering the dynamic character of a crisis: Since temporal stages of crisis are associated with certain communicative functions (e.g. decreasing uncertainty, settling responsibility; see Reynolds & Seeger, 2005), some frames may resemble these functions more than others and therefore fluctuate or prevail in salience over time. Thus, it is crucial to know how constructed frames of news media and the public evolve per actor over time.
To advance the understanding of how news media and the public may collectively make sense of a crisis, the concept of frame alignment is useful (Snow, Rochford, Worden & Benford, 1986; van der Meer, Verhoeven, Beentjes & Vliegenthart, 2014). In the context of crisis research, frame alignment can be understood as the temporary rapprochement in the actors’ framing (van der Meer et al., 2014). The temporary similarity in frame building can denote the actors’ common interpretations about the events (Snow et al., 1986; Hellsten et al., 2010). This can have essential implications for the trajectory of the crisis, as the alignment may provide issue coherence and indicate the communicative end of a crisis (van der Meer et al., 2014), outlining its relevance as a central construct to understand collective sensemaking.
However, despite the outlined relevance, extant crisis communication research has deficits when it comes to describing the dynamic framing process of news media and the public. First, although framing analysis has a prominent history in crisis communication research, the field has been dominated by organization-centric studies (Coombs, 2010; Bundy, Pfarrer & Coombs, 2011), widely neglecting the role of news media and the public in crisis framing (Kleinnijenhuis et al., 2015). Despite its gain in framing power, especially the public
has been widely overlooked as a relevant actor (Wigley & Fontenot, 2010). Additionally, news media have predominantly been studied in terms of their unidirectional effects on recipients rather than how they frame a crisis (e.g. Coombs & Holldaday, 2008; Schultz, Utz & Göritz, 2011). Given this lacking knowledge about the frame building process of news media and the public, inductive research is needed to establish present frames of these actors. Second, while crisis literature acknowledges the dynamic nature of crisis evolution and resolution in temporal crisis stages (Fink, 1986; Coombs, 2010), scholars note a lack of
research concerning the frame development of actors over time in these aspects (Schultz et al., 2012; Sung & Hwang, 2014). Hence, it remained largely unclear how the news media’s and the public’s frames evolve over different stages of a crisis. Third, scholars have just begun to describe patterns of crisis frame alignment between news media and the public. Studies have thus far linked alignment between news media and the public either solely qualitatively (van der Meer & Verhoeven, 2013) or in organizational crises to each other (van der Meer et al., 2014). Therefore, the quantitative validation of frame alignment in a public crisis case remains absent from the literature.
Filling these research gaps, this study aims to provide insight into the dynamic framing process of the news media and the public by addressing their frame building processes and frame alignment patterns during a violent crime crisis. Therefore, the research question guiding this study reads as follows: Which frames are present in the crisis communications of
news media and the public, how do they evolve over time and do they align?
To appropriately address this research question, a method innovation of
semantic-network analysis (Hellsten et al., 2010; Vlieger & Leydesdorff, 2012) was applied on the 2016 Munich shooting to study the dynamic frame building processes of German news media and public social media manifestations.
Theoretical Framework Theoretical context of crisis literature
In line with the dominance of organization-centric studies in the field of crisis communication, research has conceptualized crises most commonly as the “perception of an unpredictable event that threatens important expectancies of stakeholders and can seriously impact an organization’s performance and generate negative outcomes” (Coombs, 2010, p. 19). Consequently, a significant amount of studies has inquired how affected organizations can manage and respond to crisis situations effectively (Coombs & Holladay, 2008; Claeys & Cauberghe, 2014), how crises affect organizational reputation (Coombs, 2007; Schultz et al., 2012) or financial performance (van der Meer & Vliegenthart, 2017).
However, it can be argued that organization-centred approaches fail to adequately describe the context surrounding some complex crises in society, in which crisis meaning is not merely the concern of a single actor. Mass shootings present such a case, as they are public crisis events that constitute major sources of disruption and uncertainty in public life (Falkenheimer & Heide, 2010; Bundy et al., 2011). Given the high magnitude of mass shootings (Boin, Hart & McConnell, 2009) and the participation of diverse social actors in public life (Reese, Gandy & Grant, 2001), violent crimes are likely to be perceived as a crisis by different actors. In other words, such crises are, as Falkenheimer and Heide (2010) note, “social, political, and cultural phenomena: a crisis is a crisis due to the fact that different groups, interested parties, and institutions perceive and experience it as a crisis” (p. 514).
This study addresses the public and news media as affected actors in the 2016 Munich shooting crisis. Since mass shootings pose immediate threats on citizens’ safety (Boin & Hart, 2003), disrupt community cohesion (Wormwood, Lynn, Feldman Barrett & Quigley, 2016) and raise issues of public accountability to citizens (Boin, Stern & Sundelius, 2016), the public is one of the actors in society to be directly affected by a violent crime crisis. In crisis situations, the news media are largely expected to play a social role in providing relevant
information on the unfolding event as a response to these severe public issues (Graber & Dunaway, 2014), making crisis information both available and interpretable (Sorribes & Rovira, 2011) As the public, and hence news media’s customers, demand rapid and vast information on an unfolding crisis event (Liu & Kim, 2011), news media are hence a likely affected actor during violent crime crises.
Based on their respective perceptions of reality, crises are socially constructed by the affected actors, who make their perceptions of reality salient in their communications. (Heath, 2010). Furthermore, crises are eventually deconstructed by these affected actors, who resolve the respective crises by responding to and negotiating over crisis-meanings with each other (van der Meer et al., 2014). In the context of a violent crime crisis, this implies that news media and the public use complex interdependent communications to negotiate over crisis-meaning in order to resolve the crisis (Schultz & Raupp, 2010; Liu & Kim, 2011; Schultz et al., 2012).
Theoretical context of framing literature
The concept of framing provides a suitable theoretical framework for the inquiry into how different actors in society make sense of crisis situations. Being applied in multiple disciplines (Van Gorp, 2007; Borah, 2011), framing theory has been subject to theoretical inconsistencies because of diverging conceptualizations and different theoretical assumptions of the disciplines (Scheufele, 1999). Yet, framing has become one of the most researched concepts in communication science in the last 25 years (Vliegenthart & van Zoonen, 2011).
In essence, framing research is concerned with how actors in society attribute meaning to issues by highlighting certain aspects of them in their communication, i.e. by increasing the salience of these aspects to the recipients of their communication. In his classic definition, Entman (1993) defines framing as “select[ing] some aspects of a perceived reality and
mak[ing] them more salient in a communicating text, in such a way as to promote a problem definition, causal interpretation, moral evaluation, and/or treatment recommendation” (p. 52). Because issues can be subject to manifold perceptions of reality, the framing of issues can be similarly constituted by multiple attributions of social reality, i.e. by different frames. In short, frames can be referred to as “conceptual tools which media and individuals rely on to convey, interpret, and evaluate information” (Neuman, Just & Crigler, 1992, p. 60).
Supplementary to the outlined actor-centric paradigm, framing can also be conceptualized on a communication level that emphasizes the complex communicative interactions by which actors negotiate crisis meaning (Hellsten et al., 2010). In this respect, framing can be defined as a “set of discourses that interact in complex ways within and among domains” (van der Meer et al., 2014, p. 752). The underlying idea of this approach is that frames are semantically constituted in language and their physical expression as words in sentences (Jonkman & Verhoeven, 2013). The entirety of texts by actors can be seen as a discourse of the actor (van der Meer et al., 2014), in which patterns of words can occur across documents, between actors and over time (Leydesdorff & Hellsten, 2006). This is in line with Entman’s (1993) theoretical assertion that “frames in text are manifested by the presence or absence of certain keywords, stock phrases, stereotyped images, sources of information, and sentences that provide thematically reinforcing clusters of facts or judgments” (p. 52). Hence, meaning can be attributed to these word networks by tracing to what extent words appear with other words in these semantic structures as ‘implicit frames’ (Hellsten et al., 2010).
Crisis Communication and framing
Communication, and framing play a significant role in the evolution and resolution of crises. Communication can be used to increase the salience of issues (Chong & Druckman, 2007), and hence to bring the actors’ perception of the crisis to different agendas and thus emphasize the magnitude and consequences of unfolding events (McCombs, 2005). In the
context of crisis situations, framing refers to how the crisis is presented. In the actors’ communications, frames are part of the actors’ meaning construction and sensemaking
processes (Weick, 1988), assisting individuals to construct meaning of an ambiguous situation (Benford & Snow, 2000), and thus helping actors to form an initial understanding of the crisis (Utz, Schultz & Glocka, 2013). Based on the level of the actors’ mutual comprehension of the situation, frames can therefore cohere or dispute issues in crisis situations (Hellsten et al., 2010) and influence the perception of uncertainty amongst affected actors (Liu & Kim, 2011, van der Meer & Verhoeven, 2013).
News media framing
News media are considered a crucial actor for making crises highly visible and salient amongst the public (Chyi & McCombs, 2004), offering interpretations of the crisis to their audiences (An & Gower, 2009) and guiding the public with action recommendations (Chyi & McCombs, 2004). Maintaining relationships with diverse sources and serving broad public audiences, news media are traditionally considered powerful actors in shaping the evolution and magnitude of a crisis (Schultz et al., 2012) and influencing public opinion (McCombs, 2005).
Crisis situations disrupt routine news practices (Utz, Schultz & Glocka, 2012), yet journalists uphold their routine source selection principles during them (van der Meer et al., 2016). In their framing during a crisis, news media commonly show proximity to official sources (Olsson, Nord & Falkheimer, 2015), especially if the official crisis communicator is perceived as credible (Masters & Hart, 2012). In line with this most studies point out the dominance of ‘information frames’ (van der Meer & Verhoeven, 2013; van der Meer et al., 2014) or ‘descriptive frames’ in the news media’s framing during crises (Olsson & Nord, 2014; Falkheimer & Olsson, 2015; Zeng, Zhou & Li, 2015), which tend to closely resemble official information pertaining to the crisis. In contrast, An and Gower (2009) find news
media to frame crises predominantly based on the attribution of responsibility, while the use of other frames (such as human interest or conflict) strongly depended on the type of the crisis.
Public framing
With the rise of social media usage in recent years (Pew Research Center, 2016), the public has gained significant influence in the framing of issues online (Neuman et al., 2014). Due to their real-time communication structure, social media are found to play a particularly crucial role right after a crisis occurs: Relatively seen, social media posts account for the most reports in the initial development phase of a crisis (Sung & Hwang, 2014). Despite the fact that social media are hybrid tools also used by news media and other actors (Sung & Hwang, 2014), unaffiliated citizens account for the by far largest share of users, who additionally produce the vast majority of social media reports in crisis situations (Heverin & Zach, 2010). Although social media tools are not demographically representative of general publics, they are important tools for large parts of the public to exert considerable influence on public opinion (Neuman et al., 2014).
In its communications about the crisis, the public is constrained far less by time and cost than the news media (Neuman et al., 2014). Nor is it as much governed by regulations (Palen et al., 2009). In crisis situations, people particularly turn to social media to gather real-time information on crisis-development (Liu & Kim, 2011) to assess the relevance of the crisis and to vent their emotions (Liu et al., 2011). In line, the few existing findings on public social media frames during crisis situations predominantly find these frames to be associated with personal values of the public: Social media publics frame crises in terms of warning functions (Liu et al., 2016), personal speculations about the event (van der Meer &
Verhoeven, 2013) and their expressions of scepticism and criticism towards other actors (van der Meer & Verhoeven, 2013; van der Meer, 2014).
In sum, we know little about the role of the news media and the public in crisis framing so far, yet the preceding discussion documents existing literature, leading to following research question: RQ1: Which frames do news media and the public construct
during the crisis?
Crisis development
To describe the process of crisis evolvement and resolution in terms of temporal dimensions, scholars have identified various stages of communications during a crisis (Fink, 1986; Coombs, 2010; Graber & Dunaway, 2014). The main premise of splitting a crisis into different stages is that crises are not ‘one-moment events’ (Nikolaev, 2010), i.e. crises are dynamic processes in which they evolve and resolve. Hence, this process can be referred to as a ‘crisis life-cycle’ (Fink, 1986; Nikolaev, 2010), which serves as a standard reference in describing the dynamic character of crisis evolvement and resolution (Coombs, 2010;
Nikolaev, 2010). Despite the fact that the crisis life-cycle model has been predominantly used to test reactive communication strategies, i.e. crisis response or crisis defence strategies (e.g. Marra, 1998; Cloudman & Hallahan, 2006), it acknowledges the necessity to also investigate communicative interaction between crisis actors (Coombs, 2010).
Following Fink (1986), the ‘crisis life-cycle’ approach can be categorized into three time stages.1 First, as a crisis event unfolds, the acute stage constitutes the ‘point of no return’ after which the actual damage has occurred (Fink, 1986) and which requires the immediate attention of all affected actors (Coombs, 2010). At this stage, rapid communications are established to decrease uncertainty and emotional turmoil amongst the involved actors (Reynolds & Seeger, 2005). Second, the chronic stage of a crisis constitutes the preliminary end of the incident itself, yet it is defined by lingering concerns of consequences of, and responsibility within the crisis (Fink, 1986). Extended communication between the actors takes place to create a more accurate understanding of the crisis and foster certainty
(Reynolds & Seeger, 2005). Third, the crisis resolution stage indicates the actors’ return to normal operations after the crisis responsibility is settled (Fink, 1986). Thus, it can be best characterized as the transition to a post-crisis, in which actors respond to the ongoing crisis closure by facilitating risk prevention and drawing consequences from the solved crisis (Reynolds & Seeger, 2005).
Actor-specific crisis framing
Because actors, such as news media and the public, have different functions in society (McQuail, 2010) and personal or professional values (van der Meer et al., 2014), they are likely to differ in the frames they produce about an issue (Chong & Druckman, 2007).
Arguably, this results in actor-specific patterns of frames that may conflict between the actors (Chong & Druckman, 2007). In other words, diverging frames form the basis for the actors’ interdependent communications, which can be seen as their collective process of negotiation over meaning (Vliegenthart & van Zoonen, 2011).
The main communicative function of the acute crisis stage among the actors is to call attention the unfolding event and its consequences (Coombs, 2010). As sensemaking in this immediate evolvement of the crisis needs to occur rapidly, actors firstly need to decrease uncertainty in their specific context to create an elementary understanding about the events (Weick, 1988). In the context of this study, the public may therefore base their sensemaking on their personal interpretations of the crisis, whereas news media are rather assumed to pursue professional (i.e. journalistic) functions. Hence, the actors are more likely to assess crisis meaning self-referential immediately after a crisis occurs, visible by the production of actor-specific frames (Schultz & Raupp, 2010).
Empirical evidence supports the initial sensemaking processes for different actors immediately after a crisis occurs, such as organizations and governments (Schultz & Raupp, 2010) or news media and PR (Schultz et al., 2012). Notably, van der Meer & Verhoeven
(2013) found that the Dutch public differed in their initial crisis framing of a chemical plant explosion in 2011 from the news media: While the public initially framed the incident on social media as a terrorist attack, the news media dominantly used an information frame in their coverage (van der Meer & Verhoeven, 2013). Due to these different functions of news media and the public, initial frame alignment is not to be expected shortly after a crisis occurs. Therefore, the following hypothesis is proposed:
H1: In the acute stage of the crisis, frames of news media and the public are not aligned.
Frame alignment during crisis
Framing is part of an interactive discursive process (Benford & Snow, 2000), in which meaning can be transferred between the communications of actors (Hellsten et al., 2010). Once the actor-specific frames of different actors are constructed, they are subject to negotiation over the crisis meaning on the different media platforms (van der Meer et al., 2014). Thus, meaning can be transferred in the communicative interplay between the actors (Snow et al., 1986) by utilizing similar or dissimilar manifestations in the form of implicit frames (Hellsten et al., 2010). Hence, the concept of frame alignment can be conceptualized as the similarity in frame construction between different actors (van der Meer et al., 2014).
Crises are disruptive and system-destabilizing events (Falkenheimer & Heide, 2010), in which affected actors need to come to a mutual understanding of the crisis to not intensify the negative effects (Weick, 1988) and to re-stabilize the relationships between the actors (van der Meer et al., 2014). Given the outlined initial differences in framing, actors arguably need to achieve issue coherence by aligning their meaning provisions, i.e. their frames (Hellsten et al., 2010, van der Meer et al., 2014).
Relating this to crisis development, the settlement of responsibility and removal of lingering concerns are the main communicative functions in the chronic crisis stage (Coombs, 2010), which outlines the need to come to a mutual understanding among the actors. Hence,
extended communication between the affected actors is crucial to eventually bring coherence to the issue and to drive the resolution of the crisis (Weick, 1998, Hellsten et al., 2010). To achieve this collective sensemaking, actors therefore need to decrease their initial variability in frames, eventually resulting in frame alignment (van der Meer et al., 2014).
Such frame alignment patterns during crisis events have been found in various empirical case studies, such as the H1N1 crisis (Liu & Kim, 2011), the BP oil spill crisis (Schultz et al., 2012) and the Dutch Moerdijk crisis (van der Meer & Verhoeven, 2013). Most notably, van der Meer et al. (2014) found systematic patterns of frame alignment on four different crises, indicating the structural necessity of frame alignment in crisis situations.
Thus, the following hypothesis is proposed:
H2: In the chronic stage of the crisis, frames of news media and the public have aligned
compared to the acute stage.
Restoration of actor-specific frame building
Because issues are subject to ongoing reassessments and renegotiations by the
different actors (Snow et al., 1986), the resolution of a crisis terminates the necessity of issue coherence amongst affected actors (van der Meer et al., 2014). In other words, as uncertainty and disruption decrease, actors arguably return to assess the issues self-referentially based on their identities and social functions. Accordingly, it gets more likely that actors alter their communications to actor-specific frames again (van der Meer et al., 2014).
The resolution crisis stage does not imply extensive discursive processes between the actors anymore, since the crisis has been solved (Nikolaev, 2010), but rather assumes actor-specific facilitation of risk prevention (Reynolds & Seeger, 2005). Although risk prevention can also be discursive between actors (e.g. discussions), the re-establishment of the actors’ routine practices particularly increases their likelihood of framing the issue based on their respective personal and professional identities again (van der Meer et al., 2014).
Initial empirical evidence supports the restoration of actor-specific framing patterns in the resolution stage of a crisis. First, van der Meer et al. (2014) found systematic de-alignment of frames by different actors during the resolutions of multiple Dutch crises. Second, van der Meer and Verhoeven (2013) found that public and news media de-aligned their frames when the Dutch Moerdijk crisis resolved. While both actors came to a mutual understanding of the crisis beforehand, the public personalized the crisis with a government criticism frame during crisis resolution, whereas the news media dominantly used a safety information frame (van der Meer & Verhoeven, 2013). Given the previous review, the following hypothesis is proposed:
H3: In the resolution stage of the crisis, frames of news media and the public have de-aligned
compared to the chronic crisis stage
Method Data Collection
To empirically test the frame building and frame alignment process of the news media and the public during a violent crime crisis, an automated content analysis was conducted. According to the outlined research interest, several criteria were applied to find a suitable test case: (1) A public crisis event (rather than an organizational crisis) was requested, (2) which evoked considerable news media and public reactions (3) in a German-speaking country (4) in the recent years 2016-2017. Scanning for such events and evaluating media coverage and public reactions, the 2016 Munich shooting was eventually chosen as a suitable test case. The case was one of the severest violent crimes in Germany’s recent history, killing 10 people and injuring dozens (see a history of the crisis in Appendix 1). Accordingly, the case lead to notable news coverage and evoked considerable social media reactions (von Nordheim, 2016), eliciting a sufficient data situation that some cases in earlier studies lacked (van der Meer et al., 2014). Furthermore, the selection of a crisis from a German-speaking country contributes to the methodological aspect of semantic-network analyses, as the present study is one of the first to conduct a study for the German language (cf. Strauß & Vliegenthart, 2017)
Since computer-assisted methods of automated content analysis allow for the analysis of vast amounts of data, it was aimed to reflect the German news media discourse on the crisis extensively. Therefore, a census of available German-language regional and national, online and offline, news articles on the incident were collected using the database LexisNexis.2 Duplicates and press agency reports were excluded to solely include unique public news media articles. Social media manifestations of news media were not collected so as to obtain a robust variance of words for the analysis. Articles unrelated to the incident were removed from the selection after manually checking for the topic of each article.
Twitter was chosen as a suitable social media platform to collect public
communications on the crisis. During crisis situations, Twitter is the most commonly used social media tool by publics to disseminate short reports online (Heverin & Zach, 2010; Sung & Hwang, 2014). Citizens account for about 90% of users on Twitter that produce the vast majority of social media reports in crisis situations (Heverin & Zach, 2010). Furthermore, around 88% of Twitter users have public profiles (Twitter, 2016), allowing for a more valid data collection compared to social media platforms with predominantly private user profiles, such as Facebook (Dey, Jelveh & Ross, 2012).
Social media posts were collected via the social media monitoring tool Costoo. The tool is connected to Twitter’s Firehose, allowing access to Twitter posts in retrospect. Since Twitter’s API is usually only available in real-time, this procedure offers a great opportunity to investigate a large amount of Twitter data on an existing crisis. Tweets were filtered for German language and the hashtag ‘München’, which constitutes the most frequently used identifier on the incident (von Nordheim, 2016). A census of available tweets was obtained and checked for crisis-unrelated posts, which were manually removed.
Research units
In order to empirically assess the temporal dimension of the research questions, the actors’ communications were separated into different time stages that served as the research units for the analysis. Following the criteria of Fink (1986), crisis stages were identified based on the crisis life-cycle model. The acute stage of crisis breakout was operationalized from the beginning of the shooting until investigators announced the end of public danger. The chronic
stage was set following this announcement until initial closure over the case was declared,
which settled the responsibility for the attack. Finally, the resolution stage was
closure. Table 1 provides an overview over the different time stages and the collected data for these stages.
Table 1. Research units and number of observations Crisis stage Time N news
articles N public Description Acute 22 / 07 / 2016, 17:50 CET - 23 / 07 / 2016, 11:53 CET
75 2691 The shooting occurs and the perpetrator flees. Rumours spread about more perpetrators and shooting sites. Later, it transpires that a single shooter allegedly shot himself dead.
Chronic 23 / 07 / 2016, 11:54 CET - 24 / 07 / 2016, 12:00 CET
87 270 Investigators confirm the identity of the shooter and assume the act as “a classic killing spree.” They emphasize that motive and circumstances remain unclear and under investigation. Resolution 25 / 07 / 2016
12:01 CET - 28 / 07 / 2016 12:00 CET
163 289 Investigators announce initial closure over the case, as they found more evidence supporting a mental illness of the perpetrator. Prosecutors push forward investigations on confidants.
N total 325 3256
Semantic-network analysis
This study uses an automated content analysis to inductively identify so-called implicit frames in the communications of news media and the public (e.g. Hellsten et al., 2010). More specifically, the investigation of implicit frames is possible by means of an automated semantic-network analysis. The premise of this approach is that the meaning of words is not only provided through word choice but also through the relative position of words in word networks (Leydesdorff & Hellsten, 2006). Since frames are carriers of meaning (Entman, 1993), implicit frames can be operationalized as networks of words that can co-occur as latent dimensions in communication (Hellsten et al., 2010). As such, implicit frames are “created from spurious correlations between word (co)occurrences in
communications” (Hellsten et al., 2010), which gain meaning through the semantic context in which they occur (Leydesdorff & Hellsten, 2006). To conduct the semantic-network analysis, several successive steps based on literature by Vlieger and Leydesdorff (2010, 2012) were followed.
First, lists of the 255 most frequently used words in the discourses of the actors were created for each crisis stage, using the software FrequencyList. Stopwords were removed by applying a German stopword list that has been used in a previous study on implicit frames in the German language (Strauß & Vliegenthart, 2017).3 Additionally, the words were stemmed and special characters were removed to aggregate words that are semantically similar but syntactically different (e.g. plurals or mutated vowels).
Second, word-document matrices were built based on the produced word list, using the software Ti for social media posts and FullText for news articles. In word-document matrices, the identified words constitute the variables, whereas the cases represent their occurrences across the different news articles or tweets.
Third, these matrices served as the input for principle component factor analyses (PCA) with Varimax rotation in SPSS. Factor analysis can infer latent concepts from
observable variables (Field, 2009). Thus, PCA is a suitable method to identify implicit frames from (co-)occurring words in the discourses of the actors.
For conceptual and methodological reasons, the number of extracted components was set to six. First, a fixed number of comparable components was needed to investigate frame alignment quantitatively (see van der Meer, 2016). This was established by locating a comparable explained variance on the last retrieved components between the actors (see Appendix 2; cf. Kaiser, 1958). Second, the public’s components beyond six were sometimes not conceptually categorizable, i.e. they did not meet the interpretability criterion (Revelle & Rocklin, 1979).
Components, i.e. implicit frames, scoring the highest explained variance (R2) can be
considered the dominant frames of the actor within the given time period, whereas the
consecutive frames can be considered sub-frames (see van der Meer, 2016). Cronbach’s alpha of the identified frames ranged between .81 and .92, indicating that the scales show a high level of internal consistency (Nunnally, 1978; Tavakol & Dennick, 2011). It should be noted
that the comparison between news articles and tweets initiates an unequal distribution in variance of words in the texts, leading to a higher component diversity in the social media discourse because tweets are limited to 140 words (see Appendix 2). However, since the last retrieved components share a similar explained variance between the actors, the assumption of conceptual likeness is met.
Finally, to address RQ2 and compare frame alignment between news media and the public over time, a variation of an analytical extension of semantic-network analysis introduced by van der Meer et al. (2014) was used. Since factor loadings constitute the relative importance of words in the identified frames (van der Meer et al., 2014), they can be statistically assessed by correlating the factor loadings of mutually used words between the actors in their framing. Consequently, frame alignment can be statistically observed when mutually used words have a comparable importance for the constitution of frames, i.e. when their factor loadings on frames correlate with each other. In practice, the factor loadings of mutually used words scoring the highest on a frame were extracted for each crisis period. Finally, Spearman’s rho correlations were used to statistically address the similarities between the factor loadings for the frame building of the actors.4 Since a census of news media
coverage and social media posts were obtained, the general patterns of frame (de-)alignment are deemed as the relevant indicator for answering H1-H3, rather than significance testing (see also van der Meer et al., 2014).
Results Framing-building
A semantic-network analysis was conducted to gain insight into the frame building of news media and the public following the 2016 Munich shooting. Appropriately, principal component analyses (PCA) were run to identify the implicit frames in the communications of the actors over the three different crisis stages, thereby answering RQ1. The reported frames
can conceptually be distinguished into eight different frame clusters, which are described in Table 2. A complete overview of all identified components, i.e. frames, and their properties are reported in Appendix 2. The dynamic evolution of frame cluster presence is visualized in Figure 2.1. and 2.2.
Table 2. Conceptual frame clusters of identified implicit frames
Frame cluster Description Components
Information frames Providing basic information on
the incident or the investigations, closely resembling official information
Information frame, Traffic frame, Victim count frame, Search information frame
Interpersonal support frames Framing support for different
actors, such as victims, the general public or investigators
Police support frame, Solidarity frame, Condolences frame
Political response frames Depicting political responses on
the incident, including
incumbent and non-incumbent reactions
Political response frame, Populist frame
Discussion frames Relating the incident to a larger
societal context and discussing its implications
Government criticism frame, Comparison frame, Safety frame
Alarming frames Framing warnings on ongoing
dangers related to the crisis
Caution frame, Situation frame
Responsibility frames Attributing responsibility for the
crisis on an individual and political level
Shooting spree frame, Terrorism frame, Perpetrator frame
Human interest frames Focusing on emotional sides of
the incident, such as human involve-ment or sensational circumstances
Witness frame, Victim frame
Inter-media frames Referring to other media
activities on the crisis, such as the spread of rumours on other media
Figure 2.1. Frame cluster presence of the news media over crisis stages in %
Note. Non-present cluster in the actor’s communications (‘interpersonal support’) has been
excluded from the figure. Numerical properties can be found in Appendix 3.
Figure 2.2. Frame cluster presence of the public over crisis stages in %
Note. Non-present clusters in the actor’s communications (‘human interest’, ‘inter-media’)
have been excluded from the figure. Numerical properties can be found in Appendix 3.
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60%
Acute Crisis Stage Chronic Crisis Stage Resolution Crisis Stage
News Media
Information Political response Discussion Alarming
Responsibility Human interest Inter-media
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60%
Acute Crisis Stage Chronic Crisis Stage Resolution Crisis Stage
Public
Information Interpersonal support Political response
In the acute stage of the crisis, information frames were most present in the news media’s communications (58,6%). News media dominantly framed basic official information on the unfolding crisis (R2 = 15.98), using words like ‘police’, ‘perpetrator’ and
‘spokesperson’. Also, other sub-frames on this cluster framed basic information on traffic (R2 = 7.70) and victim count (R2 = 5.13). Furthermore, alarming frames were present in the news media’s communications (29%), warning about the severity of the situation (R2 = 9.70) and advising caution (R2 = 4.52). Also, a political response sub-frame (12,4%) captured first
political responses on the crisis (R2 = 6.08).
In the public’s communications, interpersonal support frames were most salient (35,5%) during the acute stage, dominantly calling for police support (R2 = 3.23), using words like ‘shooting’, ‘photos’, ‘online’ and ‘support’, as well as offering shelter to affected people (R2 = 2.72). Information frames were comparably present (33%), informing about the police
search (R2 = 3.00) and victim count (R2 = 2.54). Other salient sub-frames focused on a
political response (17%, R2 = 2.86) and public alarming (14,5%, R2 = 2.43).
During the chronic stage of the crisis, news media’s information frames decreased in salience (33,1%), yet the cluster remained dominant. More specifically, the dominant
information frame (R2 = 12.22) provided extended information about the ongoing
investigation’s findings about how the shooter lured his victims, using words such as ‘Facebook’, ‘invitation’ and ‘McDonald’. Political response was the second most salient cluster (26,7%), framing reactions of incumbent politicians (R2 = 5.87), as well as a populist social media response (R2 = 4.03). Furthermore, the human-interest cluster appeared in this stage (21,4%), framing emotional stories of victims (R2 = 4.13) and eyewitnesses (R2 = 3.82).
Also, an inter-media frame appeared in this stage (18,8%), framing the role of social media in the events (R2 = 6.96).
For the public, information frames evolved to the most salient cluster in the chronic
using mutual words like ‘Facebook’, ‘invitation’ and ‘McDonald’, and information on the police hunt (R2 = 2.86). The responsibility cluster firstly appeared in this stage among the
public (33,2%), attributing responsibility to the gunman (R2 = 2.96) and linking it to his intentionality (R2 = 2.75). Furthermore, interpersonal support remained similarly salient for public framing (32,5%), framing solidarity (R2 = 2.98) and support of victims (R2 = 2.61).
In the resolution stage of the crisis, information remained the most salient frame cluster among the news media (36%), dominantly framing information about the mental illness of the shooter as disclosed in the police’s press conference (R2 = 12.13), using words like ‘treatment’, ‘depression’ and ‘investigators’. Furthermore, the presence of human interest remained relatively constant (19,4%), framing emotional stories of victims (R2 = 6.52). Also, two clusters appeared for the first time among news media in this stage: The news media attributed responsibility (14,4%) to the gunman (R2 = 6.52) and commenced
discussions (10,4%) regarding the crisis and national safety issues (R2 = 12.13). Lastly, an
inter-media frame remained present in news media’s communications (7,9%), referring to the role of social media in the crisis (R2 = 2.65).
In the public’s communications during the resolution stage, the discussion cluster appeared for the first time (32,8%), providing the dominant frame discussing government criticism (R2 = 2.53), using words like ‘illegal’, ‘minister’ and ‘weapon’, and with a sub-frame discussing the shooting in the wake of other recent violent crimes (R2 = 2.34). The
interpersonal support cluster remained relatively equally present compared to the previous
stages (32,8%), with sub-frames offering condolences (R2 = 2.40) and solidarity with victims (R2 = 2.13). The information cluster was less present than in the previous stages (16,2%),
framing investigation findings (R2 = 2.23). Also, the responsibility cluster was less present (15,8%), attributing responsibility with terrorist intentions to the gunman (R2 = 2.17).
Frame alignment
In order to investigate frame alignment between the actors over the different crisis stages, Spearman’s rho correlations between the factor loadings of mutually used words in implicit frames of the news media and the public were conducted for each crisis period, answering H1-H3.
Figure 3. Patterns of frame (de-)alignment
For the acute stage of the crisis, a marginally negative correlation between the factor loadings was found, rs = -0.10. The correlation close to zero indicates a fundamentally
deviating word importance of mutually used words in the identified frames between the actors (van der Meer et al., 2014). For example, while the news media heavily used the words
‘police’ (.86), ‘casualties’ (.78), ‘public’ (.76) and ‘unclear’ (.69) as part of an information frame, these words had relatively little importance for the implicit frames on Twitter (.25, -.12, .49, .24). Conversely, Twitter users for example widely used ‘help’ (.75), ‘police forces’ (.75), ‘operation’ (.68) and ‘video’ (.60) as part of the dominant police support frame,
whereas these words had less importance for the implicit frames in the news media (<.10, .51, .44, .47). In sum, the overall disaccording importance of words for the implicit framing of
-0,10 0,47 -0,22 -0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6
Acute crisis stage Chronic crisis stage Resolution crisis stage
Fa ct o r lo ad in g co rr el at io ns ( Sp ea rma n' s rh o )
news media and the public support actor-specific framing patterns of the actors in the acute
stage of the crisis. Therefore, H1 is supported.
As for the chronic stage of the crisis, there was a moderately strong positive
correlation between the factor loadings of the actors, rs = 0.47. This correlation demonstrates
the fair equivalence of word importance for the framing among the actors. Illustratively, both news media and the public used the words ‘invitation’ (NM: .78; SM: .83), ‘McDonald’ (NM: .68; SM: .84) and ‘Facebook’ (NM: .52; SM: .75) to a comparable extent as part of
information frames on how the perpetrator lured his victims to the shooting site. Comparing the correlations of the first two stages (T1: rs = -0.10; T2: rs = 0.47), a clear pattern of
increasing mutual word emphasis in the framing process among the actors becomes visible, indicating frame alignment in the second stage of the crisis. Thus, H2 is supported.
Eventually, a weak negative correlation between the factor loadings in the resolution stage of the crisis was found, rs = -0.22. This correlation denotes a diverging importance of
mutually used words for the implicit framing of the actors (van der Meer et al., 2014).
Exemplarily, the words ‘illegal’ (.95), ‘weapon’ (.80) and ‘interior minister’ (.69) prominently formed part of the government criticism frame on social media, whereas these words had a considerably deviating importance in the implicit framing of news media (.37, .28, .12). Comparing the correlation to the previous result of H2 (T2: 0.47; T3: -0.22), a clear pattern of
de-alignment in the implicit framing between the actors becomes evident. Therefore, H3 is confirmed.
Conclusion and discussion
This study aimed at investigating the frame building and frame alignment processes of news media and the public during a violent crime crisis by addressing the research question:
Which frames are present in the crisis communications of news media and the public, how do they evolve over time and do they align? The research question can be answered by the
following findings: First, this study automatically identified implicit frames in the actors’ communications over the course of the crisis (RQ1), which were conceptually described in eight different frame functions (or clusters). Results established frame clusters of (1)
information, (2) interpersonal support, (3) political response, (4) discussion, (5) alarming, (6) responsibility, (7) human interest and (8) inter-media. It was described how these frame functions dynamically evolved over time. Second, an analytical extension of semantic-network analysis was conducted to test patterns of frame alignment and frame de-alignment amongst the actors over the course of the crisis (H1-H3). Results revealed that the framing of news media and the public was not aligned in the acute crisis stage, indicating actor-specific meaning construction (H1). In the chronic crisis stage, the results showed signs of alignment among the actors’ crisis framing, indicating collective sensemaking processes and
rapprochement of interpretations (H2). Finally, results showed de-alignment of the actors’ implicit framing in the resolution stage, indicating the restoration of actor-specific framing patterns after the crisis resolved (H3).
Little was known about the dynamic frame building processes of the news media and the public in times of crisis so far. Strikingly, the most prominent and stable frame clusters of the news media and the public identified in this study corresponded to existing findings: First, the public prominently framed interpersonal support throughout the crisis, which supports extant theory that considers emotional support as a main function of citizens to seek and disseminate information on social media during crisis situations (Liu & Kim, 2011; Liu et al., 2011). Second, the news media dominantly applied information frames over the whole course of the crisis, giving support to propositions of initial studies that news media frames closely resemble information provided by official sources during crisis situations (Olsson & Nord, 2014; Falkheimer & Olsson, 2015; Zheng et al., 2015). However, the public also prominently used information frames in the first two crisis stages, implying that framing official
response to an unsettled crisis. Relating these patterns to the larger context of crisis situations, this however assumes that there is official information available and emotional involvement of the public apparent, which makes future research worthwhile about how the actors prominently frame crisis situations that do not meet these requirements.
This study has inductively identified further frame functions, thereby making some theoretical assertions on crisis framing for further research. First, results inductively revealed frame clusters that differed in their use by actors. Exemplarily, human interest frames were exclusively found for news media and interpersonal support frames for the public. This gives support that the actors’ framing is bound to their personal and professional identities
(Scheufele, 1999; Benford & Snow, 2000). In other words, while human interest frames for example closely resemble a journalistic narrative deeply rooted in practice and storytelling routines (Semetko & Valkenburg, 2000; Cho & Gower, 2009), interpersonal support can be seen as a function of the public’s personal relevance to the crisis (van der Meer & Verhoeven, 2013). This implies that the crisis framing of the news media and the public need to be
understood in the context of their social function and respective identities. Second, results revealed time-specific frame clusters that occurred time-dependently during the crisis. Illustratively, alarming frames and discussion frames were present only in the first stage and last stage respectively. This suggests that certain frames evolve precisely in the context of the specific communicative demands that the crisis poses to the actors at that time (e.g. Reynolds & Seeger, 2005). In this sense, the actors used the alarming frames only when it was
immediately needed to call attention to the evolving crisis (Coombs, 2010). Discussion
frames, discussing the implications of the crisis for the future, solely occurred when the crisis
was resolved and actors needed to facilitate risk prevention (Reynolds & Seeger, 2005). Hence, this implies that crisis framing also needs to be understood in the context of crisis development, meaning that framing is a dynamic communicative process that is closely associated with the crisis circumstances and progress.
Little research had been conducted on patterns of frame alignment between the news media and the public in times of crisis so far (H1-H3). Generally, this study provides support to initial research in this field which finds patterns of crisis frame alignment between news media and the public both qualitatively (van der Meer & Verhoeven, 2013) and quantitatively (van der Meer et al., 2014). The statistical findings overall support that the actors aligned their meaning provisions in the chronic stage by utilizing more similar interpretations of the
situation. However, as this association only emphasizes the similarities of frames by mutual word use of the actors, it cannot claim that the frames also conflict when they are not aligned (see also van der Meer et al., 2014). Therefore, the implications of the actors’ frame building patterns enrich the general conclusion. Both actors dominantly framed official information in the chronic stage, implying the similar interpretations were also most salient in contrast to the other stages. Furthermore, the actors largely corresponded in mutual word use in these
specific frames (see results), implying they also complied in sensemaking within these frames. The focus on information frames for alignment largely corresponds to existing findings in a different case study (van der Meer & Verhoeven, 2013), which may be explained by the reliance on a single actor to announce the end of the crisis (e.g. investigators). In sum, the findings therefore imply that the news media and the public collectively made sense of the situation by making their meaning provisions of official information on ongoing crisis closure more salient and more similar in the chronic stage of the crisis. Due to the responsibility-settling and definitive character of these interpretations, the alignment therefore suggests a communicative closure over the crisis between the actors.
Remarkably, this study has found the highest degree of crisis frame (de-)alignment between actors in an empirical study so far (see van der Meer et al., 2014). This finding can be particularly discussed under aspects of crisis type and crisis involvement. Existing studies find small indications that actors tend to engage increasingly in collective sensemaking when they are directly affected (Schultz & Raupp, 2010; van der Meer et al., 2014). In this sense,
achieving issue coherence is foremost a concern of actors who need to repel the most adverse consequences of the crisis (Hellsten et al., 2010). It can therefore be argued that frame
alignment increasingly occurs where actors are heavily involved in the crisis. Since previous findings on frame alignment between news media and the public were often based on
organizational crises (van der Meer et al., 2014), this study might have provided a case in which the public and media were increasingly involved. However, further research is required to establish the link between frame alignment and the involvement of the actors in the crisis.
To inspire future research into the actors’ dynamic crisis framing process, some shortcomings of this study should be addressed. First, the findings are based on a case study, which limits generalizability. The specific context of a violent crime crisis might have resulted in crisis-specific frames that are not applicable to other crises. Therefore, future research should pick up on the established frames and frame clusters and test them in different contexts. Second, while the data census provided a remarkable amount of data, shortcomings in the media selection should be noted. Since Twitter is also an important space for news media activity (Sung & Hwang, 2014), its representativeness as a ‘pure’ public media remains limited. Future studies should control for the news media effects by also taking other public manifestations (e.g. Facebook, interpersonal discussions) into account. Lastly, while this study was able to describe patterns of crisis frame alignment, the driving mechanisms for this phenomenon, i.e. its causes and effects, remain a question of theoretical interpretation using this research design. Future research should empirically delve further into these mechanisms to explain their forces, e.g. by establishing experimental designs testing against non-crisis events or conducting time-series analyses based on Granger causality.
Despite these shortcomings, this study makes valuable contributions to the existing body of literature on crisis communication, which may stimulate future research. As one of its first, this study grasps an initial understanding of the dynamic frame building process of the news media and the public in times of crisis, answering the call for a more dynamic
investigation on the actors’ sensemaking in times of crisis (Schultz et al., 2012; Utz et al., 2013). Since inductive research is valuable where little theoretical knowledge is available, the established frames and their development patterns mark a useful theoretical starting point to test the actors’ framing in future empirical research on crisis situations. Second, the findings contribute to the aspiring field of frame alignment research that aims to understand collective sensemaking in times of crisis. By establishing frame alignment patterns in a public crisis case and suggesting influences on the magnitude of the phenomenon, this study gives directions for further research. Third, this study contributes on a methodological level by validating an innovative way to automatically identify and compare frames in large amounts of texts using automated content analysis. Increasingly pursuing such approaches in the future may help communication research to create more empirical consistency in research on the sometimes-ambiguous framing concept. Finally, the study has a societal implication. Actors in crisis communication should understand their roles in relation to the dynamic development of the crisis and in the context of other actors, who simultaneously try to make sense of the situation, in order to gain a better understanding of how crises can be collectively resolved.
Note
1 Fink’s approach originally assumes four different crisis stages. However, as the model
conceptualizes crises as events with sufficient warning signs, the prodromal stage of pre-crisis symptoms is not applicable to unexpectedly occurring violent crime crises. In the context of a violent crime, a pre-crisis stage may be reduced to a potential threat as a ‘latent issue’ in society, in which little measurable communication between the actors occurs.
2 German-language search terms equalling “Munich AND (fatalities OR casualties OR attack
OR OEZ)” were used to exclude shooting unrelated coverage, such as city or sports news.
3 Stopwords are commonly used words that have little semantic value for the constitution of
frames, e.g. short function words such as ‘is’, ‘was’, ‘will’.
4 To correlate the factor loadings, Spearman’s rho correlations is used rather than Pearson’s r
References
An, S. K., & Gower, K. K. (2009). How do the news media frame crises? A content analysis of crisis news coverage. Public Relations Review, 35(2), 107-112.
https://doi.org/10.1016/j.pubrev.2009.01.010
Backes, T., Jaschensky W., Langhans, K., Munzinger, H., Witzenberger, B. & Wormer, V. (2016). Timeline der Panik. Süddeutsche Zeitung. Retrieved from:
http://gfx.sueddeutsche.de/apps/57eba578910a46f716ca829d/www/
Benford, R. D., & Snow, D. A. (2000). Framing processes and social movements: An overview and assessment. Annual review of sociology, 26(1), 611-639.
https://doi.org/10.1146/annurev.soc.26.1.611
Benoit, W. L. (1997). Image repair discourse and crisis communication. Public relations
review, 23(2), 177-186. https://doi.org/10.1016/S0363-8111(97)90023-0
Boin, A., & Hart, P. T. (2003). Public leadership in times of crisis: mission impossible?. Public Administration Review, 63(5), 544-553.
Boin, A., Hart, P. T., & McConnell, A. (2009). Crisis exploitation: political and policy impacts of framing contests. Journal of European Public Policy, 16(1), 81-106. https://doi.org/10.1080/13501760802453221
Boin, A., Stern, E., & Sundelius, B. (2016). The politics of crisis management: Public
leadership under pressure. Cambridge University Press.
Borah, P. (2011). Conceptual issues in framing theory: A systematic examination of a decade's literature. Journal of communication, 61(2), 246-263.
https://doi.org/10.1111/j.1460-2466.2011.01539.x
Bundy, J., Pfarrer, M. D., Short, C. E., & Coombs, W. T. (2016). Crises and Crisis Management Integration, Interpretation, and Research Development. Journal of
Chyi, H. I., & McCombs, M. (2004). Media salience and the process of framing: Coverage of the Columbine school shootings. Journalism & Mass Communication Quarterly, 81(1), 22-35. https://doi.org/10.1177/107769900408100103
Claeys, A. S., & Cauberghe, V. (2014). What makes crisis response strategies work? The impact of crisis involvement and message framing. Journal of Business Research, 67(2), 182-189. https://doi.org/10.1016/j.jbusres.2012.10.005
Cloudman, R., & Hallahan, K. (2006). Crisis communications preparedness among US organizations: Activities and assessments by public relations practitioners. Public
Relations Review, 32(4), 367-376. https://doi.org/10.1016/j.pubrev.2006.09.005
Coombs, W. T. (2010). Parameters for crisis communication. In W. T. Coombs, & S. J. Holladay (Eds.). The handbook of crisis communication. (pp. 17-53). Chichester: Wiley-Blackwell.
Coombs, W. T., & Holladay, S. J. (2008). Comparing apology to equivalent crisis response strategies: Clarifying apology's role and value in crisis communication. Public Relations
Review, 34(3), 252-257. https://doi.org/10.1016/j.pubrev.2008.04.001
Dey, R., Jelveh, Z., & Ross, K. (2012, March). Facebook users have become much more private: A large-scale study. In Pervasive Computing and Communications Workshops
(PERCOM Workshops), 2012 IEEE International Conference on (pp. 346-352). IEEE.
https://doi.org/10.1109/PerComW.2012.6197508
Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of
communication, 43(4), 51-58. https://doi.org/10.1111/j.1460-2466.1993.tb01304.x
Falkenheimer, J., & Heide, M. (2010). Crisis communicators in change: From plans to improvisations. In W. T. Coombs, & S. J. Holladay (Eds.). The handbook of crisis
Falkheimer, J., & Olsson, E. K. (2015). Depoliticizing terror: The news framing of the terrorist attacks in Norway, 22 July 2011. Media, War & Conflict, 8(1), 70-85. https://doi.org/10.1177/1750635214531109
Fink, S. (1986). Crisis management: Planning for the inevitable. American Management Association.
Galtung, J., & Ruge, M. H. (1965). The structure of foreign news: The presentation of the Congo, Cuba and Cyprus crises in four Norwegian newspapers. Journal of peace
research, 2(1), 64-90.
Glik, D. C. (2007). Risk communication for public health emergencies. Annu. Rev. Public
Health, 28, 33-54. https://doi.org/10.1146/annurev.publhealth.28.021406.144123
Graber, D. A., & Dunaway, J. (2014). Mass media and American politics. Cq Press. Field, A. (2009). Discovering statistics using SPSS. Sage publications.
Harcup, T., & O'Neill, D. (2001). What is news? Galtung and Ruge revisited. Journalism
studies, 2(2), 261-280. https://doi.org/10.1080/14616700118449
Heath, R. L. (2010). Crisis communication: Defining the beast and de-marginalizing key publics. In W. T. Coombs, & S. J. Holladay (Eds.). The handbook of crisis communication. (pp. 1-13). Chichester: Wiley-Blackwell.
Hellsten, I., Dawson, J., & Leydesdorff, L. (2010). Implicit media frames: Automated
analysis of public debate on artificial sweeteners. Public Understanding of Science, 19(5), 590-608. https://doi.org/10.1177/0963662509343136
Heverin, T., & Zach, L. (2010). Microblogging for Crisis Communication: Examination of Twitter Use in Response to a 2009 Violent Crisis in the Seattle-Tacoma, Washington Area. In Proceedings of the 7th International ISCRAM Conference–Seattle (Vol. 1).
Holland, K., Sweet, M., Blood, R. W., & Fogarty, A. (2014). A legacy of the swine flu global pandemic: Journalists, expert sources, and conflicts of interest. Journalism, 15(1), 53-71. https://doi.org/10.1177/1464884913480460
Iannarino, N. T., Veil, S. R., & Cotton, A. J. (2015). Bringing home the crisis: How US evening news framed the 2011 Japan Nuclear Crisis. Journal of Contingencies and Crisis
Management, 23(3), 169-181. https://doi.org/10.1111/1468-5973.12068
Jonkman, J., & Verhoeven, P. (2013). From risk to safety: Implicit frames of third-party airport risk in Dutch quality newspapers between 1992 and 2009. Safety science, 58, 1-10. https://doi.org/10.1016/j.ssci.2013.03.012
Kaiser, H. F. (1958). The varimax criterion for analytic rotation in factor
analysis. Psychometrika, 23(3), 187-200. https://doi.org/10.1007/BF02289233
Kleinnijenhuis, J., Schultz, F., Utz, S., & Oegema, D. (2015). The mediating role of the news in the BP oil spill crisis 2010: How US news is influenced by public relations and in turn influences public awareness, foreign news, and the share price. Communication
Research, 42(3), 408-428. https://doi.org/10.1177/0093650213510940
Lerbinger, O. (2011). The crisis manager: facing disasters, conflicts, and failures. Routledge. Liu, B. F., Austin, L., & Jin, Y. (2011). How publics respond to crisis communication
strategies: The interplay of information form and source. Public Relations Review, 37(4), 345-353. https://doi.org/10.1016/j.pubrev.2011.08.004
Liu, B. F. (2010). Distinguishing how elite newspapers and A-list blogs cover crises: Insights for managing crises online. Public Relations Review, 36(1), 28-34.
https://doi.org/10.1016/j.pubrev.2009.10.006
Liu, B. F., Fraustino, J. D., & Jin, Y. (2016). Social Media Use During Disasters How Information Form and Source Influence Intended Behavioral Responses. Communication
Research, 43(5), 626-646.
Liu, B. F., & Kim, S. (2011). How organizations framed the 2009 H1N1 pandemic via social and traditional media: Implications for US health communicators. Public Relations
Marra, F. J. (1998). Crisis communication plans: Poor predictors of excellent crisis public relations. Public Relations Review, 24(4), 461-474. https://doi.org/10.1016/S0363-8111(99)80111-8
Masters, A. & Hart, P. (2012), ‘Prime Ministerial and Recession Politics: Meaning Making in Economic Crisis Management’, Public Administration. 90(3). 759–780.
McCombs, M. (2005). A look at agenda-setting: Past, present and future. Journalism
studies, 6(4), 543-557. https://doi.org/10.1080/14616700500250438
McGinty, E. E., Webster, D. W., Jarlenski, M., & Barry, C. L. (2014). News media framing of serious mental illness and gun violence in the United States, 1997-2012. American journal
of public health, 104(3), 406-413. https://doi.org/10.2105/AJPH.2013.301557
McQuail, D. (2010). McQuail’s Mass Communication Theory, 6th ed. London: Sage. Neuman, W. R., Guggenheim, L., Mo Jang, S., & Bae, S. Y. (2014). The dynamics of public
attention: Agenda-setting theory meets big data. Journal of Communication, 64(2), 193-214. https://doi.org/10.1111/jcom.12088
Neuman, W. R., Just, M. R., & Crigler, A. N. (1992). Common knowledge. News and the
construction of political meaning. Chicago: University of Chicago Press.
Nikolaev, A. G. (2010). Thirty common basic elements of crisis management plans: Guidelines for handling the acute stage of “hard” emergencies at the tactical level. The
handbook of crisis communication, 261-281.
Olsson, E. K., & Nord, L. W. (2015). Paving the way for crisis exploitation: The role of journalistic styles and standards. Journalism, 16(3), 341-358.
https://doi.org/10.1177/1464884913519032
Olsson, E. K., Nord, L. W., & Falkheimer, J. (2015). Media coverage crisis exploitation characteristics: A case comparison study. Journal of Public Relations Research, 27(2), 158-174. https://doi.org/10.1080/1062726X.2014.976827
