In times of crisis : frame alignment across company, government and online newspapers

In Times of Crisis

Frame Alignment across Company, Government and Online

Newspapers

Master’s Thesis

Graduate School of Communication, University of Amsterdam

Master’s program Communication Science

Submitted by:

Jeannine Egi

11104317

Supervisor:

Toni van der Meer

Date of Completion:

Abstract

This study analyzes the communicative interplay during a corporate crisis among the company in crisis, the corresponding government and online newspapers. This paper

contributes to the field of PR and crisis communication research by building on sense-making and framing theory and thereby stretching the notion of frame (de-)alignment across the private and the public sector as well as the news media.

Through automated semantic-network analyses, this research studies differences and similarities of implicit frames that can be found in the press releases from Malaysia Airlines and the Malaysian government and in news articles from online newspapers during the

MH370 crisis. The findings document that the government engaged in the discussion at a later point in time and imply that the government experienced an aggravated entrance into the already established communicative discourse between Malaysia Airlines and the news media. Furthermore, a belated frame alignment among all three actors was found, indicating that high uncertainty during crises can delay the process of frame alignment.

Introduction

Corporate crises are disrupting the status quo of today’s society on a regular basis and oftentimes bear severe consequences. Communication and simultaneously framing by

relevant actors such as the company, the news media or the corresponding government during these times can have a serious impact on the development of the crisis (Cornelissen, Carroll & Elving, 2010; Liu, Horsley & Levenshus, 2010; Schultz, Kleinnijenhuis, Oegema, Utz, & Van Atteveldt, 2012). Communication during crises draws on the concept of framing when

involved actors “select some aspects of a perceived reality and make them more salient in a communicating text” (Entman, 1993, p. 52). Actors thereby define problems and cause of the crisis, make moral judgments or express treatment recommendations according to their understanding of the issue, which opens up the possibility for contradicting frames. The relevance of differing frames lies in the power that established media outlets can have on stakeholders’ perception of organizations (Hallahan, 1999). Those frames are able to direct the public’s attention to certain aspects and withhold others (McCombs, Shaw & Weaver, 1997). Therefore, one goal of organizations in crisis situations is to influence the way in which mass media frame the crisis and ultimately, to trigger positive evaluations of the organization by stakeholders and the public (Coombs, 2007; Coombs & Holladay, 2011).

So far, discussions in crisis communication and previous case studies on crisis

responsibility (Coombs, 2004) or response strategies (Benoit, 1995; Millar & Heath, 2004) have been focusing on the private sector and on the interplay between company in crisis and the news media (e.g. Cornelissen et al., 2010; Bowen & Zheng, 2015). In addition, the existing research on governmental crisis communication evolves around societal crises (e.g. Lee, 2009; Schultz & Raupp, 2009; Graham, Avery & Park, 2015), rather than corporate crises that include the corresponding government. However, the exploration of this

the need of governmental support during a corporate crisis indicates that the crisis has reached a certain severity and the company is facing major issues that also concern society. On the other hand, public relations practitioners deal with unique challenges depending on the sector in which they work (Liu & Horsley, 2007; Lee, 2009). Therefore, it is important to understand the communication dynamics among the company in crisis, the corresponding government and the news media.

Furthermore, the aspect of time is an essential factor in crisis situations. With time, information about the crisis can surface and influence communication. Thus, it must be taken into account that the communicated content during crises can change together with the various stages of a crisis (Jaques, 2007).

During the last 20 years, the media landscape has developed to a decentralized and interactive system where the phenomenon of mass self-communication has become standard (Castells, 2007; Karlsson, 2012). However, the content of these alternative forms of media, such as blogs, mostly rely on traditional media, which indicates that people continue to prefer information produced on professional journalistic standards (Reese, Rutigliano, Hyun & Jeong, 2007; Karlsson, 2012). Yet, because of the digitalization, the field of traditional media has changed as well. Online news has emerged and is characterized by immediacy.

Immediacy describes the process of immediate publication of information that the online news outlet receives, which can lead to quick emergence and continuous changing of frames (Massey & Levy, 1999; Karlsson, 2011; 2012).

So far, public relation research has been studying the general influence of online news immediacy on organizations in crises and the challenges for organizations to meet the expectations of instant information provision that come with a 24/7 news cycle (Bucher, 2002; Perry, Taylor & Doerfel, 2003). Only Karlsson (2012) has examined whether and how the immediacy of online news affects the framing of crises. Therefore, this study aims to contribute to the existing research on crisis communication of online newspapers and to fill

the abovementioned gap regarding the involvement of governments in corporate crises by posing the following research question:

To what extent do the frames used by the company, the corresponding government and the online newspapers differ during a corporate crisis and how does this relate to the different

phases of a crisis?

This paper attempts to answer this research question by applying computer assisted semantic-network analyses (Hellsten, Dawson & Leydesdorff, 2010) onto organizational and governmental press releases as well as news articles from online newspapers.

Theoretical framework Communication during corporate crises

A corporate crisis can be understood 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, 2007, pp. 2-3). Such events are characterized by low probability and severe consequences for the company as their

fundamental objectives and their reputation might be at risk (Weick, 1988; Coombs, 2007). During these times, the aspect of communication is especially important for several reasons. First, crises are newsworthy as they threaten a company’s financial performance as well as its reputation, which is why most crises receive high media attention and scrutiny (Seeger, Sellnow, & Ulmer, 2003; Huang, 2006). Furthermore, one inherent characteristic of most crises is uncertainty, as the company often lacks information for explanations in the beginning of a crisis (Lu, 2013). In order to avoid speculation and incorrect media coverage that can hurt the company’s image severely (Horsley & Barker, 2002), it is important for the

company in crisis to communicate swiftly and try to guide the communication about the crisis. In addition, a corporate crisis can also have great impact on society and individuals, for instance when people are facing physical or psychological threats (Coombs, 2007). Events that are perceived as a risk require dissemination of facts to affected publics (Chakraborty & Creutzfeldt-Banda, 2011) as people seek information about the cause and want to make attributions (Coombs & Holladay, 2004). Their perceptions of events are based on the information that is salient in their environment (Heider, 1958), which again relates to the information provided by the company and displayed by the media. When considering all these factors, it becomes clear that even though crises origin from real problems, they are

constituted in the communicative interplay among several actors. It is the actors’ perception of a series of events as a crisis that leads to consequences for corporate, governmental and public actors as well as the news media (Kleinnijenhuis, Schultz, Oegema, 2015). Scholars have repeatedly shown that communication can have an impact on duration, magnitude and consequences of corporate crises (Cornelissen et al., 2010; Schultz et al., 2012; Holladay & Coombs, 2013), which is why it is essential to understand the role of communication among different actor constellations during such crises.

Framing in organizational crisis situations

Framing

Frames used in communication materials (e.g. press releases or news articles) highlight some information in order to make them more noticeable and meaningful to the audience and thereby aim to influence the receivers’ opinion about the issue (Entman, 1993). By using certain key words, information sources, and sentences it is possible to thematically reinforce constructs of facts or judgments (Entman, 1993). The presence of a certain kind of

information about an issue for a longer period of time can then result in an unconscious foundation to which individuals resort to when they form opinions (Higgins, 1996). Thus, it

can be said that framing refers to the construction of social reality as it provides the perspectives through which people see a certain subject (Hallahan, 1999).

As mentioned, frames are build trough the specific use of language (Entman, 1993), which is based on the choice of particular words (Leydesdorff & Hellsten, 2006). The

network in which these words are embedded gives the words their specific meaning (Hellsten et al., 2010). This understanding is based on the analytical approach that communication materials can contain both implicit as well as explicit frames. An explicit frame can be found if an author decides to focus on a specific perspective by using certain words to characterize the topic (Hellsten et al., 2010). These explicit frames can be identified with the help of traditional content analysis, where texts are coded to predefined categories (Krippendorf, 1980). However, in communication materials explicit frames are less represented; the

majority of frames are implicit (Hellsten, 2002). In implicit framing, the meaning of words is created by the semantic context in which they are used. The implicit frames can then be recognized “in latent dimensions of the communication, and they are generated because of spurious correlations between word (co-)occurrences in communications” (Hellsten et al., 2010, p. 5). Therefore, unlike explicit frames, implicit frames are not directly observable and need to be discovered by using a computed-assisted method of automated content analysis. As only the analysis of implicit frames allows to study the development of frames over time and to compare the frame usage of different actors (Jonkman & Verhoeven, 2013; Leydesdorff & Hellsten, 2005), the focus of this paper lies on the implicit frames that are used by the

abovementioned three actors.

Sense-making process and competing frames

When communication during an event or a crisis involve different actors it is possible that they establish competing interpretations and meanings (Chong & Druckman, 2007). This is attributed to the fact that these actors all make sense of the event themselves first, before

they start communicating about it (Weick, 1995). Sense-making can be described as the interpretation, categorization, labeling, and explanation of a certain issue by individuals within an organization (Weick, Sutcliffe & Obstfeld, 2005; Cornelissen et al., 2010). Circumstances are thereby “turned into a situation that is comprehended explicitly in words and that serves as a springboard to action” (Taylor & Van Every, 2000, p. 40). This sense-making process is influenced by the actor’s professional identity, which involves different values and motives that can affect the way issues are perceived and understood. This means that each actor refers to different beliefs and standards when they assess the significance of an event or issue to the public (Cornelissen et al., 2010).

Therefore, the aspects and details that they deem relevant might differ from each other, which ultimately results in different understandings and competing frames in their

communication. The problem of competing frames arises from the roles of the different actors when it comes to opinion formation of the public. In general, and especially during crises, the frames that are communicated by the media have a severe impact on the public’s perception of the organization in crisis (Hallahan, 1999). Hence, it is crucial for corporate and

governmental organizations to understand how the media interacts with and takes over the frames of other actors.

Interplay between company, government and news media

A crisis usually involves multiple actors, whose communication and framing can heavily influence the development and consequences of the crisis (Schultz et al., 2012). This study focuses on the interplay of three specific actors that are constitutional for reestablishing the social order after a crisis (Kleinnijenhuis et al., 2015): the company going through the crisis, the corresponding government and the news media.

In some corporate crises the government plays an important role, for example when corrective actions and resources coming from the company only are not sufficient to

overcome the problem and cooperation between the company and the government is needed. The involvement of the government predicts severe societal impacts resulting from the corporate crisis, making evident that the company is facing a substantial issue. Interactions between these two actors are usually accompanied by the news media. A corporate crisis is considered newsworthy as it threatens a company’s financial performance as well as its reputation (Seeger et al., 2003; Huang, 2006). Moreover, the involvement of the government increases media scrutiny and coverage (Kim & Liu, 2012), which results in even more intense media attention throughout the crisis.

Even though in literature the suggested crisis communication for companies and government show little differences – both are supposed to communicate instantly,

transparently, reactively, and adjusted to the crisis type accordingly (Benoit, 1997; Horsley & Barker, 2002; Coombs, 2007; Lee, 2009) – public relations practitioners still deal with unique challenges depending on the sector in which they work (Liu & Horsley, 2007; Lee, 2009). For instance, governments are exposed to a higher impact of legal regulations and politics, which lead to several restrictions and responsibilities for governmental crisis managers that

corporate crisis managers do not face (Kim & Liu, 2012). These aspects coincide with the aforementioned notion of professional identity, which drives the sense-making process of these actors. While both company and government want to ensure that accurate information is distributed, there might still be differences in the way they portray certain aspects of the issue because they are guided by diverging environments and societal functions (Liu & Horsley, 2007; Cornelissen et al., 2010). In line with their profit oriented interests, central

characteristics in the framing by companies could be described as illustrating the issue in a way that is likely to be accepted by stakeholders and representing the company in the best possible light (Cornelissen et al., 2010). Governments however, are responsible for the societal well-being (Liu et al., 2010) and might not focus on the company’s interest when communicating about a crisis.

Regarding the sense-making process of the news media, it can be said that their professional identity is two fold. On the one hand, news media fulfill the role of the “watchdog”; they serve as main and critical source of information for society (Johnson, 2014). On the other hand, their professional identity is characterized by economic based interests that they try to pursue, such as gaining attention of the audience and increasing financial profits (Castells 2007; Harcup & O'Neill, 2016). These organizational and economic factors influence the media’s choice of topics and their framing behavior (Galtung & Ruge, 1965; Cornelissen et al., 2010). Therefore, it is likely that in the beginning of a crisis, the communication material from these three actors show conflicting frames, as they follow different interpretation patterns.

However, in order to overcome a crisis, it is necessary to achieve collective sense-making and coherent understanding of the complex situation (Weick, 1988; Seeger, 2002; Hellsten et al., 2010). Actors try to reduce ambiguity and uncertainty, which is why they might start to mutually influence, borrow and align their sense-making and frame usage over the course of a crisis (Leydesdorff & Ivanova, 2014; van der Meer, Verhoeven, Beentjes & Vliegenhart, 2014). As company, government and news media start to engage in a more collective process where they react to each other and try to cooperatively make sense of the crisis, the outcome can be a coherent understanding of the elements that constitute a frame, which is considered constitutional for a common crisis understanding and shared implicit frames (Steinberg, 1998). As a consequence, actors apply more consistent words occurrence patterns in their implicit framing of the crisis. Even though this does not result in a complete overlap of implicit frames, it indicates an alignment in word importance when it comes to meaning construction.

Framing literature suggests that frame construction is often an emergent and dynamic process, constructed through interactions between actors (Steinberg, 1998). When the actors have reached a common understanding of the crisis (i.e. frame alignment), they can return to

following their interests and professional goals. This means that their frames would diverge and move away from one another, focus on new communicative aspects that relate to their professional identity (van der Meer et al., 2014). This process would ultimately lead to frame diversity again.

In order to further address the dynamic character of framing and the influence that time can have on communication in crisis situations, this study analyzes frame development over the course of different crisis stages.

Framing and crisis stages

The frames that are used by involved actors during crises can change together with the different stages of a crisis as the amount of available information changes (Jaques, 2007; Lu, 2013). In order to study the development of frames over time, this paper categorizes crises into a four-stage framework according to Fink (1996). The first stage covers the period before the crisis and deals with crisis prevention. It is followed by the acute crisis stage, where the crisis is most noticeable to outsiders of the organization due to high media coverage. The third stage is called chronic crisis stage, during which the organization tries to deal with and

overcome the crisis damages. The last stage is characterized by organizational activities directed to image restoration and returning to its pre-crisis status. It is called crisis resolution stage. As the purpose of this study is to analyze the interplay of the communication between organization, government and media during an existing crisis, only stages two to four are relevant and the first stage is going to be disregarded.

Acute crisis stage

Especially early phases of a crisis are important for crisis managers, as this phase is characterized by confusion and high media interest (Chakraborty & Creutzfeldt-Banda, 2011). This is why the impact on the public perception of the frames used by the media but also the

possibility of the organization to influence these frames are very high during this stage (Regester & Larkin, 2008).

However, one inherent characteristic of crises causing initial crises responses to be difficult is uncertainty. Uncertainty can occur due to unknown causes and nature of the crisis, unknown involvement of stakeholders and unknown scope and duration of crisis impacts (Lu, 2013). This lack of information can lead to rumors and speculations, which have crucial consequences on the company’s image (Horsley & Barker, 2002). Together with the above mentioned sense-making process, it is conceivable that in the beginning of a crisis, corporate as well as governmental organizations and the media use different frames in their

communication as they start to make sense of the event by processing all kinds of information and cues. Moreover, uncertainty regarding the situation allows incorrect material to surface. Empirical findings have shown that during the initial crisis phase of the BP oil spill the frame usage indeed differed between companies and the media (Schultz et al., 2012). Furthermore, the analysis of several crises of Dutch companies that occurred between 2000 and 2011 disclosed competing implicit frames from companies, news media and the public (van der Meer et al., 2014). In addition, research has also found differing implicit frames used by private and public organizations during the initial crisis communication about the financial crisis in 2008 (Schultz & Raupp, 2009). Therefore, the first hypothesis reads as follows:

H1: The implicit frames used by the company, government and online newspapers differ from

each other during the acute stage of a crisis. Chronic stage

During this intermediate stage of the crisis, temporary procedures are set up until normal organizational operations can resume (Myers, 1993). This also applies to the communication of the involved actors, as they are trying to achieve a consistent understanding of the crisis that is necessary to overcome the crisis (Weick, 1988; Seeger, 2002; Hellsten et al., 2010).

Due to the above outlined interactions and mutual influencing among all actors, it is expected that the usage of implicit frames of these two actors becomes more similar over the course of a crisis, as more information is available and uncertainty decreases (Leydesdorff & Ivanova, 2014; van der Meer et al., 2014). Previous research showed that during the economic crisis, companies and government developed a common narrative over time, even though their frames initially differed from each other (Schultz & Raupp, 2009). Furthermore, the analysis of the communication about the explosion of a Dutch chemical plant revealed that the news media and the public engaged in an alignment of frames as well (Van der Meer & Verhoeven, 2013). Therefore, this study anticipates that the same applies to the implicit framing by company, government and the online news. This leads to the expectation that their implicit frames are more aligned during the second stage of the crisis compared to the first one.

H2: The implicit frames used by the company, government and online newspapers differ less

from each other during the chronic stage compared to the acute stage of a crisis. Resolution stage

Finally, during the last stage of a crisis, organizations try to restore their reputation and return to their daily routine. This phase usually lasts the longest, as the restoration of a damaged organizational image is a process and takes time (Fombrun & van Riel, 2004). Ideally, all uncertain factors about a crisis have been removed at this point and the company as well as the government can focus on rebuilding their images. However, this last stage of the crisis gives the media space for assessment and evaluation of the crisis handling, as the initial anxiety and focus on the crisis itself has faded. Especially the involvement of the government triggers intensified media scrutiny which means that the governmental crisis managers face more scrutiny during and evaluation after the crisis (Kim & Liu, 2012). Therefore, it can be said that again, all actors follow their professional identity when communicating during a phase where the crisis is not acute anymore. This means that the

actors might adopt new perspectives and aspects in their communication that are not comparable to each other anymore (van der Meer et al., 2014).

So far, research has showed that during the last phase of a crisis, the public started to criticize the government for their crisis management through social media whereas the media focused on safety aspects. Thus, the implicit frames between media and public differed accordingly (van der Meer & Verhoeven, 2013). Furthermore, in their study about several Dutch crises, van der Meer et al. (2014) found that the implicit frames from companies, news media and the public were no longer aligned at the end of these crises. Therefore, it is not far to seek that during the last stage of a crisis, company, government and news media also start to produce different frames according to their professional identities and interests. This would lead to different frame usage by all actors.

H3: The implicit frames used by the company, government and online newspapers differ less

from each other during the resolution stage compared to the chronic stage of a crisis.

Method Case

By studying a corporate crisis case, the implicit frames in the communication material provided by company, corresponding government and online newspapers are going to be explored.

The selected case for this study is the disappearance of the airplane MH370 from

Malaysia Airlines on March 8, 2014. This crisis serves as a good illustration of the theoretical framework because (1) it involved the Malaysian government as a central actor in overcoming the issue, (2) it has triggered high media attention and coverage, (3) it lies in the past and therefore allows a complete overview of the course of the crisis and (4) the selected case was

characterized with high uncertainty regarding cause and nature as well as scope and duration of the crisis.

Flight MH370 departed from Kuala Lumpur and was supposed to land in Beijing after a six-hour flight. However, shortly after departure, the plane disappeared from the radar and all contact to the aircraft broke off (Quinn & Branigan, 2014). After two weeks of many dead-end cues, Malaysia Airlines assumed that the plane must have dead-ended in the Indian Ocean, however, they did not have clear proof of that fact (Smith-Spark, 2014). The search of the plane developed into a major project, which required not only the involvement of the

Malaysian government, but also the help of several other countries (McDonell, 2014). Despite all efforts, it took more than one year until a first debris of the plane was found (“MH370: Reunion wing debris”, 2015). The unknown cause of the disappearance and location of the plane naturally triggered very high media attention. Furthermore, the investigation of the case led to a drop-by-drop information discovery. The successive release of new findings was also reflected in the media coverage, which resulted in an ongoing presence of the issue in the media. Even though the reason for the disappearance and the location of the plane has not been found until today, it can be argued that the crisis itself has calmed down to such a degree that it does not match the definition of a corporate crisis (see above) in the narrower sense anymore. In sum, this crisis serves as suitable case for the purpose of this research.

Sample

This study analyzes the frames of press releases provided by Malaysia Airlines and the Malaysian government as well as the frames that can be found in news articles from three of the most read US online newspapers, namely ‘The Huffington Post’, ‘The New York Times’ and ‘The Wall Street Journal’ (“Top 15”, 2016).

The computer assisted semantic-network analysis chosen for this study allows examining large numbers of press releases and news articles. Therefore, it is possible to analyze a census

sample, which in turn facilitates the study of frame development over time. The census consists of the entire communication material about MH370 that has been published by these three actors (Malaysia Airlines, Malaysian government and the three online newspapers combined) between March 8, 2014, the day of the disappearance of the airplane, and August 13, 2015, one week after the publishing of the last press release by Malaysia Airlines (status: 25.02.16). The sample size amounts to N = 658. The press releases from the company were downloaded through the Malaysia Airlines website. The press releases from the Malaysian government were accessed through the official site of the MH370 plane crash provided by the government.

The news articles were collected from the websites of the newspapers. The Huffington Post, a free online newspaper, The New York Times, a broadsheet newspaper, and The Wall Street Journal, which is a financial newspaper, were chosen as they enable a broad impression of the discourse.

The articles were identified by using the search string <“MH370” OR “MH 370” OR “Flight 370” OR “Malaysia Airlines”>. However, the term “Malaysia Airlines” led to an inclusion of news articles about the crash of MH17, another Malaysia Airlines crash, which happened four months after MH370. Because of the different nature of the crash, all news articles concerning MH17 were removed in order to avoid biased results. The exclusion was conducted while manually downloading the relevant news articles.

Research units

In order to examine the role of crisis phases and the development of frames over time (H1-3), the data was split into different groups according to their date of publication and analyzed separately. The first phase is defined as the first 14 days (08.03.2014 – 24.03.2014), starting from the day of the disappearance and ending on the day of the announcement by Malaysia Airlines that there are no survivors of MH370. In this specific case, two weeks

represent an accurate time frame to grasps the initial crisis communication, as Malaysia Airlines was completely lacking information on the cause of the disappearance and the whereabouts of the plane. The second period, the chronic crisis stage, lasts for 79 days after the initial phase of the crisis (25.03.2014 – 11.06.2014). During this phase, temporary procedures were set up until normal operations could resume. In the case of MH370, these temporary procedures can be operationalized as the actions directed to the search of the plane. The last phase covers the process of image restoration. It combines the ongoing search with activities that intended to rebuild Malaysia Airlines’ reputation and starts on the day of the first announcement of interim compensation for each passenger on board by Malaysia Airlines until one week after the last press release of the company (12.06.2014 -13.08.2015).

While collecting the sample, it became evident that the involvement of the government only started during the second crisis stage. Therefore, the three actors together with the three crisis stages can be summarized in eight research units, which are analyzed separately and compared to each other.

Table 1.

Overview research units hypotheses 1-3.

N Malaysia Airlines N Government N Online News Acute crisis stage

Chronic crisis stage Crisis resolution stage

32 26 15 -- 5 29 220 210 121

Automated semantic-network analysis

This study examines implicit frames in the communication material about the MH370 crisis published by Malaysia Airlines, the Malaysian government and the media. As mentioned before, implicit frames cannot be analyzed with traditional, manually coded content analysis (Hellsten et al., 2010). To understand the meaning construction of crisis situations through words and their relative position to each other, this study applies a type of

automated content analysis. This so-called automated semantic-network analysis determines implicit crisis framing as word networks by following an algorithmic approach. By measuring the meaning of words in their context, this method can identify semantic changes in

communication materials of different actors. The degree of similarity in occurrence patterns of words between the words can then be measured with similarity measures such as

correlation coefficients (Hellsten et al., 2010). Another advantage of this method is that it allows to look for frames in an inductive, exploratory way, so that unlike in traditional content analysis, frames do not have to be predefined in order to be found.

Following the guidelines for the construction of semantic maps in a publication by Vlieger and Leydesdorff (2011), several steps involving different software programs were taken in order to obtain the data. As this paper compares different crisis phases and different actors, the following process had to be conducted for each research unit separately.

Firstly, the sample of every research unit was run through the program FrequencyList. Words that are not adding any value to the analysis of frames, such as “then”, “me”, “with” etc., were removed automatically with a standardized stopwordlist. As an output, this program created a list of all the words that were used in the communication material ranked from most used to the least used. At this point, additional words could be removed manually if

necessary. Usually, this kind of analysis ends with a visualization of the implicit frames as semantic fields, where a maximum of 75 words is advisable to avoid difficulties when interpreting the data. However, in this study, several actors and phases are compared to each other and therefore, it is more convenient and clearer to present the data in a table instead of visualized networks. The omission of the visualization allowed the analysis of more words, which is why a maximum of 255 most used words was selected. This number is based on limitations determined by the computer programs. Secondly, this list of words together with the sample of each research unit served as input for the program called FullText. This program generated word/document occurrence matrices. In order to analyze latent structures

that exist in these word/document occurrence matrices, principal component factor analyses with varimax rotation were conducted. As factor analysis enables the inductive exploration of latent constructs (Field, 2009), the identified factors can be seen as implicit frames found in the analyzed communication materials. Furthermore, factor analyses enable discovering statistical correlations between words by comparing factor loadings (van der Meer et al., 2014). However, prior to the factor analysis, the variance of the variables (the 255 words) was calculated and all variables with a variance of zero were excluded from the analysis. For better comparability of the communication material, a fixed number of twelve components for all research units of hypothesis 1-3 was selected.

After obtaining the factors, in a third step, an overall image of the frame usage from Malaysia Airlines, the Malaysian government and the newspapers was distinguished. Because of the large number of different frames, the qualitative interpretation focused on the three most dominant frames with the highest eigenvalues (EV) and the highest explained variances in each research unit. Eigenvalues are an indicator of the substantive importance of each factor (Field, 2009), which is why it is possible to focus the most important frames by

following the highest eigenvalues. Variables with a factor loading above 0.3 were included in the corresponding factor, as these variables are regarded as important (Field, 2009). However, this method is based on an interpretative labeling approach and thus, is a subjective

comparison of frames. Therefore, in a fourth step, this study compared the implicit frames based on statistical indicators. Following a newly developed analytical process (van der Meer et al., 2014), the factor loadings of words that are used by two actors at a certain stage of the crisis were compared to each other. The factor loadings imply the extent of involvement, i.e. importance of the word in the implicit frame. The correlation coefficient Spearman’s rho indicates the degree to which the words had the same importance in meaning provision and framing for these two actors in that particular stage of the crisis. The correlations can be deemed relevant rather than the significance level, as this paper analyzed a census sample.

Results

In the following section, the identified frames used in the communication material by Malaysia Airlines, the Malaysian government and the three online newspapers were compared to each other. This analysis was conducted for each crisis stage separately in order to identify frame development across crisis stages.

Table 2.

Correlations among all actors.

Malaysia Airlines / Online News Government / Online News Malaysia Airlines / Government Spearman’s rho correlation coefficient ρ Acute crisis stage Chronic crisis stage Crisis resolution stage

0.037 0.137 0.243 -- -0.158 0.142 -- 0.109 0.322

Hypothesis 1 stated that the frames used by the company, government and online newspapers would differ from each other during the acute stage of a crisis. However, while gathering the sample, it became evident that the government was not yet involved in the communication during the acute stage of the crisis. Therefore, only the implicit frames used by Malaysia Airlines and the online newspapers could be compared. For this comparison, the frames were first interpreted qualitatively, by comparing the three most dominant frames (with the highest eigenvalue) from both actors and second, assessed quantitatively by calculating and comparing factor loadings correlation coefficients.

When looking at the frames from Malaysia Airlines and the three online newspapers, it is visible that some of their thematic priorities differ. Malaysia Airlines made use of frames that can be named “search information frame” (EV = 62.12, Cronbachs α = 0.95), “country cooperation frame” (EV = 22.93, Cronbachs α = 0.96) and “Malaysian ministries frame” (EV = 21.24, Cronbachs α = 0.96). The frames from the newspapers partially overlap with those

from Malaysia Airlines, especially when it comes to the search of the plane. However, they also direct attention to other topics. The three most dominant frames are named “Australia involvement frame” (EV = 25.55, Cronbachs α = 0.9), “search information frame” (EV = 13.19, Cronbachs α = 0.88) and “speculation frame” (EV = 9.75, Cronbachs α = 0.87).

During this first stage of the crisis, the words attributed to the frames differ between these two actors. Malaysia Airlines framed the crisis with words such as ‘surveillance’ (0.97), ‘nautical’ (0.94), ‘committed’ (0.91) or ‘navy’ (0.91), indicating that Malaysia Airlines initially focused on the search aspects. By contrast, for the online newspapers, words like ‘object’ (0.82), ‘debris’ (0.78), ‘stolen passport’ (0.92) or ‘travel’ (0.83) formed the dominant frames, demonstrating that the media put an emphasis on search cues and speculations about what might have happened to the plane.

To address H1 quantitatively, the correlations of factor loadings between the two actors were compared with the correlation coefficient Spearman’s rho. Table 2 shows that the correlation between Malaysia Airlines and the newspapers during the acute stage of the crisis is ρ = 0.037. This coefficient leaning towards zero indicates a weak correlation between the two actors. Low factor correlations between actors suggest that the mutually used words are ascribed different degrees of importance when it comes to framing (van der Meer et al., 2014). By taking a closer look at specific words, weak correlations of word importance between the implicit frames from Malaysia Airlines and the online newspapers can be

illustrated. For example, the words ‘surveillance’, ‘square’ and ‘analysis’ scored considerably higher factor loadings in the frames of the Malaysia Airlines’ communication material (0.97, 0.85, 0.78) than in the frames of the online newspaper articles (0.34, 0.39, 0.33).

When combining the qualitative and quantitative interpretation of the data it can be concluded that the frames from Malaysia Airlines and the online newspapers differed in the first phase of the crisis, however, the government was not communicatively present yet.

Therefore, it is not possible to compare the frames of all three actors and H1 can only be partially supported.

For H2, the similarity of the frames used by Malaysia Airlines, the Malaysian

government and the online newspapers during the chronic stage of the crisis were compared to the similarity of the frames in the acute stage of the crisis. For this comparison, the most dominant frames were again first interpreted qualitatively. The government joined the

communication during this stage of the crisis, which is why for this hypothesis all three actors could be compared to each other.

When looking at the frames in general, it is noticeable that the frames from two actors evolved around the search of the plane. Malaysia Airlines focused on the technical and political aspects of the search, and the frames are labeled “search activities frame” (EV = 39.87, Cronbachs α = 0.96), ‘ICAO frame’ (International Civil Aviation Organization) (EV = 30.64, Cronbachs α = 0.94), or ‘political frame’ (EV = 20.11, Cronbachs α = 0.93), whereas the newspapers put an emphasis on the cues and information that come from the investigation of the missing plane but also included political aspects. Their frames are named ‘search instruments frame’ (EV = 24.76, Cronbachs α = 0.93), ‘region frame’ (EV = 15.77, Cronbachs α = 0.81), or ‘political frame’ (EV = 10.35, Cronbachs α = 0.89). The frames from the

government are more diverse. The implicit frames can be labeled ‘Condolences frame’ (EV = 42.61, Cronbachs α = 0.97), ‘political frame’ (EV = 37.67, Cronbachs α = 0.98) or

‘investigation frame’ (EV = 20.6, Cronbachs α = 0.98). The words that formed these dominant frames in the newspapers are ‘signal’ (0.82), ‘shield’ (0.72), ‘maritime’ (0.72) or ‘object’ (0.72), indicating that the media continued to focus on search cues. Malaysia Airlines also concentrated on the investigation, as the most prominent words are ‘analysis’ (0.97), ‘station’ (0.97), ‘ground’ (0.97) or ‘AAIB’ (Air Accident Investigation Branch) (0.95). The prominent words used in the frames of the government, however, show different fields of

attention: the words ‘families’ (0.95), ‘arrangements’ (0.95), ‘importance’ (0.95),

‘investigation’ (0.97) or ‘report’ (0.98) tend to focus on the victims’ families as well as the search of the plane.

As the government was not active in their communicative role during the first stage of the crisis, it is necessary to look at the correlation of the factor loadings of mutually used words by company and online newspapers first, in order to be able to compare the similarity of these implicit frames to the ones from the first stage of the crisis. During the second stage of the crisis, the correlation between Malaysia Airlines and the online newspapers is ρ = 0.137, demonstrating a weak correlation between the two actors concerning word importance in meaning provision. Compared to the correlation coefficient from the acute stage of the crisis (ρ = 0.037), it can be said that the frames were slightly more similar in the chronic stage.

As can be seen from table 2, the factor loading correlations during the chronic phase between government and online newspapers is ρ = -.0.158, whereas between government and Malaysia Airlines ρ = 0.109. These correlations suggest that the same words used in the communication material from the government and those from the news media did not have the same importance when it came to the establishment of meaning and frames, whereas the word importance of governmental and corporate communication material showed a slightly positive correlation. A listing of factor loadings of some words can illustrate these

correlations: the words ‘communication’, ‘ocean’ and ‘Perth’ load highly similar on the frames from Malaysia Airlines (0.41, 0.62, 0.58) and on the frames from the online

newspapers (0.43, 0.64, 0.59). However, words such as ‘accident’, ‘board’ or ‘underwater’ have different factor loadings when the government (0.98, 0.92, 0.17), and the online newspapers (0.34, 0.21, 0.58) are compared. Finally, words like ‘accordance’, ‘analysis’ or ‘kilometer’ indicate more similar factor loadings between the company (0.84, 0.97, 0.86) – government (0.73, 0.84, 0.87) pair. Overall, the correlations across all actors is

ρ = (0.137+-.0.158+0.109)/3 = 0.029, representing a very weak correlation of word importance among all actors.

Because just the factor loadings correlation of the frames from Malaysia Airlines and online newspapers during the chronic stage can be compared to the correlations from the acute stage of the crisis, H2 is only partially supported. The inclusion of the government, however, shows that the implicit frames among all actors are not aligned during the chronic stage of the crisis.

H3 expected the frames used by Malaysia Airlines, the Malaysian government and the online news media to differ more from each other during the resolution stage compared to the chronic stage of the crisis. In order to assess this hypothesis, the correlation of factor loadings of frames from the pairs newspapers – Malaysia Airlines, newspapers – Malaysian

government and Malaysia Airlines – Malaysian government were compared across these two crisis phases. The evaluation of the overall impression of the three most dominant frames from all three actors show a mixed picture. On the one hand, the ‘debris discovery frame’ was prominent in communication material from the online newspapers (EV = 23.67, Cronbachs α = 0.92) as well as from the government (EV = 17.1, Cronbachs α = 0.94). On the other hand, the frame ‘search frame’ was used by Malaysia Airlines (EV = 65.59, Cronbachs α = 0.95) and the government (EV = 36.16, Cronbachs α = 0.94). Furthermore, the newspapers started to extend the coverage on the topic, for example by connecting the disappearance of MH370 to the crash of the Germanwings airplane ‘Germanwings frame’ (EV = 19.78, Cronbachs α = 0.89) or by applying the frame ‘safety regulations frame’ (EV = 16.12, Cronbachs α = 0.92), whereas the company and the government stayed focused on technical and economic aspects of the search. The dominant words attributed to these frames of the online newspapers were ‘debris’ (0.81), ‘piece’ (0.73), ‘Réunion island’ (0.73) but also ‘Lubitz’ (0.93), ‘cockpit’ (0.92) and ‘pilot’ (0.91). This shows that on the one hand that the media was focused on the

progress of the plane search and on the other hand, that they started to relate the MH370 case to other subjects, such as the Germanwings crash. The most important words that formed the frames in the communication material of Malaysia Airlines were ‘deepsea’ (0.99), ‘arc’ (0.99), ‘conclusion’ (0.99), ‘ATSB’ (Australian Transport Safety Bureau) (0.94) or ‘interim procedure’ (0.99), demonstrating the company’s focus on the investigation. The words used by the government show some overlap, as for example words as ‘deepsea’ (0.97) and ‘ATSB’ (0.86) loaded highly on the factors as well.

The average correlation of factor loadings among all three actors is ρ =

(0.243+0.142+0.322)/3 = 0.236. When compared to the chronic stage of the crisis (ρ = 0.029), it becomes evident that there is an overall alignment of word importance during the resolution stage of the crisis. For example, there is clear similarity of factor loadings for words such as ‘operation’, ‘found’ or ‘public’ in the frames of the government (0.65, 0.75, 0.50), Malaysia Airlines (0.67, 0.75, 0.55) and the newspapers (0.7, 0.62, 0.49). Thus, hypothesis 3 could not be supported.

Conclusion and discussion

This paper set out to answer the question whether and to what extent the frames used by online newspapers, the company going through a crisis and the corresponding government differ from each other and how this (dis-)similarity relates to the different phases of a crisis. By means of semantic-network analyses, the entire communication material that was

published during the MH370 crisis from Malaysia Airlines, the Malaysian government and three of the most read online newspapers was analyzed.

The study showed that during the acute stage of the crisis, the implicit frames used by Malaysia Airlines and the online newspapers were not aligned. This result indicates that in the

beginning of the crisis, these two actors tried to make sense of the crisis and at the same time followed their professional identity, which resulted in competing frames. The online

newspapers, for example, started to speculate about the cause of the disappearance, which is a way of gaining the audience’s attention. This finding supports the theoretical framework of sense-making (Weick, 1988; Cornelissen et al., 2010) and is in line with previous research on the dynamics between these two actors (Schultz et al., 2012; van der Meer et al., 2014). However, it is interesting to see that during the first 14 days, the government did not release any official statements on their website, which implies that the government was not yet active in their communicative role. This discovery might be explained by the fact that the crisis started out to be a corporate crisis. The involvement of the government stipulates a certain magnitude of the crisis, indicating that the company is not able to solve the issue without external help. However, the decision to bring in the government takes time, as the company initially tries to solve the problem itself. This course of events can also be observed in other corporate crises that involved the corresponding government. For example, during the BP oil spill crisis in 2010, the US government released their first statement 10 days after the

explosion, in which they explain governmental arrangements and actions that are going to be performed (“Press Briefing on the BP Oil Spill in the Gulf Coast“, 2010). By then, it has become apparent that BP needed support to overcome the crisis and its consequences. In sum, the belated participation of the government in the MH370 crisis can be seen as a characteristic of the initial communication dynamic among all actors. The analysis showed that the frames between online newspapers and the company in crisis differed and that the government was not communicatively present in the beginning of the crisis, lending partial support to H1.

Furthermore, the research revealed that during the second phase, the chronic stage of the crisis, the frames of the media and the company slightly converged. This finding supports the existing research on frame alignment and may indicate that these two actors began to engage in mutual influencing and borrowing of their frames (Leydesdorff & Ivanova, 2014; van der

Meer et al., 2014). The government started to actively participate in the communication about the crisis during this period. However, the inclusion of the government shows that the word importance for meaning provision strongly differ when compared to the media and imply that there is no frame alignment during this phase of the crisis. It could be argued that as the government started to take part in the communication later than the other two actors, it has also engaged in sense-making at a later stage. As organizational sense-making primarily deals with the question of how something has become an event for organizational members and what the meaning of the event is (Weick et al., 2005), it is logical that the sense-making process only starts after an event has become relevant to an organization. As a result, the frames that are used by the government “lag behind” and are dealing with other aspects that may relate to their societal function. This finding also illustrates the aggravated entrance of the government into the existing discourse between company and media, underlining the importance of timely communication during crisis situations. In conclusion, the frames used by online newspapers and the company slightly aligned in the second phase of the crisis, however, the government did not engage in frame alignment. Therefore, H2 is partially supported.

Another interesting finding is that during the last stage of the crisis, the frames of all three actors clearly aligned, contrary to the expectation that they would de-align (H3). This finding could be related to the nature of this specific crisis. The disappearance of MH370 triggered high uncertainty for a very long time, as there was no information on the

whereabouts of the plane or the cause of the disappearance at all (Hodal, 2014). It seems as if such a high degree of uncertainty could influence and delay the process of frame (de-)

alignment. If the actors lack information they can process in order to make sense of the issue, it is difficult to find common ground and to engage in mutual influencing and orienting when making sense and creating frames. Therefore, the frames become more comparable, once there is more information available. In the case of MH370, at the time of the resolution stage,

the investigation had already persisted for several weeks. Even though the plane had not yet been found, there was more information available because of the search activities, which in turn made it easier for the actors to cooperate and involve in a more collective process of sense-making. When following the belated frame alignment consequently, it is imaginable that the expected de-alignment phase was delayed as well, and thus, may have taken place after the last phase that was analyzed in this paper.

This study adds to the crisis communication literature by empirically analyzing the communication dynamics of the company in crisis, the corresponding government and online newspapers. The exploration of the concept of implicit framing made it possible to show that the processes of sense-making and frame alignment might be influenced by the delayed appearance of governments during corporate crises. By means of automated semantic-network analyses, this paper contributes methodologically to a less prominent field of

empirical methods. By validating a newly developed analytical technique (van der Meer et al., 2014), it was possible to quantitatively assess the development of implicit frames over time among different actors and crisis stages.

Despite the theoretical and methodological contributions, this study faces some

limitations. First, the case selection might have influenced the results, as the disappearance of MH370 was an extreme case in terms of uncertainty. However, as almost every crisis is characterized by some degree of uncertainty (Lu, 2013), it is still relevant for crisis managers to be aware of the possible impacts that uncertainty can have on the dynamics among actors regarding crisis communication. Second, also related to the case study is the aspect of generalizability. Case studies come with a restriction as they are never completely

transferable to other cases (Easton, 2010). Yet, every crisis is unique and therefore, studies on specific crisis cases can only enrich the body of knowledge by adding different perspectives and aspects regarding this field. Furthermore, causal explanations that have been produced for one case can serve as a foundation for developing theory beyond that case (Easton, 2010).

The insights resulting from the study allow to draw practical implications for corporate and governmental organizations. First, it is crucial for organizations in crises to engage in immediate communication in order to keep up with the fast moving environment of news media in general, and online newspapers in specific. It can be difficult for a third party to join an established discourse between key actors at a later point. Second, crisis managers should take into consideration how information provision can influence and promote frame

alignment. Especially in situations with high uncertainty, actors need information in order to start a more collective process of meaning construction, which ideally would evolve around the provided information.

Future research should further investigate the role of governments in corporate crises, as they seem to have an influence on the communication dynamics that is unlike the influence of other actors, such as the public (e.g. van der Meer & Verhoeven, 2013; van der Meer et al., 2014). In addition, scholars should also continue to direct attention to the aspect of

uncertainty during crises, as it can complicate the interplay among different actors and impede effective communication. Lastly, it would be interesting to extend existing research on the differences in framing behavior between online and print news media, as in today’s society immediate online access to information for everyone has become a challenge for

Reference list

Benoit, W. L. (1995). An Analysis of Sear’s Repair of its Auto Repair Image: Image restoration discourse in the corporate sector. Communication Studies, 46(1), 89-105. Benoit, W. L. (1997). Image Repair Discourse and Crisis Communication. Public Relations

Review, 23(2), 177-186.

Bowen, S. A., & Zheng, Y. (2015). Auto recall crisis, framing, and ethical response: Toyota’s missteps. Public Relations Review, 41(1), 40-49.

Bucher, H. J. (2002). Crisis Communication and the Internet: Risk and Trust in a Global Media. First Monday, 7(4), retrieved from

http://firstmonday.org/ojs/index.php/fm/article/view/943

Castells, M. (2007). Communication, Power and Counter-power in the Network Society.

International Journal of Communication, 1, 238-266.

Chakraborty, S., & Creutzfeldt-Banda, N. (2011). Initial Phase Crisis Communications Following High Perceived Risk Events: The Volcanic Ash Crisis and the Japanese Tsunami as Examples. European Journal of Risk Regulation, 2(2), 265-267. Chong, D., & Druckman, J. N. (2007). A Theory of Framing and Opinion Formation in

Competitive Elite Environments. Journal of Communication, 57, 99-118.

Coombs, T. (2007). Protecting Organization Reputations During a Crisis: The Development and Application of Situational Crisis Communication Theory. Corporate Reputation

Review, 10(3), 163-176.

Coombs, W. T., & Holladay, S. J. (2004). Reasoned action in crisis communication: An attribution theory-based approach to crisis management. In D.P. Millar & R.L. Heath (Eds.), Responding to Crisis: A Rhetorical Approach to Crisis Communication (pp. 95-115). Mahwah, NJ: Lawrence Erlbaum Associates.

Coombs, W. T., & Holladay, S. J. (2011). An exploration of the effects of victim visuals on perceptions and reactions to crisis events. Public Relations Review, 37(2), 115-120. Cornelissen, J. P., Carroll, C. E., & Elving, W. J. L. (2010). Making Sense of a Crucial

Interface: Corporate Communication and the News Media. In L. Chouliaraki & M. Morsing (Eds.), Media, Organizations and Identity (pp. 129-145). New York: Palgrave Macmillan.

Easton, G. (2010). Critical realism in case study research. Industrial Marketing Management,

39, 118-128.

Entman, R. M. (1993). Framing: Toward a Clarification of a Fractured Paradigm. Journal of

Communication, 43(4), 51-58.

Field, A. (2009). Discovering statistics using SPSS. London: Sage Publications Limited. Fink, S. (1996). Crisis management: Planning for the inevitable. New York: American

Management Association.

Fombrun, C. J., & van Riel, C. B. M. (2004). Fame & Fortune. How Successful Companies

Build Winning Reputations. New York: Pearson Education.

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.

Graham, M. W., Avery, E. J., & Park, S. (2015). The role of social media in local government crisis communications. Public Relations Review, 41(3), 386-394.

Hallahan, K. (1999). Seven models of framing: Implications for public relations. Journal of

Public Relations Research, 11(3), 205-242.

Harcup, T., & O'Neill, D. (2016): What is news? News values revisited (again). Journalism

Studies, 1-19. doi: 10.1080/1461670X.2016.1150193

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

Tampere: Tampere University Press.

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.

Higgins, E. T. (1996). Knowledge activation: accessibility, applicability, and salience. In E. T. Higgins, & A. W. Kruglanski (Eds.), Social Psychology: Handbook of Basic

Principles (pp. 133-168). New York: Guilford.

Hodal, K. (2014, March 14). Flight MH370: a week of false leads and confusion in hunt for missing plane. The Guardian. Retrieved from

https://www.theguardian.com/world/2014/mar/14/flight-mh370-false-leads-confusion-hunt-missing-plane

Holladay, S. J., & Coombs, W. T. (2013). Successful prevention may not be enough: A case study of how managing a threat triggers a threat. Public Relations Review, 39(5), 451-458.

Horsley, J. S., & Barker, R. T. (2002). Toward a Synthesis Model for Crisis Communication in the Public Sector. An Initial Investigation. Journal of Business and Technical

Communication, 16(4), 406-440.

Huang, Y. H. (2006). Crisis situations, communication strategies, and media coverage: a multicase study revisiting the communicative response model. Communication Research,

33(3), 180-205.

Jaques, T. (2007). Issue management and crisis management: An integrated, non-linear, relational construct. Public Relations Review 33(2), 147-157.

Johnson, D. E. (2014). The mass media’s dual role: ‘watchdog’ and guardian of their own interests. Journal of Public Affairs, 14(3), 369-378.

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.

Karlsson, M. (2011). The immediacy of online news, the visibility of journalistic processes and a restructuring of journalistic authority. Journalism, 12(3), 279-295.

Karlsson, M. (2012). The online news cycle and the continuous alteration of crisis frames: A Swedish case study on how the immediacy of online news affected the framing of the swine flu epidemic. Journal of Organisational Transformation & Social Change, 9(3), 247-259.

Kim, S., & Liu, B. F. (2012). Are All Crises Opportunities? A Comparison of How Corporate and Government Organizations Responded to the 2009 Flu Pandemic. Journal of Public

Relations Research, 24(1), 69-85.

Kleinnijenhuis, J., Schultz, F., & Oegema, D. (2015). Frame Complexity and the Financial Crisis: A Comparison of the United States, the United Kingdom, and Germany in the Period 2007–2012. Journal of Communication, 65(1), 1-23.

Krippendorff, K. (1980). Content Analysis: An Introduction to Its Methodology. Thousand Oaks, CA: Sage.

Lee, K. (2009). How the Hong Kong government lost the public trust in SARS: Insights for government communication in a health crisis. Public Relations Review, 35(1), 74-76. Leydesdorff, L., & Hellsten, I. (2005). Metaphors and diaphors in science communication

mapping the case of stem cell research. Science Communication, 27(1), 64-99. Leydesdorff, L., & Hellsten, I. (2006). Measuring the meaning of words in contexts: An

automated analysis of controversies about 'Monarch butterflies, 'Frankenfoods,' and 'stem cells'. Scientometrics, 67(2), 231-258.

Leydesdorff, L., & Ivanova, I. A. (2014). Mutual redundancies in inter-human communication systems: Steps towards a calculus of processing meaning. Journal of the American

Society for Information Science and Technology, 65(2), 386-399.

Liu, B. F., & Horsley, J. S. (2007). The Government Communication Decision Wheel: Toward a Public Relations Model for the Public Sector. Journal of Public Relations

Research, 19(4), 377–393.

Liu, B. F., Horsley, J. S., & Levenshus, A. B. (2010). Government and Corporate Communication Practices: Do the Differences Matter?, Journal of Applied

Communication Research, 38(2), 189-213.

Lu, X. (2013). Managing Uncertainty in Crisis: Exploring the Impact of Institutionalization

on Organizational Sensemaking. (Unpublished doctoral dissertation). Utrecht University,

Utrecht.

Massey, B. L., & Levy, M. R. (1999). “Interactive” online journalism at English-language web newspapers in Asia. A dependency theory analysis. Gazette, 61(6), 523-538. McCombs, M., Shaw, D. L., & Weaver, D. (1997). Communication and democracy:

Exploring the intellectual frontiers in agenda-setting theory. Mahwah, NJ: Lawrence

Erlbaum Associates Inc.

McDonell, S. (2014, March 17). Missing flight MH370: Malaysia calls on 25 countries in massive search for Malaysia Airlines plane. ABC. Retrieved from

http://www.abc.net.au/news/2014-03-16/malaysia-calls-on-25-countries-in-search-for-malaysia-airlines/5324596

MH370: Reunion wing debris 'certainly' from missing flight. (2015, September). BBC News. Retrieved from http://www.bbc.com/news/world-asia-34145127

Millar, D. P., & Heath, R. L. (2004). Responding to crises. A rhetorical approach to crisis

Myers, K. (1993). Total contingency planning for disasters: Managing risk...minimizing

loss...ensuring business continuity. New York: John Wiley.

Perry, D. C., Taylor, M., & Doerfel, M. L. (2003). Internet-based communication in crisis management. Management Communication Quarterly, 17(2), 206-232.

Press Briefing on the BP Oil Spill in the Gulf Coast (2010, April, 29). The White House. Retrieved from https://www.whitehouse.gov/the-press-office/press-briefing-bp-oil-spill-gulf-coast

Quinn, B., & Branigan, T. (2014). Malaysia Airlines loses contact with plane carrying 239 people. The Guardian. Retrieved from

http://www.theguardian.com/world/2014/mar/08/malaysia-airlines-loses-contact-plane Reese, S., Rutigliano, L., Hyun, K., & Jeong, J. (2007). Mapping the blogosphere:

Professional and citizen-based media in the global news arena. Journalism, 8(3), 235-261.

Regester, M., & Larkin, J. (2008). Risk issues and crisis management in public relations: A

casebook of best practice. London: Kogan Page.

Schultz, F., Kleinnijenhuis, J., Oegema, D., Utz, S., & Van Atteveldt, W. (2012). Strategic framing in the BP crisis: A semantic network analysis of associative frames. Public

Relations Review, 38(1), 97-107.

Schultz, F., & Raupp, J. (2010). The social construction of crises in governmental and corporate communications: An inter-organizational and inter-systemic analysis. Public

Relations Review, 36(2), 112-119.

Seeger, M. W. (2002). Chaos and crisis: propositions for a general theory of crisis communication. Public Relations Review, 28(4), 329-337.

Seeger, M., Sellnow, T., & Ulmer, R. (2003). Communication and organizational crisis. Westport, CT: Praeger.

Smith-Spark, L. (2014, August 05). How Malaysia missteps on MH370 eroded families' trust.

CNN. Retrieved from

http://edition.cnn.com/2015/08/06/asia/mh370-malaysia-investigation-missteps/

Steinberg, M. W. (1998). Tilting the frame: Considerations on collective action framing from a discursive turn. Theory and Society, 27(6), 845-872.

Taylor, J. R., Van Every, E. J. (2000). The Emergent Organization: Communication as Its Site

and Surface. Mahwah, NJ: Lawrence Erlbaum Associates.

Top 15 Most Popular News Websites (2016). The eBusiness Guide. Retrieved from http://www.ebizmba.com/articles/news-websites

Van der Meer, T., Verhoeven, P., Beentjes, H., & Vliegenhart, R. (2014). When frames align: The interplay between PR, news media, and the public in times of crisis. Public Relations

Review, 40(5), 751-761.

Van der Meer, T., & Verhoeven, P. (2013). Public framing organizational crisis situations: Social media versusnews media. Public Relations Review, 39(3), 229-231.

Vlieger, E., & Leydesdorff, L. (2011). Content analysis and the measurement of meaning: The visualization of frames in collections of messages. The Public Journal of Semiotics,

3(1), 28-50.

Weick, K. E. (1988). Enacted sense making in crisis situations. Journal of Management

Studies, 25, 305-317.

Weick, K. E. (1995). Sensemaking in Organizations. Thousand Oaks, CA: Sage.

Weick, K. E., Sutcliffe, K. M., & Obstfeld, D. (2005). Organizing and the process of sense making. Organizational Science, 16, 409-421.

Appendix Rotated varimax component matrix

This matrix is a small excerpt of the rotated varimax component matrix of the online newspapers during the first phase of the crisis, showing all absolute values above 0.1.

Components Words 1 2 3 4 5 6 7 8 9 10 11 12 Australian Object Debris Authority Image Ocean Maritime Search System Commu-nication Investi-gator Plane Data Ground Strait Civilian Chief Radar Etc. 0.84 0.82 0.78 0.68 0.68 0.67 0.65 0.65 0.15 0.11 -0.1 0.23 0.71 0.62 0.61 0.6 0.58 0.58 0.11 0.19 0.38 -0.1 0.16 0.14 0.1 0.33 0.17 0.34 0.72 0.70 0.67 0.63 0.19 0.23 0.23 0.17 0.31 0.34 0.45 0.12 0.17 -0.1 0.21 -0.1 0.14 0.12 0.24 0.31 0.12 0.11 0.21 -0.15 0.24 -0.12 -0.1 0.22 0.14 0.27 0.24 -0.11 0.12 -0.12 0.21 0.15 0.16 0.3 -0.18 -0.1 -0.21 0.26 -0.13 -0.12 -0.14 0.28

Dataset comparison factor loadings

This table is an illustration of a small part of the dataset constructed for factor-loading comparison of mutually used words between the online newspapers and Malaysia Airlines in the first phase of the crisis.

Words Factor loading

online newspapers Factor loadings Malaysia Airlines Australian Object Authority Image Ocean Search Investigator Plane Data Navy Force Aircraft China Etc. 0.61 0.75 0.56 0.85 0.65 0.51 0.63 0.88 0.78 0.91 0.87 0.78 0.76 0.84 0.82 0.37 0.68 0.67 0.66 0.37 0.59 0.58 0.67 0.51 0.47 0.67