Exploring Variations in Corporations’ Communication After a CA Versus CSR Crisis: A Semantic Network Analysis of Sustainability Reports

Abstract

The study attempts to understand corporations’ efforts to communicate their values and commitment to stakeholders after a crisis. Specifically, the study explores the characteristics of communication efforts that may differ depending on the reputational crisis types: corporate ability (CA) and corporate social responsibility (CSR) crises. Employing a series of semantic network analyses, the study examined the sustainability annual reports of two Korean airlines (i.e., Korean Air and Asiana Airlines) published before and after their recent crises. Results showed how sustainability reports’ central keywords, social issues the companies support, and prioritized stakeholders varied in response to the different types of crises. Word frequency results showed that there was an increasing trend in emphasizing the word “safety” after both types of crisis, while a noticeable decrease in emphasis on the word “ethics” was observed after CA crisis. The results of semantic network analyses showed that Korean Air’s sustainability reports seemed to focus more on aspects of the relationship with stakeholders after the CSR crisis, while Asiana Airlines appeared to place more emphasis on business-related notions after the CA crisis. Theoretical and practical implications are discussed.

Keywords

reputational crisis CA crisis CSR crisis sustainability reporting semantic network analysis

Crisis is a threat for organizational performance and sustainability. Yet it also offers an opportunity to reshape and reform its governance (Jaques, 2007; Seeger & Ulmer, 2001). Scholars have emphasized the importance of post-crisis management and communication to rebuild the tarnished reputations of corporations and increase the trust of stakeholders (e.g., Coombs, 2007; Y. Kim, 2016). Despite the importance of this issue, the vast majority of crisis communication research has focused on examining the effects and the effectiveness of the immediate crisis response strategies (e.g., J. Kim et al., 2009; Liu et al., 2011).

To better understand corporations’ efforts to communicate their values and commitment to stakeholders in the long-term perspective, this study proposed to conduct semantic network analysis to examine the sustainability annual reports of companies that experienced a reputational crisis. Mainly concerned about a consequence of an incident that can damage a specific aspect of reputation, Sohn and Lariscy (2014) suggested two types of crises: corporate ability (CA) crises and corporate social responsibility (CSR) crises. A CA crisis refers to a “critical event that adversely affects reputation associated with expertise of product and service,” whereas a CSR crisis refers to an “event that poses a threat to reputation associated with norms and values cherished by society, and socially expected obligations” (Sohn & Lariscy, 2014, p. 25). For example, a product-harm crisis involving explosion of laptop computer batteries is considered a CA crisis and an ethical-harm crisis involving the leader’s misbehavior that violates ethical standards is considered a CSR crisis. Past research surveying European communication professionals indicated that a product-harm (i.e., CA) crisis and management or leadership (i.e., CSR) crises were the types of crises they encountered most frequently, following institutional crises (Verhoeven et al., 2014). However, there are only a few studies focused on comparing these subtypes of crises.

The purpose of this study is to examine characteristics of communication efforts that may differ depending on the reputational crisis type: CA crisis or CSR crisis. As an example of CA crises, the study selects the Asiana Airline crash at the San Francisco International Airport. For CSR crises, the study selects Korean Air’s “Nut Rage Return,” where the vice president of Korean Air violated airline safety laws and mistreated a cabin crew member.

Given that crisis type and attributions of crisis may inform corporate governance and communication efforts including CSR commitment (e.g., Coombs, 2007; H. S. Kim & Lee, 2015; S. Kim & Choi, 2018), the study analyzes the pre- and post-crisis sustainability reports of two corporations. A company uses the report to communicate what it has achieved with regard to environmental, social, and economic issues, and what it intends to achieve in the future (Willis, 2003). By adopting semantic network analysis, which allows understanding of semantic relations among words based on their co-occurrence (S. Lee et al., 2009), this study also documented evidence about how an organization’s sustainability communication efforts changed over time.

Literature Review

Organizational Crisis and Crisis Type

A crisis is an unexpected event that results in high uncertainty and threatens an organization’s goals and performance as represented by the amount of financial, physical, environmental, or emotional harm (Coombs, 2007, 2015). A crisis is also a threat or challenge to an organization’s legitimacy (Jaques, 2007). In exploring a reputational crisis, Sohn and Lariscy (2014) defined it as “a major event that has the potential to threaten collective perceptions and estimations held by all relevant stakeholders of an organization and its relevant attributes” (p. 24) and suggested two types of reputational crises (i.e., CA and CSR). Building on Brown and Dacin’s (1997) two dimensional systems of corporate association, Sohn and Lariscy (2014) argue that corporate moral aspects conflicted with social norms and values cause a distinct crisis from CA associated with negative evaluations from corporations’ products or services.

Specifically, a CSR crisis is caused by a corporate morality defection, which violates socially shared norms and expected obligations including environmentally friendly business practices and human rights protections (Sohn & Lariscy, 2014). The present study considers CSR crises more broadly, which embrace ethical-harm crises, such as workplace violence, an organization’s inappropriate practices deception, law/regulation violation, and organizational misdeeds (e.g., Coombs, 2007). According to an article by Forbes, of the 10 biggest business scandals, five crises were attributed to unethical business operations or the unethical behavior of CEOs (Shen, 2017). In contrast, corporations experience ability-related crises (i.e., CA crisis) when expertise about their products and services is associated with negative evaluations or accidents (Sohn & Lariscy, 2014). Dawar and Pillutla (2000) defined product-harm crises as well-publicized crises involving products that are defective or dangerous to consumers. The specific causes of the CA crises include a technical error (i.e., mega damage, technical breakdown accidents, and recalls), and human errors (i.e., human breakdown accidents and recalls: Coombs, 2007). Sohn and Lariscy (2014) pointed out that stakeholders use the crisis dimensions as a cue to understand the situation and process information accordingly. Because different attributes of crises influence the way stakeholders interpret the events, it may damage a particular aspect of corporate reputation, corporate morality and ability.

In dealing with crisis situations, effective communications are crucial to restore organizational reputation and to rebuild trust with stakeholders (Coombs, 2007; J. Kim et al., 2009; Y. Kim, 2016). To optimize the effects of crisis communication, crisis strategies should be specified according to crisis type. Although the specific guidelines are still controversial, findings of previous research imply that corporations expect different levels of communication effectiveness, depending on which crisis strategies they choose for the crisis typology. For example, P. H. Kim et al. (2004) showed that a denial might be the best solution in an integrity violation matter, while an apology is the most effective strategy for repairing trust from issues of competence. On the other hand, Coombs and Holladay (2002) suggested that the more responsibility stakeholders attribute to the organization, the more the crisis strategy must seem to accept responsibility for the situation.

Sustainability Reporting

An organizational crisis may influence its strategic goals and business objectives, and the value it provides for the stakeholders. Etzion and Ferraro (2010) note that sustainability reports reflect a corporation’s values and commitment. Therefore, this study uses sustainability reports to gauge how the organization’s communication efforts after a crisis event. Unlike other communication channels that provide an immediate corporate response after a crisis occurs, such as a press release or social media posts, sustainability reports are published on an annual basis in a consistent manner. Corporations also publish their reports with fully controlled messages and manners (S. Y. Lee, 2016). Thus, the sustainability reports reflect the organization’s deliberative communication strategy without external interference.

Organizations disclose their CSR efforts and social commitments voluntarily through diverse media channels, including annual CSR or sustainability reports (Hahn & Kühnen, 2013; S. Y. Lee, 2016). Sustainability reports have been used as a channel between a corporation and its stakeholders to communicate its CSR activities, such as preventing human rights violations and environmental pollution and supporting fair international trade and production (Kolk, 2003). Sustainability reports also include the code of ethics for the organization (Ulmer, 2001), which builds social legitimacy for its actions. Sustainability reports tend to be written following a template with instrumental logic and institutional topics that industries share (Kolk, 2003; Willis, 2003), while deviations from the sustainability reports template might be shown when a corporation intends to present innovative values or strengthen existing values (Kolk, 2003). In doing so, sustainability reporting may reflect a corporation’s values and communication intentions.

This study assumes corporations not only differently describe the two crises but also have variations in mentioning CA and CSR crises on sustainability reports because of the degrees of expected damages. In these regard, the following research questions are proposed:

Research Question 1: How does a crisis type (i.e., CA vs. CSR) differently emphasize the crisis-related terms (i.e., ethics and safety) in sustainability reports, and how do they differ by before and after the crisis?

Research Question 2: How does a crisis type (i.e., CA vs. CSR) relate to the central keywords in sustainability reports, and how do they differ by before and after the crisis?

Corporate Engagement as Crisis Management Strategy

As a crisis management strategy, this study highlighted on corporate sustainability or CSR efforts described in sustainability reports. Beyond responding reactively to the crisis, growing attention has also been given to a longer-term process of organizational renewal (Jaques, 2007). Given that crisis provides opportunities to reshape organizational structure, the forward-looking commitment is critical in the aftermath of a crisis. For example, if a company went through a CA crisis, post-crisis management and communication should be centered on how the organization will make its products safer in the future to regain legitimacy.

A corporation’s sustainability or CSR efforts, defined as the way corporations participate in nonfinancial activities to meet social needs (Davis, 1973), are motivated by a desire to get approval from the stakeholders for organizational legitimacy. Stakeholders refer to individuals who are considered important in evaluating organizations, such as suppliers, stockholders, customers, and employees (Ulmer, 2001). They are related closely to corporate behaviors with regard to setting norms, experiencing effects, and evaluating performance.

Previous research has noted that active engagement in, and communication of, CSR can be used as an instrumental method to facilitate effective crisis communications, which includes building a concrete reputation prior to a crisis (H. S. Kim & Lee, 2015; J. Kim et al., 2009), and countering negative publicity after a crisis (Vanhamme & Grobben, 2009). S. Kim and Choi (2018) suggested that when a company launches a CSR initiative congruent with a crisis, consumers demonstrate a favorable attitude about its activity. Past research also indicated that response to public pressure, the need to legitimize business activities, reputation protection, crisis management, and need to increase transparency are all motives companies employ to implement CSR (Luhmann & Theuvsen, 2016).

In addition, an organization’s crisis may influence not only its communication strategies but also its CSR program visions, such as an issue domain, stakeholders targeted, and beneficiaries. Scholars have emphasized the importance of selecting an issue domain and the target stakeholders when performing CSR practices (Du et al., 2010), especially when a company engages in CSR after a crisis (Rim & Ferguson, 2017). Even if the social issues a company supports with CSR tend to focus on idealistic values (Du et al., 2010), they imply that which the company wishes to be acknowledged for and referred to by its stakeholders when evaluating its business (J. Lee & Rim, 2018). Thus, corporations may select social issues that reflect their core values strategically for their CSR practices. Those discussions lead this study to the following research question:

Research Question 3: What social issues do corporations emphasize in sustainability reports before and after crisis, and how do they differ by a crisis type (i.e., CA vs. CSR)?

Communication Network Represented in Sustainability Report

To examine each organization’s sustainability reports corresponding to their crisis, the current study focuses on the communication semantic relations by exploring the semantic networks of sustainability reports with the eigenvector centrality approach. Going beyond merely counting the frequency of words or concepts, semantic network analysis reveals paired relations based on shared meaning (Doerfel, 1998), which enables researchers to examine how words are used and interpreted to form a web of meaning (Carley, 1993; Shumate et al., 2013). Furthermore, eigenvector centrality shows the extent a specific node (i.e., word) is central in a network based on its connectivity with other well-connected nodes in a network (Aktamov & Zhao, 2014). Eigenvector centrality is chosen among other centrality measures because the most dominant themes appeared in text should show higher levels of connectivity among others in the network. Given the advantages of the semantic network analysis, the study analyzes pre- and post-crisis communication networks represented in the organization’s sustainability reports. Scholars have argued that a company that publishes sustainability reports regularly has a greater tendency to scrutinize stakeholders’ expectations and may try to adopt relational methods to improve its social legitimacy (Etzion & Ferraro, 2010; Hahn & Kühnen, 2013). Therefore, as discussed earlier, corporations may specify and prioritize stakeholders in their post-crisis communication efforts, which could be reflected in sustainability reports in their communication networks.

Sustainability reporting is communication targeted to specific stakeholders who have direct connections with a corporation, rather than mass audiences. Those who read sustainability reports are likely to be the active public and stakeholders who are highly engaged with the organization compared to the general public (i.e., mass media audiences) and are willing to participate in communication (J. N. Kim et al., 2010). These active stakeholders are more likely to be interested in corporate governance and their sustainability efforts, especially in the aftermath of a crisis (Y. Kim, 2016). For these reasons, a corporation can stand actively for its stakeholders’ values and relationships through its sustainability reports (Schaltegger & Figge, 2000). In addition, different types of crises may influence the way in which they prioritize stakeholders who may be more or less important in receiving post-crisis responses. Therefore, this study proposes the research question as follows:

Research Question 4: Which stakeholders do corporations prioritize in sustainability reports before and after crises, and how do they differ by a crisis type (i.e., CA vs. CSR)?

Method

Data Collection Scope

This study conducted a multiple case study, which examines a collection of cases having common characteristics (Stake, 2013). Two crisis cases involving Korean Airline companies in the same crisis cluster (i.e., preventable) but representing different types of crisis (i.e., CSR and CA crisis) were selected. The two cases shared many commonalities: the time of crises, industry sectors, headquarter locations, and crisis locations (i.e., occurred in the United States).

Korean Air’s “Nut Rage Return,” which occurred on December 5, 2014, was selected as an exemplar ethical crisis. The vice president of Korean Air violated airline safety laws and insulted a member of the cabin crew because she was unsatisfied with the way Macadamia nuts in first class were served (Myung & Choi, 2017). That employee was asked to kneel down and beg for forgiveness, and the vice president ordered the aircraft to taxi back to the airport’s gate to remove the chief cabin crew at the JFK international airport. In turn, Korean Air received harsh criticism. This crisis demonstrated the way a leader’s unethical behavior influenced a corporation’s relationship with its stakeholders (Fearn-Banks, 2016). Such crisis was different from Asiana Airlines’ crash at the San Francisco International Airport on July 6, 2013, which was caused by pilots’ error. While landing, Asiana Airlines Flight 214 hit the airport’s seawall, spun out and slid down the runway. Then a fire erupted. Among the 307 passengers and crew members on board, the crash killed three girls and more than 180 were injured (Hamilton, 2017).

Both these crises received considerable media coverage and public attention, and each corporation invested a tremendous amount of effort in communicating with both the United States and Korean publics. In addition, scholars consider airline crises a good subject for crisis study (Helm & Tolsdorf, 2013; B. K. Lee, 2005). Haruta and Hallahan (2003) argued that airline crashes generate wide media coverage because the accidents provide highly visible and dramatic images and stories related to death. They also asserted that “a crash or on-the-ground mishap” are important issues for public relations practitioners in the airline industry because those events are related to diverse stakeholders who are waiting for corporate responses (Haruta & Hallahan, 2003).

The scope of the data analysis was set to before and after each crisis’ occurrence. The date on which each crisis occurred served as an anchor point. The selected reports were published before the crisis (t1), immediately after the crisis (t2), and 2 years after the crisis (t3). Since the exact publication dates of each report were unavailable, this study used third-party assurance dates (i.e., the dates when the third-party experts verify the information) for data selection. Considering the assessment is completed when a publication is ready, the assurance date may be used as a proxy of the report’s publication date.

Thus, we selected sustainability annual reports from two Korean airlines that were published before and after their recent crises. Korean Air has published a sustainability report annually since 2006. The reports published in August 2014, 2015, and 2016 were selected for data analysis. The contents of these reports are business philosophy and ethics, governance, crisis management, growth and employment, and environmental and social responsibility. Asiana Airlines has also published sustainability reports annually since 2010, and the reports published in April 2013, July 2014, and September 2015 were selected for data analysis. The contents include sustainability management, financial and risk management, customer service and marketing, and environmental and social responsibility.

Data Preparation

First, sustainability reports were downloaded in .pdf format from each corporation’s website. Words in the reports were preprocessed using R with the “tm” package, which Meyer et al. (2008) developed. The tm package is a popular text mining package that provides corpus handling, data preprocessing, term-document matrix creation, and so on. (Meyer et al., 2008). All words were converted to lowercase and stemmed, and stopwords (i.e., “a” or “the”) and numbers were removed. In preprocessing the data, some transition words (e.g., also) were removed from the frequency analysis. For the final data set, the researchers manually reviewed the postprocessed data and transformed similar words that had the same meanings into the form of a noun. For example, the words “ethical,” “ethic,” and “ethics” are combined to “ethics.” The tm package was used to extract the frequencies of each word and create co-occurrence matrices based on the top 100 words used most frequently in each of the reports. For each set of the top 100 prominent words, term-document matrices with K (the number of words included in the top 100 ranking) * 1 were transformed to K * K co-occurrence adjacency matrices.

Term Frequency and Augmented Normalized Term Frequency (ANTF)

Term frequency (i.e., word frequency) is a simple and intuitive text-mining technique method that counts how many times a word/term repeatedly occurs in a document (Lv & Zhai, 2011). Term-frequency analysis detects to what extent a particular word appears, and in turn, dominates a document (Hou, 2006), which indicates the degree of emphasis of the work in a document. This analysis is often adopted for document component analysis in the business communication context (Benhardus & Kalita, 2013; Ghiassi et al., 2013; Hou, 2006; Liew et al., 2014). For example, Liew et al. (2014) adopted the term-frequency analysis on the sustainability reports to investigate the sustainability trends in the process industry. Other studies conducted the term-frequency analysis with Twitter data, to explore which words were pervasive in brand-related discussions in tweets (Benhardus & Kalita, 2013; Ghiassi et al., 2013).

In the term-frequency analysis, a term (i.e., word) is defined as a lexical item included in a document (Premalatha & Natarajan, 2010). Since term frequency can highly depend on document length (Lertnattee & Theeramunkong, 2004), scholars often perform frequency normalization. Augmented normalized term frequency (ANTF) is one of the popular term frequency normalization techniques (Gadge et al., 2013). Croft (1983) proposed this approach and defined it as 0.5 + 0.5 * tf/max tf, where tf is the term’s occurrence frequency (i.e., word frequency) and tfmax is the maximum term frequency in a document (Lertnattee & Theeramunkong, 2004). This formula assigns a weight of 0.5 for appearance in the document, and extra weight to words that appear more frequently (Gadge et al., 2013). Croft (1983) suggested that the weight must be set at a low level (0.3) for large documents and a high level (0.5) for shorter documents. Each term frequency is divided by the term’s maximum frequency in the text (Croft, 1983), so that each term frequency is guaranteed to lie between 0.5 and 1.0 (Crouch, 1988), which indicates a word’s relative emphasis in a document. By restricting the maximum word frequency to 1.0, this technique compensates for the problem of word frequency dependence on document length (Gadge et al., 2013).

Semantic Network Analysis

A series of semantic network analyses was conducted to explore the different communication patterns in each crisis. A semantic network refers to the way words are semantically related to each other. Thus, semantic network analysis determines the structural differences among texts with regard to word frequency and clusters (Doerfel & Barnett, 1999; S. Lee et al., 2009). Semantic network analysis reveals texts’ semantic organization by examining the relations among words based on their co-occurrence, frequency, and distance (S. Lee et al., 2009). Thus, the map of the relations among words contained in a text can be visualized (Carley, 1993). Semantic network analysis can complement the term-frequency analysis listed above by revealing text structures and meanings (Carley, 1993; Yang & Veil, 2017). For these reasons, semantic network analysis used in this study can shed light on discovering what the dominant words were and how the words are structured and related to each other in creating meanings and identifying subtle patterns in the text (Yang & Veil, 2017).

Particularly in public relations and crisis communication research, semantic network analysis helps demonstrate “. . . mutually associated issues, stakeholders, attributions such as causes, consequences, and solutions” (Schultz et al., 2012, p. 97). Schultz et al. (2012) also used semantic network analysis to study framing strategies for the BP oil spill crisis. The semantic analyses were conducted using UCINET.

To answer RQ2, eigenvector centrality was examined for each word. Eigenvector centrality is a relative measure indicating to what extent a node is central in a network based on its connectivity with other nodes it has connection with (Aktamov & Zhao, 2014; Borgatti et al., 2002). A higher eigenvector centrality score means a node is connected to other well-connected nodes (Mazzola et al., 2016). Therefore, eigenvector centrality indicates the importance, popularity, or prominence of an individual node in a network (Dang et al., 2019). In this study, the eigenvector centrality scores are normalized, divided by the maximum centrality score observed.

Results

Descriptive Statistics

A total of 103,208 words extracted from the six reports was analyzed. For Korean Air (CSA crisis), 14,537 words were drawn from the 2014 report, 21,250 from the 2015 report, and 18,055 from the 2016 report, respectively. For Asiana Airlines (CA crisis), 16,346 words were derived from the 2013 report, 16,954 words were selected from the 2014 report, and 16,066 words were from the 2015 report.

Before applying the term-frequency analysis and semantic network analysis, two preliminary analyses (i.e., document similarity analysis and quadratic assignment procedure [QAP] correlation) were conducted to see whether the sustainability report sets for each airline were meaningfully different from each other to capture the differences between pre-and post-crisis. Similarities among the sustainability reports were assessed by computing cosine similarities between each pair of documents. Cosine similarity is a measure of similarity used to gauge the overlap between documents, which ranges from 0 (no match) to 1 (the greatest match), based on the frequency of word occurrence (Han et al., 2011; Jang et al., 2019). The text mining R package, “stylo,” developed by Eder et al. (2019) was used. The similarity analysis results showed that Korean Air’s sustainability reports were the most similar between post-crisis years (cosine similarity between 2015 and 2016 = 0.89), followed by pre- and post-crisis years (cosine similarity between 2014 and 2015 = 0.88), and pre-crisis and crisis after 2 years (cosine similarity between 2014 and 2016 = 0.85). As for Asiana Airline, the results were different. Asiana Airline’s sustainability reports were the most similar between pre- and post-crisis years (cosine similarity between 2013 and 2014 = 0.91), followed by post-crisis years (cosine similarity between 2014 and 2015 = 0.83), and pre-crisis and crisis after 2 years (cosine similarity between 2013 and 2015 = 0.81). The document similarity analysis results indicated that there were small differences among pre-and post-crisis times for each airline, yet at least, those reports were not identical.

In addition, QAP correlation analysis was conducted to explore the similarities among the semantic networks for each airline (i.e., pre-and post-crisis networks for each airline). QAP correlation analysis is an algorithm-based nonparametric technique that performs the Pearson’s correlation between corresponding cells in two networks (Jiang et al., 2016). Regarding Korean Air, no significant network correlation between the pre- and post-crisis years was observed. However, weak network correlations were observed between pre-crisis and crisis after 2 years (r = 0.18, p < .05), and post-crisis years (r = 0.33, p < .01). For Asiana Airline, there was a moderate semantic networks correlation between pre-and post-years (r = 0.57, p < .001), but weak correlations between pre-crisis and crisis after 2 years (r = 0.21, p < .05) and post-crisis years (r = 0.19, p < .05) were found. The QAP correlation results indicated that, for each airline, the semantic networks among the pre- and post-crisis times did not repeat and showed meaningful differences.

After conducting the preliminary analyses, the top 100 words based on word frequency were selected as the final data set and compared across reports as the analysis sample. For further comparison, each word frequency was normalized with the ANTF approach.

RQ1 asked how a crisis type differently emphasized the crisis-related terms in sustainability reports. Among the top 100 words lists mentioned above, we selected “ethics” as the ethical-harm related word and “safety” as the product-harm related word. The word “safety” is selected as this is the main attribution of Asiana Airlines’ CA crisis. These results showed that the CSR crisis (i.e., Korean Air’s Nut Rage crisis) was not associated with the emphasis of the notion of “ethics” in the reports, but the CA crisis (i.e., Asiana Airlines’ crash crisis) was. There was no noticeable difference for “ethics” observed in Korean Air’s reports before or after the CSR crisis. The word “ethics” appeared 31 times before and 42 and 34 times after the crisis. The difference in the ANTF for “ethics” in Korean Air’s reports was trivial (0.55 before and 0.54 and 0.55 after the crisis).

However, Asiana Airlines’ reports showed noteworthy changes in their use of the word “ethics” with reference to the CA crisis. “Ethics” appeared 49 times before and 23 and 27 times after the crash, and the ANTF confirmed that the frequency of the word “ethics” decreased after the crash. The ANTF for “ethics” before the crash was 0.57, and it decreased to 0.53 and 0.54 after the crash.

Regardless of the crisis type, the emphasis on the notion of “safety” showed increasing trends in the word’s emphasis after the crash. The word frequency of “safety” increased in Korean Air’s reports after the CSR crisis. The word “safety” appeared 89 times before and 121 and 159 times after the crisis. Similarly, the ANTF of “safety” in Korean Air’s reports was 0.65 before and 0.63 and 0.74 after the crisis. With regard to Asiana Airlines’ CA crisis, the word “safety” appeared 91 times in the reports before and 123 and 112 times after the crash. The ANTF of the word “safety” was 0.64 before and 0.65 and 0.72 after the crash.

RQ2 asked how the central keywords in sustainability reports are related to a crisis type. A series of separate semantic network analyses was performed with the top 100 frequently appearing words in each report. Table 1 and 2 summarize the eigenvector centrality of the top 20 words among those. Each word’s eigenvector centrality was examined to find influential words that are connected to other well-connected words in each semantic network, which may play an important role in a corporation’s communication strategies with its stakeholders.

Table 1.

The Top 20 Eigenvector Centrality of the Words Appeared in the Korean Air’s Sustainability Reports.

Korean Air
Pre-crisis (2014)		Post-crisis (2015)		Post-crisis (2016)
Word	Eigenvector centrality	Word	Eigenvector centrality	Word	Eigenvector centrality
Korean	1.000	air	1.000	Korean	1.000
air	0.864	Korean	0.923	air	0.886
management	0.469	manager	0.556	management	0.698
operate	0.455	employee	0.497	safety	0.489
service	0.448	service	0.357	operate	0.440
report	0.441	business	0.337	service	0.434
employee	0.409	environment	0.326	employee	0.428
sustain	0.374	report	0.322	business	0.415
business	0.360	system	0.309	report	0.409
customer	0.325	customer	0.284	customer	0.394
aircraft	0.322	operate	0.274	flight	0.375
excel	0.318	safety	0.263	sustain	0.323
safety	0.308	person	0.260	person	0.311
global	0.294	flight	0.258	program	0.302
flight	0.262	sustain	0.236	system	0.295
market	0.248	aircraft	0.232	aircraft	0.283
new	0.241	include	0.221	issue	0.274
system	0.241	transport	0.217	perform	0.274
internal	0.238	global	0.206	develop	0.255
program	0.234	education	0.201	environment	0.252

Table 2.

The Top 20 Eigenvector Centrality of the Words Appeared in the Asiana Airline’s Sustainability Reports.

Asiana Airline
Pre-crisis (2013)		Post-crisis (2014)		Post-crisis (2015)
Word	Eigenvector centrality	Word	Eigenvector centrality	Word	Eigenvector centrality
Asiana	1.000	Asiana	1.000	Asiana	1.000
airline	0.896	airline	0.903	airline	0.988
management	0.618	report	0.485	management	0.879
service	0.546	flight	0.472	service	0.531
report	0.454	management	0.472	flight	0.527
flight	0.432	service	0.421	operate	0.508
employee	0.404	sustain	0.413	report	0.496
operate	0.391	safety	0.310	system	0.477
company	0.388	operate	0.308	safety	0.438
sustain	0.338	customer	0.256	employee	0.383
safety	0.287	system	0.249	customer	0.355
business	0.265	business	0.241	business	0.332
period	0.262	employee	0.241	environment	0.328
customer	0.256	company	0.238	year	0.316
year	0.256	passenger	0.221	sustain	0.313
passenger	0.243	environment	0.213	financial	0.289
system	0.227	period	0.213	activity	0.281
develop	0.211	total	0.195	company	0.270
total	0.208	year	0.185	period	0.270
corporation	0.196	train	0.179	train	0.266

The results showed that, before the crisis, the most influential nodes were “management,” “operate,” “service,” “employee,” “sustain,” and so on, excluding brand and report-related words (i.e., Korean, air, report). After the crisis, the most influential nodes found were “manager,” “employee,” “service,” “business,” “environment,” and so on, excluding brand and report-related terms. Two years after the crisis, the most influential nodes were “management,” “safety,” “operate,” “service,” and “employee,” excluding brand and report-related terms. However, after 2 years from the crisis, their communication focus moved to performance-related terms (i.e., management, safety, operate, service).

Whereas, before the Asiana crash, the most influential words were “management,” “service,” “flight,” “company,” “employee,” “operate,” “sustain,” and so on, excluding brand and report-related words (i.e., Asiana, airline, report). After the crash, the most influential nodes were “flight,” “management,” “service,” “sustain,” “safety,” “operate,” “customer,” and so on, excluding brand and report-related words. Two years after the crash, the most influential nodes were “management,” “service,” “flight,” “operate,” “system,” and so on, excluding brand and report-related words.

RQ3 asked what social issues corporations emphasized in sustainability reports, and how they differ by a crisis type. To gauge how frequently the social issue words defined in the typology appeared in each report, the study adopted J. Lee and Rim’s social issue types (2018): (1) environment, (2) art/culture/sports, (3) health/disease/safety, (4) education, (5) support for animals, (6) equal rights/diversity/family/women, (7) hunger, (8) homelessness/housing, (9) crime/violence prevention, and (10) others. It appears that Korean Air’s reports did not overtly emphasize any particular social issues after its CSR crisis, while Asiana’s reports put emphasis on safety issues after its CA crisis. Table 3 summarizes the term frequency and ANTF of the social issue words that appeared in each report.

Table 3.

The Term Frequency and Augmented Normalized Term Frequency of the Social Issue Words Appeared in Both Airlines’ Sustainability Reports.

Korean Air
Pre-crisis (2014)			Post-crisis (2015)			Post-crisis (2016)
Word	TF	ANTF	Word	TF	ANTF	Word	TF	ANTF
Environment	93	0.661	Environment	150	0.660	Safety	159	0.739
Safety	89	0.654	Safety	121	0.629	Environment	88	0.632
Emission	57	0.599	Education	94	0.600	Fuel	63	0.595
Education	30	0.552	Emission	46	0.549	Emission	49	0.574
			Health	42	0.545	Culture	43	0.565
						Water	41	0.562
						Education	33	0.550
Asiana Airline
Pre-crisis (2013)			Post-crisis (2014)			Post-crisis (2015)
Word	TF	ANTF	Word	TF	ANTF	Word	TF	ANTF
Safety	91	0.638	Safety	123	0.653	Safety	112	0.719
Environment	84	0.627	Environment	111	0.638	Environment	95	0.686
Green	58	0.588	Health	47	0.558	Education	54	0.605
Health	45	0.568	Green	42	0.552	Fuel	50	0.598
Culture	42	0.564	Education	40	0.550	School	40	0.578
Education	32	0.548	Air	36	0.545	Health	36	0.570
Emission	32	0.548	Culture	32	0.540	Culture	35	0.568
Air	30	0.545	Fuel	31	0.538	Emission	34	0.566
Family	29	0.544

Note. TF = term frequency; ANTF = augmented normalized term frequency.

With regard to the Korean Air’s CSR crisis, “environment,” “safety,” “education,” and “emission,” were the words most frequently observed in the term frequency and ANTF. In particular, “education” increased slightly after the CSR crisis and the numbers dropped again after 2 years in terms of their absolute (i.e., term frequency) and relative emphasis (i.e., ANTF) while “emission” showed a decreased pattern first, and the number increased again after 2 years. The relative emphasis of “environment” remained the same before and after the crisis, but it dropped slightly after 2 years. The relative emphasis of “safety” slightly dropped after the CSR crisis, but it significantly increased after 2 years of crisis.

For the Asiana Airline crash, “safety,” “environment,” “health,” “culture,” and “education,” were the common words that appeared frequently. Interestingly, the emphasis of “safety” gradually increased after the crash, and “education” and “school” were highly emphasized in the 2 years after the crash.

RQ4 asked which stakeholders the corporations prioritize in sustainability reports and how those differ by a crisis type. There appeared to be some noticeable differences after the CSR crisis, but the CA crisis was not related to the way in which relationships with stakeholders were built to any great degree. Table 4 summarizes the term frequency and ANTF of the stakeholder-related words among the top 100 words that appeared most frequently in each report.

Table 4.

The Term Frequency and Augmented Normalized Term Frequency of the Stakeholder-Related Words Appeared in Both Airlines’ Sustainability Reports.

Korean Air
Pre-crisis (2014)			Post-crisis (2015)			Post-crisis (2016)
Word	TF	ANTF	Word	TF	ANTF	Word	TF	ANTF
employee	117	0.700	manager	264	0.780	employee	149	0.724
customer	93	0.660	employee	227	0.740	customer	128	0.692
passenger	67	0.620	customer	132	0.640	person	102	0.653
supplier	59	0.602	person	119	0.630	stakeholder	44	0.566
stakeholder	53	0.590	supplier	73	0.580	passenger	41	0.562
person	32	0.560	stakeholder	72	0.580	director	36	0.554
director	27	0.550	passenger	68	0.570	partner	31	0.547
manager	27	0.550	director	46	0.550
Asiana Airline
Pre-crisis (2013)			Post-crisis (2014)			Post-crisis (2015)
Word	TF	ANTF	Word	TF	ANTF	Word	TF	ANTF
employee	130	0.700	customer	100	0.620	employee	98	0.691
customer	81	0.620	employee	94	0.620	customer	91	0.678
passenger	77	0.620	passenger	86	0.610	passenger	59	0.615
partner	58	0.590	stakeholder	44	0.550	partner	40	0.578
stakeholder	44	0.570	director	42	0.550	stakeholder	39	0.576
director	36	0.550	crew	35	0.540	director	33	0.564
people	35	0.550	partner	34	0.540
			member	33	0.540
			people	33	0.540

Note. TF = term frequency; ANTF = augmented normalized term frequency.

With regard to the Korean Air’s CSR crisis, some noticeable changes in stakeholder-related words were observed. “Employee,” “customer,” “passenger, “stakeholder,” “person,” and “director” were the stakeholder-related words used frequently in both the term frequency and ANTF, before and after the crisis. In particular, there were noteworthy changes in the words that indicated internal stakeholders. “Manager” showed the most remarkable increase in both term frequency and ANTF right after the crisis. Term frequency of the word “employee” showed an increasing tendency of emphasis. Before and after Asiana Airlines’ product-harm crisis, there were minimal differences in the appearance of stakeholder-related words. “Employee,” “customer,” “passenger,” “stakeholder,” “partner,” and “director” were those used most frequently both pre- and post-crash. Interestingly, Asiana Airlines also tended to mention the word “crew” in its reports after the crash, but the tendency was not observed 2 years after the crash.

Discussion

Corporations’ sustainability reports are one of the primary communication channels with their close stakeholders, where they deliver messages about corporate missions, goals, values, and commitment to the stakeholders and communities (Du et al., 2010; Perrini & Tencati, 2006). Accordingly, this study analyzed the sustainability reports to explore how CA and CSR crises are related to their stakeholder communications in terms of the use of words representing the attributes of crises, central keywords, social issue supports, and prioritized stakeholders. Applying Sohn and Lariscy’s (2014) categorization of a reputational crisis, the results of this study provide empirical evidence for how corporations employed different communication efforts to protect and recover their tarnished reputations after CSR and CA crises.

Our RQ1 examined how a company’s CSR and CA crisis differently emphasized the crisis-related terms in sustainability reports. The word frequency analysis focusing on “ethics” and “safety,” which are selected as representing words of the CSR and CA crisis, respectively, revealed that after Korean Air’s CSR crisis, there were not many differences in the emphasis of the word “ethics,” while there was an increasing trend in the emphasis of the word “safety.” Asiana Airlines’ reports showed a gradual increase in “safety,” but there was a noticeable decrease in the emphasis on the word “ethics” after the crash in 2 years.

Given that stakeholders take the crisis dimensions as a cue and that may damage a particular aspect of corporate reputation, communication efforts should be tailored based on the attributes of a crisis event (e.g., Coombs, 2007; Sohn & Lariscy, 2014). However, our results showed that, regardless of crisis types, both companies’ communication efforts tend to focus on emphasizing safety issues. As a crisis event in the same industry sparks not only the consumer public’s attention but also the media’s attention, it often considers as an industrial crisis. Our findings perhaps reflect the industry’s collective efforts in reducing the public’s fear on airline safety. It should be noted that after the Korean Air’s CSR crisis, there were no meaningful changes for the use of “ethics” terms found in the company’s sustainability reports. When corporations are suspected of unethical behaviors, they are pressured to make appropriate responses to mitigate negative public opinions and defend their image by choosing optimal communication responses (Coombs, 2007). Communication professionals should be mindful that appropriate post-crisis communication efforts are required for the CSR crisis to minimize reputational damage.

With regard to the changes in the sustainability reports’ central keywords (RQ2), Korean Air showed a tendency to emphasize words related highly to its internal stakeholders, such as “manager” or “employee,” but did not show the keywords directly related to ethical crisis. In response to its crisis, Asiana Airlines appeared to place more emphasis on words related to business, such as “flight” or “service.” Given that Korean Air’s CSR crisis was related closely to the leader’s unethical behavior and the employee–employer relationship, it may imply the company’s attempts to communicate with its stakeholders by emphasizing the words indicating internal relationships (e.g., manager, employee) right after the crisis. However, when it comes to the sustainability reports published 2 years after the crisis, Korean Air did not maintain the changes it made right after its crisis. In contrast, Asiana Airlines’ crisis was related strongly to its service and management and showed a propensity to emphasize more performance-related terms (e.g., flight, service, safety) to communicate with its stakeholders right after the crisis. The emphasis of communication appears to remain the same after 2 years from the crisis considering the influential nodes with performance-related terms.

The different emphasis on words by crisis type might indicate that a corporation that experienced a CSR crisis might not want to explicitly refer to the words associated closely with the crisis in communicating with its stakeholders. Similarly, the results of Korean Air’s sustainability reports also show no meaningful change in mentioning the term ethics. The 2015 Korean Air’s sustainability report was the only one that explicitly mentioned its corresponding crisis (i.e., “We deeply regret that the incident involving Korean Air stirred up a huge social controversy and caused a great deal of distress to our stakeholders. However, I would like to assure you that we have learned our lessons and that we remain determined to take this opportunity to transform the way we do our business, rather than merely viewing it as a case of crisis management, so that we can regain your trust and confidence”). Interestingly, when a corporation which experienced a CA crisis might use explicit and direct communication in relation to their own crisis. The explicit communication might be because organizations are likely to consider the crisis related to product safety as the critical factor affecting their corporate reputation (Burson-Marsteller & Penn Schoen Berland, 2011) so they might want to directly address it.

Our RQ3 examined whether there are any changes in social issues the company supports before and after crises and how those differ by the crisis type. The results also showed that there were no changes in Korean Air’s reports before and after its CSR crisis. It is possible that Korean Air, which experienced an ethical-harm crisis, did not tweak or integrate their CSR efforts in crisis communication actively, although scholars have discussed the important role of CSR in crisis management (e.g., H. K. Kim & Yang, 2009). However, Asiana Airlines consistently emphasized safety and education issues after its CA crisis. The results appeared to show that those social issue words seemed to have some relationship to their CA crisis.

With regard to stakeholder relationships (RQ4), the occurrence of the ethical crisis was found to influence the approach to building relationships with internal stakeholders and showed a noteworthy leap in the frequency of the word “manager,” as well as a slight increase in “employee” in 2 years after crisis. In contrast to Korean Air’s crisis, the communication strategy with regard to Asiana Airlines’ stakeholders remained largely unchanged, except for the frequent mention of “crew” right after its product-harm crisis. Based on previous research, we expected that corporations would make situationally sensitive stakeholders a priority group in their communication activities (Lin et al., 2011). Our results supported this argument; Korean Air tended to mention words related to those who might be considered the victims of its CSR crisis (i.e., “manager” or “employee”). On the other hand, in responding to the CA crisis, Asiana Airlines’ reports mentioned “crew” who were responsible for the CA crisis more frequently only right after the crisis.

The study provides both theoretical and practical implications. Comparing a corporation’s sustainability reports written before and after crises, this study attempted to investigate corporate crisis communication strategies in the long run. According to the situational crisis communication theory, corporations can protect and rebuild their reputations by selecting optimal response strategies based on crisis types, attributions of crisis responsibility, and initial reputation damage (Coombs, 2007). The optimal strategies should be specified into short-term (primary crisis response) and long-term (secondary crisis response) strategies. Along with the line of research, this study suggested empirical evidence on how corporations take long-term crisis response strategies in the real-world situation, which reinforces the body of crisis communication management literature.

Our results showed a more explicit communication style after the CA crisis. Perhaps, Korean Air’s response strategy to its CSR crisis can be considered as ingratiation, a deal response option, which indicates that the organization reminded stakeholders of its past commitment to society (Coombs, 2007) rather than mentioning the crisis explicitly. However, communication professionals should be mindful that if a corporation does not suggest any communication response to its unethical behavior, people develop a negative image toward the corporation (Bradford & Garrett, 1995). An organization’s unethical image can threaten not only its reputation but also its economic performance (Lindenmeier et al., 2012). Similarly, in their experiment research, Shon and Lariscy (2014) found that people tend to be more sensitive to moral-hazard cue than an ability-defection cue in evaluating attitude to and trust in a company, which implies the importance of managing ethical-harm or CSR-related crises.

Using computational analysis approaches (i.e., term frequency and semantic network analysis), which allowed the researchers to explore the real-world crisis communication cases without human coding biases, the findings of this study revealed the corporations’ crisis communication efforts that gradually changed over pre- and post-crisis years. In addition, this study showed the viability of corporate sustainability annual reports in examining the crisis response strategies that gradually change over the years. This study posited that sustainability reports reflect the organization’s deliberative communication strategy because corporations publish sustainability reports on an annual basis with well-controlled and managed content (S. Y. Lee, 2016). As findings showed, this study confirmed corporations adopt communication strategies differently depending on their crisis type (i.e., CA and CSR crisis) and when they publish each report. Therefore, as an initial attempt to investigate corporate crisis strategies using sustainability reports, this study not only provided the empirical evidence for what values corporations prioritize after crises but also confirmed that corporations strategically use sustainability reports in crisis communication.

This study has limitations that call for readers’ caution when interpreting the findings. First, this study focused on discrete words as a unit of analysis, which did not allow the researchers to explore the sentence structures of the words in a continuous context or the conceptual frames emerged from the word groups. Future research is strongly encouraged to explore the way different types of crises influence the contiguous sequence of each word included in a report (e.g., n-gram analysis or co-occurrence of terms), or explore what concepts/frames emerge depending on crisis types by grouping the words in a meaningful way that goes above the word level. Second, the scope of data collection for this study was limited to recent crises in each airline. Furthermore, as mentioned in the method section, there was high similarity among the sustainability reports, which might have been drawn from the genre conventions of sustainability in wording and structure. To better understand how corporations reform or shape the governance aftermath of a crisis, further research should consider examining the research questions with a broader scope of data to explore trends in the long term. Last, although sustainability reporting provides evidence of the way corporations’ communication approaches and CSR programs change over time after a crisis, the readers are limited to active publics. In addition, the contents and formats can be generic due to the institutional pressure. For more robust comparisons and generalizations, we urge future scholars to extend the scope of reports and investigate other types of corporate publications (e.g., press releases or internal newsletters). Exploring such research questions in other industries or in other types of crises related to ethical issues may advance the crisis communication literature further.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Keonyoung Park

Hyejoon Rim

Author Biographies

Keonyoung Park is a doctoral candidate in the S. I. Newhouse School of Public Communication at Syracuse University. Her research areas include corporate social advocacy, social media activism, public segmentation, and ethics.

Hyejin Kim is as an assistant professor in the College of Communication, at DePaul University. Her research interests include social media advertising and electronic word of mouth (eWOM) with computational research approaches. Her work has appeared in the Journal of International Advertising, Journal of Advertising Research, and Journal of Current Issues and Research in Advertising, among others.

Hyejoon Rim is an associate professor in the Hubbard School of Journalism and Mass Communication at the University of Minnesota. Her research focuses on corporate social responsibility, consumer skepticism, and cross-cultural perspectives. She has published her research in Journal of Business Research, Journal of Communication, Journal of Public Relations Research, and Public Relations Review among others.

References

Aktamov

Zhao

(2014). Impact of network centrality positions on innovation performance of the firm: Evidence from China automobile industry. Business Management and Strategy, 5(1), 164-178. https://doi.org/10.5296/bms.v5i1.5183

Benhardus

Kalita

(2013). Streaming trend detection in Twitter. International Journal of Web Based Communities, 9(1), 122-139. https://doi.org/10.1504/ijwbc.2013.051298

Borgatti

S. P.

Everett

M. G.

Freeman

L. C.

(2002). UCINET for windows: Software for social network analysis. Analytic Technologies.

Borgatti

S. P.

Everett

M. G.

Freeman

L. C.

(2014). Ucinet. In Alhaji

Rokne

(Eds.), Encyclopedia of social network analysis and mining (pp. 2261-2267). Springer.

Bradford

J. L.

Garrett

D. E.

(1995). The effectiveness of corporate communicative responses to accusations of unethical behavior. Journal of Business Ethics, 14(11), 875-892. https://doi.org/10.1007/bf00882067

Brown

T. J.

Dacin

P. A.

(1997). The company and the product: Corporate associations and consumer product responses. Journal of Marketing, 61(1), 68-84. https://doi.org/10.2307/1252190

Burson-Marsteller & Penn Schoen Berland. (2011). 2011 crisis reputation preparedness study. https://www.ropella.com/talon/pdfs/B-M_and_P_S_B_Crisis_Preparedness_Study_2011.pdf

Carley

(1993). Coding choices for textual analysis: A comparison of content analysis and map analysis. Sociological Methodology, 23, 75-126. https://doi.org/10.2307/271007

Coombs

W. T.

(2007). Protecting organization reputations during a crisis: The development and application of situational crisis communication theory. Corporate Reputation Review, 10(3), 163-177. https://doi.org/10.1057/palgrave.crr.1550049

10.

Coombs

W. T.

(2015). The value of communication during a crisis: Insights from strategic communication research. Business Horizons, 58(2), 141-148. https://doi.org/10.1016/j.bushor.2014.10.003

11.

Coombs

W. T.

Holladay

S. J.

(2002). Helping crisis managers protect reputational assets: Initial tests of the situational crisis communication theory. Management Communication Quarterly, 16(2), 165-186. https://doi.org/10.1177/089331802237233

12.

Croft

W. B.

(1983). Experiments with representation in a document-retrieval system. Information Technology-Research Development Applications, 2(1), 1-21.

13.

Crouch

C. J.

(1988, May). A cluster-based approach to thesaurus construction. In Proceedings of the 11th annual international ACM SIGIR conference on Research and development in information retrieval (pp. 309-320). ACM.

14.

Dang

M. Y.

Zhang

G. Y.

Amer

(2019). Social networks among students, peer TAs, and instructors and their impacts on student learning in the blended environment: A model development and testing. Communications of the Association for Information Systems, 44(1), 764-782. https://doi.org/10.17705/1CAIS.04436

15.

Davis

(1973). The case for and against business assumption of social responsibilities. Academy of Management Journal, 16(2), 312-22.

16.

Dawar

Pillutla

M. M.

(2000). Impact of product-harm crises on brand equity: The moderating role of consumer expectations. Journal of Marketing Research, 37(2), 215-226. https://doi.org/10.1509/jmkr.37.2.215.18729

17.

Doerfel

M. L.

(1998). What constitutes semantic network analysis? A comparison of research and methodologies. Connections, 21(2), 16-26.

18.

Doerfel

M. L.

Barnett

G. A.

(1999). A semantic network analysis of the International Communication Association. Human Communication Research, 25(4), 589-603. https://doi.org/10.1111/j.1468-2958.1999.tb00463.x

19.

Bhattacharya

C. B.

Sen

(2010). Maximizing business returns to corporate social responsibility (CSR): The role of CSR communication. International Journal of Management Reviews, 12(1), 8-19. https://doi.org/10.1111/j.1468-2370.2009.00276.x

20.

Eder

Rybicki

Kestemont

Pielstroem

(2019). Package ‘stylo’. https://cran.r-project.org/web/packages/stylo/stylo.pdf

21.

Etzion

Ferraro

(2010). The role of analogy in the institutionalization of sustainability reporting. Organization Science, 21(5), 1092-1107. https://doi.org/10.1287/orsc.1090.0494

22.

Fearn-Banks

(2016). Crisis communications: A casebook approach. Routledge.

23.

Gadge

Sane

Kekre

H. B.

(2013). N-layer approach to web information retrieval. International Journal of Applied Information Systems (IJAIS), 5, 45-49. https://doi.org/10.5120/ijais12-450840

24.

Ghiassi

Skinner

Zimbra

(2013). Twitter brand sentiment analysis: A hybrid system using n-gram analysis and dynamic artificial neural network. Expert Systems with Applications, 40(16), 6266-6282. https://doi.org/10.1016/j.eswa.2013.05.057

25.

Hahn

Kühnen

(2013). Determinants of sustainability reporting: A review of results, trends, theory, and opportunities in an expanding field of research. Journal of Cleaner Production, 59, 5-21. https://doi.org/10.1016/j.jclepro.2013.07.005

26.

Hamilton

(2017, July 6). New video shows dramatic Asiana Airlines crash in San Francisco and frantic rescue effort. Los Angeles Times. https://www.latimes.com/local/lanow/la-me-ln-asiana-crash-video-20170706-story.html

27.

Han

Pei

Kamber

(2011). Data mining: Concepts and techniques. Elsevier.

28.

Haruta

Hallahan

(2003). Cultural issues in airline crisis communications: A Japan-US comparative study. Asian Journal of Communication, 13(2), 122-150. https://doi.org/10.1080/01292980309364841

29.

Helm

Tolsdorf

(2013). How does corporate reputation affect customer loyalty in a corporate crisis? Journal of Contingencies and Crisis Management, 21(3), 144-152. https://doi.org/10.1111/1468-5973.12020

30.

Hou

J. L.

(2006). A knowledge component analysis model based on term frequency and correlation analysis. Journal of Computer Information Systems, 46(4), 64-77. https://doi.org/10.1080/08874417.2006.11645915

31.

Jaques

(2007). Issue management and crisis management: An integrated, non-linear, relational construct. Public Relations Review, 33(2), 147-157. https://doi.org/10.1016/j.pubrev.2007.02.001

32.

Jiang

Barnett

G. A.

Taylor

L. D.

(2016). Dynamics of culture frames in international news coverage: A semantic network analysis. International Journal of Communication, 10, 3710-3736.

33.

Kim

H. K.

Yang

S. U.

(2009). Cognitive processing of crisis communication: Effects of CSR and crisis response strategies on stakeholder perceptions of a racial crisis dynamics. Public Relations Journal, 3(1), 1-39.

34.

Kim

H. S.

Lee

S. Y.

(2015). Testing the buffering and boomerang effects of CSR practices on consumers’ perception of a corporation during a crisis. Corporate Reputation Review, 18(4), 277-293. https://doi.org/10.1057/crr.2015.18

35.

Kim

J. N.

Grunig

J. E.

(2010). Reconceptualizing the communicative action of publics: Acquisition, selection, and transmission of information in problematic situations. International Journal of Strategic Communication, 4(2), 126-154. https://doi.org/10.1080/15531181003701913

36.

Kim

H. J.

Cameron

G. T.

(2009). Making nice may not matter: The interplay of crisis type, response type and crisis issue on perceived organizational responsibility. Public Relations Review, 35(1), 86-88. https://doi.org/10.1016/j.pubrev.2008.09.013

37.

Kim

P. H.

Ferrin

D. L.

Cooper

C. D.

Dirks

K. T.

(2004). Removing the shadow of suspicion: The effects of apology versus denial for repairing competence-versus integrity-based trust violations. Journal of Applied Psychology, 89(1), 104-118. https://doi.org/10.1037/0021-9010.89.1.104

38.

Kim

Choi

S. M.

(2018). Congruence effects in post-crisis CSR communication: The mediating role of attribution of corporate motives. Journal of Business Ethics, 153(2), 447-463. https://doi.org/10.1007/s10551-016-3425-y

39.

Kim

(2016). Understanding publics’ perception and behaviors in crisis communication: Effects of crisis news framing and publics’ acquisition, selection, and transmission of information in crisis situations. Journal of Public Relations Research, 28(1), 35-50. https://doi.org/10.1080/1062726X.2015.1131697

40.

Kolk

(2003). Trends in sustainability reporting by the Fortune Global 250. Business Strategy and the Environment, 12(5), 279-291. https://doi.org/10.1002/bse.370

41.

Lee

B. K.

(2005). Hong Kong consumers’ evaluation in an airline crash: A path model analysis. Journal of Public Relations Research, 17(4), 363-391. https://doi.org/10.1207/s1532754xjprr1704_3

42.

Lee

Rim

(2018). Evolution of corporate social responsibility: A content analysis of United States Magazine Advertising, 1980–2009. Journal of Promotion Management, 24(4), 555-577. https://doi.org/10.1080/10496491.2017.1380111

43.

Lee

S. Y.

(2016). How can companies succeed in forming CSR reputation? Corporate Communications: An International Journal, 21(4), 435-449. https://doi.org/10.1108/ccij-01-2016-0009

44.

Lee

Kim

J. H.

Rosen

(2009). A semantic network and categorical content analysis of Internet and online media research. Open Communication Journal, 3(1), 15-28. https://doi.org/10.2174/1874916x00903010015

45.

Lertnattee

Theeramunkong

(2004). Effect of term distributions on centroid-based text categorization. Information Sciences, 158, 89-115. https://doi.org/10.1016/j.ins.2003.07.007

46.

Liew

Adhitya

Rajagopalan

(2014). Sustainability trends in the process industries: A text mining-based analysis. Computers in Industry, 65(3), 393-400. https://doi.org/10.1016/j.compind.2014.01.004

47.

Lin

C. P.

Chen

S. C.

Chiu

C. K.

Lee

W. Y.

(2011). Understanding purchase intention during product-harm crises: Moderating effects of perceived corporate ability and corporate social responsibility. Journal of Business Ethics, 102(3), 455-471. https://doi.org/10.1007/s10551-011-0824-y

48.

Lindenmeier

Tscheulin

D. K.

Drevs

(2012). The effects of unethical conduct of pharmaceutical companies on consumer behavior: Empirical evidence from Germany. International Journal of Pharmaceutical and Healthcare Marketing, 6(2), 108-123. https://doi.org/10.1108/17506121211243040

49.

Liu

B. F.

Austin

Jin

(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

50.

Luhmann

Theuvsen

(2016). Corporate social responsibility in agribusiness: Literature review and future research directions. Journal of Agricultural and Environmental Ethics, 29(4), 673-696. https://doi.org/10.1007/s10806-016-9620-0

51.

Zhai

(2011, October 24-28). Lower-bounding term frequency normalization. In Proceedings of the 20th ACM international conference on Information and knowledge management (pp. 7-16). ACM. https://doi.org/10.1145/2063576.2063584

52.

Mazzola

Perrone

Kamuriwo

D. S.

(2016). Network positions and the probability of being acquired: An empirical analysis in the biopharmaceutical industry. British Journal of Management, 27(3), 516-533. https://doi.org/10.1111/1467-8551.12174

53.

Meyer

Hornik

Feinerer

(2008). Text mining infrastructure in R. Journal of Statistical Software, 25(5), 1-54. https://doi.org/10.18637/jss.v025.i05

54.

Myung

J. K.

Choi

Y. H.

(2017). The influences of leaders’ dark triad trait on their perception of CSR. Asian Journal of Sustainability and Social Responsibility, 2, 7-21. https://doi.org/10.1186/s41180-017-0013-8

55.

Perrini

Tencati

(2006). Sustainability and stakeholder management: The need for new corporate performance evaluation and reporting systems. Business Strategy and the Environment, 15(5), 296-308. https://doi.org/10.1002/bse.538

56.

Premalatha

Natarajan

A. M.

(2010). A literature review on document clustering. Information Technology Journal, 9(5), 993-1002. https://doi.org/10.3923/itj.2010.993.1002

57.

Rim

Ferguson

M. A. T.

(2017). Proactive versus reactive CSR in a crisis: An impression management perspective. International Journal of Business Communication. Advance online publication. https://doi.org/10.1177/2329488417719835

58.

Schaltegger

Figge

(2000). Environmental shareholder value: economic success with corporate environmental management. Corporate Social-Responsibility and Environmental Management, 7(1), 29-43. https://doi.org/10.1002/(sici)1099-0925(200003)7:1<29::aid-ema119>3.0.co;2-1

59.

Schultz

Kleinnijenhuis

Oegema

Utz

van Atteveldt

(2012). Strategic framing in the BP crisis: A semantic network analysis of associative frames. Public Relations Review, 38(1), 97-107. https://doi.org/10.1016/j.pubrev.2011.08.003

60.

Seeger

M. W.

Ulmer

R. R.

(2001). Virtuous responses to organizational crisis: Aaron Feuerstein and Milt Colt. Journal of Business Ethics, 31(4), 369-376. https://doi.org/10.1023/A:1010759319845

61.

Shen

(2017, December 31) The biggest Business Scandals of 2017. Fortune. http://fortune.com/2017/12/31/biggest-corporate-scandals-misconduct-2017-pr/

62.

Shumate

Pilny

Catouba

Kim

Pena-y-Lillo

Rcooper

Yang

(2013). A taxonomy of communication networks. Annals of the International Communication Association, 37(1), 95-123. https://doi.org/10.1080/23808985.2013.11679147

63.

Sohn

Y. J.

Lariscy

R. W.

(2014). Understanding reputational crisis: Definition, properties, and consequences. Journal of Public Relations Research, 26(1), 23-43. https://doi.org/10.1080/1062726x.2013.795865

64.

Stake

R. E.

(2013). Multiple case study analysis. Guilford Press.

65.

Ulmer

R. R.

(2001). Effective crisis management through established stakeholder relationships: Malden Mills as a case study. Management Communication Quarterly, 14(4), 590-615. https://doi.org/10.1177/0893318901144003

66.

Vanhamme

Grobben

(2009). “Too good to be true!”. The effectiveness of CSR history in countering negative publicity. Journal of Business Ethics, 85(2), 273-283. https://doi.org/10.1007/s10551-008-9731-2

67.

Verhoeven

Tench

Zerfass

Moreno

Verčič

(2014). Crisis? What crisis? How European professionals handle crises and crisis communication. Public Relations Review, 40(1), 107-109. https://doi.org/10.1016/j.pubrev.2013.10.010

68.

Willis

(2003). The role of the global reporting initiative’s sustainability reporting guidelines in the social screening of investments. Journal of Business Ethics, 43(3), 233-237. https://doi.org/10.1023/A:1022958618391

69.

Yang

Veil

S. R.

(2017). Nationalism versus animal rights: A semantic network analysis of value advocacy in corporate crisis. International Journal of Business Communication, 54(4), 408-430. https://doi.org/10.1177/2329488415572781