A balance theory approach to stakeholder network and apology strategy

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS Public Relations Review xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Public Relations Review

Full Length Article

A balance theory approach to stakeholder network and apology strategy Aimei Yang a,∗ , Josh Bentley b a Annenberg School for Communication and Journalism, University of Southern California, 3502 Watt Way, Los Angeles, CA 90089, United States b Bob Schieffer College of Communication, Texas Christian University, TCU Box 298040, Fort Worth, TX 76129, United States

a r t i c l e

i n f o

Article history: Received 13 June 2016 Received in revised form 4 February 2017 Accepted 10 February 2017 Available online xxx Keywords: Crisis communication Social network perspective Stakeholder Network Management Theory Cognitive Balance Theory Apology

a b s t r a c t Apology is an important area of research in crisis communication. Scholars have largely explored apology from an organization-centric, dyadic approach. We argue that this type of research has made unrealistic assumptions about a much more complex social system and may be challenged by increasingly interconnected social reality. This paper uses Structural Balance Theory and Stakeholder Network Management Theory to develop a model and several testable propositions to guide the way organizations respond to a crisis. © 2017 Elsevier Inc. All rights reserved.

1. Introduction Apologies are symbolic rituals performed by those who have committed offensive acts to show their regret for those acts and earn forgiveness from people who have been offended (Benoit, 2015; Goffman, 1971; Tavuchis, 1991). Public apologies often receive significant attention from the mass media, public, and scholars. Apologies are an especially important topic in crisis communication (Coombs, Frandsen, Holladay, & Johansen, 2010), and researchers have analyzed apologies by many individuals and organizations, including corporations, politicians, celebrities, and governments (Benoit, 2015). Scholars generally agree that determining when and how to apologize is a strategic communication decision with considerable consequences (Benoit, 2015; Bisel & Messersmith, 2012; Coombs, 2007). In general, genuine apologies must contain acknowledgements of responsibility and expressions of remorse (Benoit, 2015; Lazare, 2004). Some argue that apologies should include other elements too, such as promises not to repeat the offense or offers of reparations (Bisel & Messersmith, 2012; Hearit, 2006). Researchers have observed that many public statements of remorse are actually pseudo-apologies or non-apologies (Gruber, 2011; Kampf, 2009; Lazare, 2004). Such statements may use words like “I’m sorry,” but they also try to minimize offenses or avoid taking responsibility for them (Lazare, 2004; Kampf, 2009). There are even some situations where pseudo-apologies are more effective than genuine apologies (Bentley, 2015; Eisinger, 2011). Indeed, simply offering a genuine apology when someone expresses anger is not necessarily the most effective crisis response. For instance, in some situations full acknowledgements of responsibility create liability problems (Hearit, 2006);

∗ Corresponding author. E-mail addresses: [email protected] (A. Yang), [email protected] (J. Bentley). http://dx.doi.org/10.1016/j.pubrev.2017.02.012 0363-8111/© 2017 Elsevier Inc. All rights reserved.

Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

2

in other cases, the organization may need to consider the conflicting interests of different stakeholders, and an apology may please one group but offend another stakeholders (Bentley, 2015). In this article we advocate for a move away from a sender-centered, dyadic understanding of apology, which only considers the relationship between a pair of actors at a time. As noted by Rowley (1997), organizations do not merely respond to one stakeholder group at a time. Rather, organizations are subject to stakeholder influence within the networks comprised of organizations and multiple stakeholders. Therefore, under many situations, apology decisions need to consider competing and even conflicting stakeholder interests. Coombs and Holladay (2010) also note that crisis communication as a field needs to move away from a sender perspective to a receiver/stakeholder perspective. Further, the digitized communication context challenges organizations and scholars to embrace new theoretical models to guide their crisis response, as the prevalence of social media often requires organizations to formulate appropriate responses quickly within the networked context of social media, which is complex and sometimes difficult to navigate. This article draws upon Structural Balance Theory and Stakeholder Network Management Theory (Rowley, 1997) to suggest how organizations can and do select the most effective apology response to a crisis. We argue that understanding the interconnected relationships between organizations and their various stakeholder groups enables crisis managers to determine whether it is more strategic to issue a genuine apology, a pseudo-apology, or no apology at all. We also propose a theoretical model to describe the most strategic crisis responses in complex social systems. We illustrate our propositions using examples from previous studies and actual apologies. Moreover, we maintain that a stakeholder network management approach is made possible by the increasing convenience and availability of social media data. These data make it easier to analyze connections and improve our understanding of how networks mediate organization-stakeholder relationships and use such data to aide organizations’ decision making. Our model and testable propositions contribute to theory building and practice in the field of crisis communication in the age of the networked society. One additional observation is worth making here. Offensive actions and public apologies have important ethical dimensions that deserve careful consideration. While we argue that genuine apologies are not always strategic or necessary, we reject any attempt to deceive or take advantage of stakeholders. Further research on the ethics of apologies and pseudoapologies is warranted, but it lies beyond the scope of this article. 2. Apology and pseudo-apologies In times of crisis, communication is a key element of stakeholder relationship management (Ulmer, 2001). Public apologies are an important form of crisis communication (Coombs et al., 2010; Ihlen, 2010). Apologies are rhetorical acts aimed at repairing one’s public image (Benoit, 2015) and relationships (Tavuchis, 1991) after one has violated social norms. This article focuses on organizational apologies as a form of stakeholder relationship management. According to Benoit (2015), there are many different strategies for responding to accusations of wrongdoing. Most of these strategies are defensive, seeking to deny or minimize one’s guilt. Apologies are unique because they involve accepting guilt and showing remorse for offensive acts, a process Benoit (2015) calls mortification. As such, apologies are a particularly complex rhetorical strategy (Benoit, 2015; Coombs, 2007). Goffman (1971) described an apology as “a gesture through which an individual splits himself [sic] into two parts, the part that is guilty of an offense and the part that dissociates itself from the delict and affirms a belief in the offended rule” (p. 113). In essence, apologies seek to demonstrate that offenses are not representative of an offender’s true nature (Schlenker & Darby, 1981). Situational Crisis Communication Theory (Coombs, 2007) suggests that apologies are an appropriate response when an organization is clearly responsible for a crisis. However, because genuine apologies may invite lawsuits, and because they are normally accompanied by reparations or corrective action, genuine apologies are often the most costly type of crisis response. Thus, when organizations are not responsible, or only partly responsible for a crisis, it is usually more appropriate for them to defend themselves. We believe organizations have an ethical duty to apologize for harms they know they have caused. As Coombs and Holladay (2008) have stated, “It is unethical to evade responsibility when it is known” (p. 256). However, Situational Crisis Communication Theory is primarily based on stakeholder perceptions of the situation, not the organization’s perception. No matter how convinced leaders of an organization are that they are not to blame, if key stakeholders insist on blaming the organization, an apology may be necessary. Sometimes activist groups who oppose an organization’s mission are quick to demand an apology for any crisis, even if the organization is not the cause of that crisis. If organizations cannot defend themselves successfully, they may need to apologize whether they agree with their critics or not. Organizations often respond to public pressure for apologies with pseudo-apologies (Gruber, 2011). According to Lazare (2004), pseudo-apologies may involve: 1) offering a vague and incomplete acknowledgement; 2) using the passive voice; 3) making the offense conditional; 4) questioning whether the victim was damaged; 5) minimizing the offense; 6) using the empathic “I’m sorry”; 7) apologizing to the wrong party; or 8) apologizing for the wrong offense (p. 86). Offenders may issue pseudo-apologies because they do not really feel responsible for a situation. They may also issue pseudo-apologies because they do not want to incur liability (Hearit, 2006) or because genuine apologies are too humiliating (Tavris & Aronson, 2007). Bentley (2012) and Kampf (2009) have suggested that pseudo-apologies can be used strategically in political contexts to avoid showing weakness toward one’s political opponents. Indeed, pseudo-apologies can be effective in certain situations. Often, the goal of an apology is to bring closure to an unfortunate episode and convince the news media to move on to other stories (Hearit, 1994). Pseudo-apologies may accomPlease cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

3

plish these goals. However, most stakeholders will only accept an apology if they perceive it to be sincere (Benoit, 2015). When a situation justifies a sincere apology, a failure to express genuine regret may further antagonize key stakeholders and escalate or prolong the crisis (Coombs, 2007). Thus, choosing which type of apology to offer can be complicated. Bentley (2015) has argued that apologizing is essentially a process of shifting identification between offenders, victims, and offenses. Drawing upon Heider’s (1946) Balance Theory and Newcomb’s (1953) co-orientation approach, Bentley explained that offenders are automatically identified with their offenses, and victims naturally seek dissociation from these offenses. Because offenders and victims have different attitudes toward the offense, they cannot normally identify with each other. However, when offenders apologize they shift their identification away from their offense and toward their victim, agreeing with the victim that the offense was wrong. This process of shifting identification through apologizing is fairly straightforward in the context of dyadic relationships. Offenders simply need to persuade victims that they are sincerely apologetic. However, Bentley (2015) noted that many public apologies involve third parties who are watching events play out and forming judgments about the offender. Sometimes the judgments of these third parties are more important to the offender than the judgments of the victim. If the third parties and the victims are not identified with each other, the offender may have to choose between identifying with the victim and identifying with the third parties. As noted by Yang and Taylor (2015), organizations often interact with a network of stakeholders rather than one stakeholder group at a time. How does this network of stakeholders influence organizations to offer apologies or pseudo-apologies? To understand this question, it is necessary to consider stakeholder network management theory. 3. Stakeholder management and stakeholder network theory 3.1. Stakeholders and stakeholder management Freeman (1984) defines a stakeholder as “any group or individual who can affect or is affected by the achievement of the firm’s objectives” (p. 53). For instance, investors, employees, customers, suppliers, activist groups, and governmental agencies can all be considered stakeholders for an organization (Friedman & Miles, 2002). Freeman further argues that the performance of a corporation largely depends on its capacity to negotiate, balance, and satisfy the demands and interests of various internal and external stakeholders. Stakeholder management research has since evolved into a prominent management theory (Donaldson & Preston, 1995; Mitchell, Agle, & Wood, 1997; Phillips, Freeman, & Wicks, 2003; Wood & Jones, 1995) that influences disciplines such as organizational communication, communication management, and public relations (Babrow, 2001; Heath, 1994; Jones, Watson, Gardner, & Gallois, 2004; Lewis, 2007; Lewis, Hamel, & Richardson, 2001). According to Freeman (1984), a major management goal for corporations is to strategically assess the importance of stakeholders and their demands, and to respond properly to stakeholder expectations while advancing the corporations’ goals and objectives. Nevertheless, given that an organization often faces a large number of internal and external stakeholders, it is impossible for managers to cater to every stakeholder’s demands simultaneously. Instead, managers make decisions based on stakeholder salience, defined as “the degree to which managers give priority to competing stakeholder claims” (Mitchell et al., 1997, p. 869). Therefore, Mitchell et al. (1997) proposed that stakeholders possessing one or more of the following attributes are likely to be viewed as salient: 1) power, referring to the stakeholder’s capacity to influence the organization; 2) legitimacy, referring to the validity of the stakeholder’s relationship with the organization; and 3) urgency, referring to the imminence of the stakeholder’s claim to the focal corporation. Despite its prominence, stakeholder theory is not without criticism. For example, Rowley (1997) noted that the dominant stakeholder literature tends to focus on stakeholder relationships as dyads, considering one pair of relationships at a time. Such an approach makes the unrealistic assumption that organizations can deal with one stakeholder group in isolation, and there are no interactions among different stakeholder groups. The emerging complexity of the organizational environment may eventually render the dyadic approach of studying stakeholder relationships obsolete, as this approach cannot keep pace with the dynamic relationships that constitute society (Castells, 2007). A new approach to understanding stakeholder influence—the network approach—can incorporate multiple relationships and more precisely describe the relationship structures that are needed to advance the stakeholder management scholarship. 3.2. Stakeholder network management theory Rowley (1997) developed a network theory of stakeholder management. This theory introduces the social network paradigm to accommodate multiple, interdependent stakeholder interests and predicts organizations’ reactions to competing stakeholder demands. An important development in the network theory of stakeholder influence is that corporations do not respond to one stakeholder group at a time. Rather, corporations are subject to stakeholder influence within the network that consists of corporations and their stakeholders. The interactions and relationship structures among corporations and their stakeholders exert considerable influence over management decisions, such as alliance building strategies, corporate social responsibility reporting, and so forth (Habisch, Patelli, Pedrini, & Schwartz, 2011). According to this theory, corporations are not always the nexus of interactions. Rather, both internal and external social networks among stakeholders may affect organizations’ behaviors. Rowley (1997) further argues that “the focal organization is more than simply the Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

4

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

central point of its own stakeholders: it is also a stakeholder of many other focal points in its relevant social system” (p. 892). This network approach goes beyond dyadic relationships between individual stakeholders and a focal organization to a macro-level perspective that accounts for the interactions among stakeholders. As noted earlier, in the context of crises where organizations need to formulate strategic apology decisions, organizations need to respond in ways that create structural balance between the organizations’ and important stakeholders’ perspectives. Therefore, in addition to the Stakeholder Network Management Theory, another relevant theoretical framework is Balance Theory and its recent network developments. 3.3. Cognitive balance and social network balance process 3.3.1. Balance theory Balance Theory (Heider, 1946, 1958) posits that a person’s attitudes toward other people and objects are interrelated. Heider represented this relationship with his P-O-X model in which P represents the first person, O represents the other person, and X represents some object. If P has a positive opinion of O and X, and O has a positive opinion of X, then P will enjoy cognitive balance. Balance Theory holds that when P and O disagree about X, P will experience an unbalanced cognitive state. The relative importance of X will affect how much cognitive imbalance P experiences. If X has a strongly held religious belief or political view, P may not be able to keep liking O if P and O disagree about X. Newcomb’s (1953) co-orientation approach emphasizes how people use communication to pursue agreement. Newcomb focused on people’s natural desire for symmetry in relationships. This symmetry is achieved through agreement or coorientation. Thus, when two actors in a relationship agree on something, the relationship is symmetrical. However, when they disagree they experience “strain toward symmetry” (1953, p. 395), which motivates them to communicate with one another. One tries either to influence the other or to change one’s own opinion so that both people have the same orientation toward the object. This balance theory perspective is foundational to persuasion research, particularly cognitive dissonance theory (Festinger, 1957). However, it does not take into account the fact that individuals’ opinions about an object are often formed in social networks. A broader perspective is necessary to account for the complexities of modern life. 3.3.2. Structural balance theory and network extension An important development from Heider’s (1958) original theory is Cartwright and Harary’s (1956) application of Balance Theory to networks of more than three nodes using graph theory. In this approach, graphs are used to represent the social structural arrangements that are the outcome of cognitive dissonance. Social structure can be defined as likes and dislikes among a group of people. A graph is considered balanced if all of its nodes can be partitioned into two subsets (also known as plus-sets). The relationship ties in one subset should be all positive and the ties in the other subset all negative. This theoretical and methodological extension is significant (Hummon & Doreian, 2003). Structural balance is no longer considered the outcome of changing minds (Heider, 1958), but rather a process of changes taking place in a group of actors. Further, the representation of social structures as graphs has led to developments in both structural Balance Theory and social network analysis (Norman & Roberts, 1972). Additionally, Cartwright and Harary’s approach was further developed by a line of more recent social network research (Hummon & Doreian, 2003; Kleinnijenhuis & de Nooy, 2013), where new developments have been made on the partition of signed, complex social networks. For the purpose of this paper, we now focus on a partitioning approach to structural balance. 3.3.3. Partitioning approach to structural balance Doreian and Mrvar (1996) developed a network partitioning approach to structural balance that is able to partition a signed network into a structurally balanced pattern, measure the extent to which a network has imbalanced structure, and identify ties that contribute to a lack of balance in social structure. A signed network is one in which each connection is either positive or negative (Leskovec, Huttenlocher, & Kleinberg, 2010). A signed graph is balanced when a network can be partitioned into two subsets, A and B, such that all paths within each subgroup have the same sign. Several insights can be derived from the literature on network structural Balance Theory: 1) structural imbalance is a fundamental tension affecting actors’ decisions and behaviors; 2) characteristics of network members significantly influence the outcome of network change; and 3) Two-mode network (i.e., networks with two sets of nodes and connections such as a network of issues and organizations) is an effective way to study structure balance. Each of these insights is discussed below. First, structural imbalance causes discomfort, tension or strain. This is a driving force that leads actors to take actions to reduce imbalance. In the context of an apology, Bentley (2015) used Balance Theory to explain why offenders and victims are naturally at odds with one another, and how apologies can restore relationships by dissociating an offender from an offense and identifying an offender with a victim. Although Bentley (2015) considered the possibility of a third party influencing this process, he did not explore the implications of network theory for these relationships. In fact, for controversial issues especially, it is natural for different stakeholders to disagree with each other. We suggest that the structure balance network perspective may be especially helpful for understanding the relationships between organizations, their various stakeholder groups, and offensive acts. Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

5

Fig. 1. Organizational Crisis Response Based on Stakeholder Reaction and Salience.

Second, on the network structural Balance Theory, Stokman and Doreian (1997) noted that “the underlying process for network change is assumed to be located in characteristics of network members” (p. 237). In other words, the effect of structure balance takes place in complex networks consist of actors with different attributions. In this process, networked actors’ attributes matter. Indeed, to assume all network actors have equal influence over a focal actor’s decision is unrealistic. In different contexts, it is important to differentiate actors’ characteristics according to certain criteria. For instance, when applying structure Balance Theory to the study of stakeholder relationships, stakeholders’ importance can be differentiated according to Mitchell et al.’s (1997) stakeholder salience indicators: power, legitimacy and urgency. Thus, when actors commit offensive acts, their responses are likely to be influenced by some stakeholders more than others, and these different levels of influence will be determined by the different characteristics of the stakeholders. Third, as explored in many studies (Doreian, 2002), structural imbalance can be caused by conflicting views towards actors or events. As such, not only actors, but social events such as an offensive act, play a significant role in affecting network structure. Indeed, Doreian et al. argued that “network processes are series of events that create, sustain and dissolve social structures” (p. 3). To incorporate this insight, when applying structural Balance Theory to social networks, scholars have utilized the actor-networks as an effective way to capture actors’ positions in regard to issues (Kleinnijenhuis & de Nooy, 2013). In other words, not only can individuals and organizations be part of a network, but social events and people’s attitudes towards those events can also be considered. In an effort to bring together insights from the previously discussed theoretical frameworks to understand strategic apology, the next section presents our model and propositions. 3.3.4. Theoretical model: a stakeholder network approach to apology So far, we have discussed three sets of concepts and theories: apology in crisis communication, stakeholder network management theory, and Balance Theory and its network extensions. In this section, we propose a conceptual model of a network approach to apology that takes into consideration the needs for stakeholder relationship management that goes beyond simplistic dyadic relationships. This model assumes that an organization’s goal in crisis communication is to preserve or repair its relationship and reputation with multiple stakeholders. Organizations do not respond to one stakeholder at a time, but respond to stakeholders connected by social networks. This model is descriptive (See Fig. 1). It describes a decision-making process that considers three sets of variables: stakeholder reaction to an event, stakeholder salience, and structural balance among stakeholders. Together, stakeholder salience and structural balance partition create three scenarios, which further lead to genuine apologies, pseudo-apologies or no apology at all. We believe organizations that appreciate the complexity of their social networks (and are strategically focused on their reputation and stakeholder relationships) follow a process like this one (although not always consciously). In the reminder of this section, we discuss each stage of this decision-making process and offer propositions accordingly. 3.3.5. Stakeholder reactions In times of crisis, stakeholders may have different perceptions and responses. As suggested by SCCT, the nature of a crisis and the nature of stakeholder responses affect the appropriateness of crisis responses (Coombs, 2007; Coombs & Holladay, 2012). Specifically, SCCT posits that crisis responses should vary depending on how much responsibility (i.e., Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

6

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

blame) stakeholders ascribe to the organization. Recognizing the influential role of stakeholder responses, our model starts with stakeholder responses and assumes that stakeholder responses should be the foundation of strategic organizational decision making. In today’s media environment, with abundant and immediately available data, it is possible to assess how people in different stakeholder groups react to a crisis (Coombs & Holladay, 2012). The value of such an assessment is that decision makers can better select and adjust crisis response strategies according to stakeholder reactions (Coombs, 2007). As discussed earlier, not all crisis response strategies involve apology (Coombs, 2007). Some response strategies (e.g., attacking the accuser, denial, excuses, and justification) are more defensive in nature. In contrast, because apologies generally involve conciliation, mortification, and the acceptance of responsibility, they are accommodative in nature (Hargie, Stapleton, & Tourish, 2010). SCCT holds that accommodative responses like apologies are only appropriate when organizations are responsible for causing a crisis. We propose that when there is a low risk of stakeholders blaming an organization for a crisis, the organization does not need to offer an apology to those stakeholders. One reason stakeholder groups may not blame the organization is because they do not think the organization was responsible for the crisis. Another possible reason is because the crisis has low salience for those groups (i.e., they do not care about it). In either case, the organization’s relationship with those stakeholders will not be strengthened by an apology, and a public apology may draw unwanted attention to this issue. However, if an organization is, in fact, responsible for a crisis, it should consider a pre-emptive apology. Not only does the organization have an ethical obligation to take responsibility in such cases, but it should also realize that as stakeholders learn more about the situation they may blame the organization more. Thus, our first proposition applies to situations in which organizations do not think, or are not sure, if they are truly guilty. Proposition 1. An organizational apology is (only) necessary when stakeholder groups believe, or are likely to believe, that an organization has done wrong. This proposition is quite similar to a hypothesis tested by Coombs and Holladay (1996). They conducted an experiment involving different types of organizational crises. For a one-time accident, they found that stakeholders blamed the organization less than they did for repeated accidents or intentional transgressions. In the case of a one-time accident, an excuse-making strategy produced a more positive image of the organization than a mortification strategy involving compensation for victims. In short, our model does not expect organizations to offer apologies without good reason (see Fig. 1). However, under some circumstances, the nature of a crisis and the reactions of certain stakeholders may call for an apology. The strategic decision about whether to offer a genuine apology, a pseudo-apology, or some type of defensive response is based on consideration of two sets of variables: (1) stakeholder salience and (2) stakeholder network structural balance (i.e., attitude toward the crisis). As discussed previously, stakeholder salience depends on power, legitimacy, and urgency (Mitchell et al., 1997). Organizations will pay more attention to the opinions of stakeholders who are more salient. However, stakeholders’ level salience does not necessarily predict how they will react to a crisis. Stakeholders’ attitudes/reactions towards an organization’s action determine the stakeholder structural balance. Techniques for measuring and representing this structural balance are discussed later in the article. For now, it is enough know that stakeholders will be grouped based on whether they have a negative view or a positive view of the organization’s actions. Stakeholders can be partitioned into two subgroups using Doreian and Mrvar’s structural balance partitioning method (1996). As discussed later in the procedure section, this partition method takes into consideration stakeholders’ attitudes/reactions towards the organization’s action, and stakeholders’ relationship with the focal organization as well as with each other, and makes a network level regrouping. Based on each group’s attitudes towards the event, and its salience level, there can be four different scenarios. 3.3.6. Scenario one: genuine apology This is a straightforward scenario. When all stakeholder groups, regardless of their salience level, generally hold a negative, strong reaction to an event or action, then the best way for an organization to rebuild its reputation and relationship with these stakeholders is to offer a genuine apology, as suggested by the following proposition (see Fig. 1): Proposition 2. When there is consistent agreement among stakeholder groups that an organization has done wrong, the organization should offer a genuine apology. The 2010 explosion of BP’s Deepwater Horizon oil rig in the Gulf of Mexico was a crisis that required a genuine apology. Benoit (2015) has called BP’s response to this situation, “very well designed” (p. 58). BP’s CEO apologized on television and the company ran newspaper ads promising to make things right. BP then followed through by spending millions of dollars to clean up the Gulf Coast. Although it did not explicitly admit to causing the oil spill in its apology, BP did take responsibility for correcting the problem, and it did not try to deny or shift the blame. Given that none of BP’s stakeholders could credibly argue that BP was not to blame, this was the only viable strategy. The BP crisis was a tragedy, but it was also a relatively straightforward situation. Many situations are more complex, with some stakeholders supporting an organization’s actions and other stakeholders being offended by those actions. In such a Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

7

Fig. 2. No Apology (Supportive stakeholders have higher salience than offended stakeholders).

complex situation, careful assessments of stakeholder salience levels as well as stakeholders’ network structural balance become important. 3.3.7. Scenario two: no apology This is the scenario depicted by Fig. 2. First, when the stakeholder subgroup that supports an organization’s action is much more salient to the organization than the stakeholder subgroup offended by the action, the organization will want to defend its actions rather than apologize for them. While this response may further antagonize the offended stakeholders, it will also reinforce the support of those stakeholders who agree with the action. Antagonizing stakeholders with low salience poses little risk to an organization’s goals, but apologizing for an action that is supported by the most salient stakeholders might cost the organization valuable support. We must keep in mind that circumstances or stakeholder actions can change the salience of a particular stakeholder group. For instance, if stakeholders who appear to have low salience start holding protests or airing their grievances in the media, this group could suddenly become more salient. Therefore it is not enough for an organization to evaluate the current salience of a stakeholder group. The organization must also consider the potential salience of that group. Still, if a stakeholder group does indeed have very low salience, then apologizing to that group does not make strategic sense. Proposition 3. When the supportive stakeholder subgroup has high salience for the organization, and the offended stakeholder subgroup has low salience, the organization should use defensive strategies rather than apologies. For example, when People for the Ethical Treatment of Animals (PETA) put up a billboard criticizing the eating of pork next to a Chicago butcher shop, the shop did not apologize for its practices. Instead, the shop posted a response on its Facebook page, explaining to its customers how it viewed the slaughter and eating of animals (Publican Quality Meats, 2014). Because the shop’s customers were a much more salient stakeholder group than PETA, and these customers supported the shop’s position, a defensive strategy was appropriate. To apologize would have forced the butcher shop to abandon its organizational goals and risk offending its customers. 3.3.8. Scenario three: genuine apology This is the scenario depicted by Fig. 3. In this scenario, the supportive stakeholder subgroup has a relatively low level of salience for the organization. In contrast, the offended stakeholder groups have a relatively high level of salience for the organization. In this situation, the organization needs to offer a genuine apology to satisfy its most important stakeholders. Even though doing so may irritate the supportive stakeholder subgroup, this particular subgroup is not salient enough, and the organization should prioritize communication strategies to repair its most important relationships. Proposition 4. When the offended stakeholder subgroup has high salience for the organization, and the supportive stakeholder subgroup has low salience, the organization should use genuine apologies rather than other responses.

Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

8

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

Fig. 3. Genuine Apology (Offended stakeholders have higher salience than supportive stakeholders).

Fig. 4. Pseudo-Apology (Offended and supportive stakeholders are both salient).

To illustrate this proposition, we point to Senator Trent Lott’s apology for comments he made at Senator Strom Thurmond’s 100th birthday party in 2002. Thurmond had run for president in 1948 on a political platform that included support for racial segregation. At the 2002 birthday party, Lott remarked that the United States would have been better off if Thurmond had won that election. Many observers interpreted this remark as implying that Lott, who was from Mississippi, supported segregation. As reported by The New York Times, the governor of Mississippi defended Lott, but other officials, including former Vice President Al Gore, called on Lott to apologize (Stolberg, 2002). Ultimately Lott did apologize more than once. Lott’s choice seems to reflect an effort to please offended stakeholders who were much more salient than supporters. 3.3.9. Scenario four: pseudo-apology This is the scenario depicted by Fig. 4. In this situation, both the supportive and the offended stakeholders have similar levels of salience for the organization or their potential salience levels are unclear but organizations need to respond swiftly. The organization needs to maintain a good relationship and reputation with both subgroups. On the one hand, a genuine apology would please the offended stakeholders, but might alienate the supportive stakeholders. On the other hand, a purely defensive response would make supportive stakeholders happy, but it would further antagonize offended stakeholders. Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

9

Under these circumstances, the organization could use a pseudo-apology to placate the offended stakeholders without completely repudiating the position of the supportive stakeholders. Proposition 5. When the supportive stakeholder subgroup and the offended stakeholder subgroup enjoy a similar level of salience to an organization, the organization should use pseudo-apologies rather than other responses. This was the type of situation conservative talk-show host Rush Limbaugh faced in 2012 after he made sexist comments on his radio program about a Georgetown law student named Sandra Fluke (Bentley, 2012). Limbaugh was criticizing Fluke for her testimony on Capital Hill in support of mandatory insurance coverage for contraceptives. He claimed that she wanted other people to pay for her to have sex, and called her a slut and a prostitute. These comments sparked a public outcry, and some of Limbaugh’s sponsors pulled their advertising. In response, Limbaugh issued a pseudo-apology in which he apologized for stooping to the level of his political opponents by engaging in name-calling. Bentley (2012) suggests this strategy allowed Limbaugh to placate some of his advertisers and media critics without actually agreeing with Fluke or the Democratic Party’s position. To be seen agreeing with the Democrats might have alienated Limbaugh’s conservative fans, since they looked to him as a spokesperson for conservatism. His pseudo-apology was apparently an attempt to put the incident behind him and to avoid making the situation any worse. So far we have detailed the propositions of the model, since the application of the model requires analytic procedures of collecting and analyzing stakeholder relationships and responses, next we offer a section on methodological procedures that organizations and practitioners can take to draw insights. 3.4. Analytic procedures When using the proposed model to analyze stakeholder responses and relationships, practitioners need to partition signed, weighted social networks. This requires three steps: 1) at the outbreak of a crisis, analyze the issue at hand, and identify all issue-relevant stakeholders (Hallahan, 2001), and assign weights to each stakeholder according to stakeholder salience indicators (Mitchell et al., 1997). 2) Construct a signed, two-mode network based on each stakeholder groups’ position on the issue. 3) Perform the partition procedure to identify two subgroups (or sub-networks) of stakeholders so that in one subgroup all stakeholders hold negative views towards the issues while in the other subgroup all stakeholders hold positive views. Each of these analytic steps is further discussed below. 3.4.1. Step I (stakeholder identification and weight-assigning) It is important to note that in times of crisis, not all organizations’ stakeholders would necessarily respond to the crisis. Therefore, in the stakeholder identification process, we recommend an issue-based identification framework, where practitioners identify stakeholders that are influenced or potentially influenced by a crisis (for information on how to identify issue-relevant stakeholders, see Hallahan, 2001). Information about stakeholders can be obtained using social media data-mining, media coverage, and internal records from focal organizations. Further, previous studies have demonstrated methods of collecting organizational stakeholder information using interviews, surveys, and focus groups (Illia, Lurati, & Casalaz, 2013; Rawlins, 2008). Next, practitioners can apply Mitchell et al.’s (1997) framework to assign weights to each group. According to Mitchell et al., in terms of power, stakeholder groups that possess critical resources for an organization are highly powerful (value = 2), others that possess useful but not critical resources are powerful (value = 1), and groups that do not possess any useful resources are relatively powerless (value = 0). In terms of urgency, stakeholders that require immediate action from the organization (e.g., consumers asking for refunds) pose urgency (value = 2), while stakeholders that require organizations to take long-term actions pose less urgency (value = 1); stakeholders that make no requirement on organizational actions possess no urgency (value = 0). Finally, in terms of legitimacy, when a corporation and its stakeholders are in contractual relationships, when stakeholders have a legal or moral claim on the corporation, or when stakeholders have something at risk, these stakeholders and their demand can be considered legitimate (value = 1). Meeting two or more criteria yields a legitimacy value of 2. In contrast, stakeholders that possess none of the above attributions are illegitimate (value = 0). Through analyzing each stakeholder group, practitioners can assign each group a value according to power, urgency and legitimacy, and add all categories to assign a weight to each group. 3.4.2. Step II (Construction of a network) To construct a signed network, we recommend following the procedure used by Kleinnijenhuis and de Nooy (2013). First, the relationship ties in the network are constructed based on stakeholders’ interconnections and their positions towards the crisis-related issues. Second, following Kleinnijenhuis and de Nooy (2013), valence is assigned depending on whether a stakeholder group takes a negative view of an issue or a positive/neutral view. These procedures would then generate a weighted, signed two-mode network. 3.4.3. Step III (Network partitioning) The last step requires network partitioning. This partition procedure is based on years of research and is routinely offered in most network analysis software such as Ucinet (Borgatti, Everett, & Freeman, 2002), Pajeck (Batagelj & Mrvar, 2011), and Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

10

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

the R SNA package. Practitioners or researchers can load their network data into one of these programs and obtain a network that is partitioned into two parts based on stakeholder reactions to the crisis event. 4. Conclusion This article aims to contribute a new theoretical model to guide future research in the field of crisis communication. This theoretical model makes several significant contributions. First, the model advocates for a stakeholder-centric approach to consider and evaluate crisis response strategies. Responding to the call from Coombs and Holladay (2012) for a move away from a sender’s perspective in crisis response, our model bases strategic choices mainly on stakeholder reactions and stakeholder salience, and therefore, theoretically articulates a stakeholder-centric perspective that has been largely missing in the crisis literature. Our model also maintains that organizations often need to deal with multiple and even competing stakeholder interests. At times, satisfying every interest group may not be possible, and it is necessary for organizational communication to prioritize the most salient stakeholders’ demands. Second, against the backdrop of the emerging complexity of a network society (Castells, 2007), the model suggests a shift away from an unrealistic, dyadic approach to studying organizational crisis response that only examines how organizations deal with one stakeholder group at a time. The model recognizes that organizations and their various stakeholders are embedded in interconnected social networks, and their relationships with each other and with the organization matters when it comes to critical decisions such as crisis response. Third, we have offered innovative approaches to data collection and analysis. In this fast-changing era of technological innovation, communication practitioners and scholars need to actively embrace opportunities made available by new software and methods such as social network analysis and the big data approach. Our model builds a theoretical framework that embraces new methodological approaches and provides a fertile ground for future development. Indeed, there are many case-by-case analyses of how organizations should apologize in times of crisis. That approach, however, is limited because it is based on anecdotes rather than theory. Studies often reach contradictory conclusions or recommendations that are not helpful to theory building or make little effort to embrace methodological innovation. We believe our model helps to move the field of crisis communication forward by encouraging critical refinements in crisis communication research and practice, and by identifying empirically testable concepts and propositions. With a continued effort, organizations can be better prepared for challenges and opportunities ahead. Finally, we reiterate the fact that this model does not account for ethical issues in the decision-making procedure. This omission is a function of space limitations, not a signal that ethics are irrelevant. In addition to the strategic considerations described in our model, communication decisions must also be guided by practitioners’ moral compass. References Babrow, A. S. (2001). Uncertainty value, communication, and problematic integration. Journal of Communication, 51, 553–573. Benoit, W. L. (2015). Accounts, excuses, and apologies (2nd ed.). Albany, NY: State University of New York Press. Bentley, J. M. (2012). Not the best: What Rush Limbaugh’s apology to Sandra Fluke reveals about image restoration strategies on commercial radio. Journal of Radio and Audio Media, 19(2), 221–238. http://dx.doi.org/10.1080/19376529.2012.722474 Bentley, J. M. (2015). Shifting identification: A theory of apologies and pseudo-apologies. Public Relations Review, 41, 22–29. Bisel, R. S., & Messersmith, A. S. (2012). Organizational and supervisory apology effectiveness: Apology giving in work settings. Business Communication Quarterly, 75(4), 425–448. Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). Ucinet for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies. Cartwright, D., & Harary, F. (1956). Structural balance: A generalization of Heider’s theory? Psychological Review, 63(5), 277–293. Castells, M. (2007). Communication power and counter-power in the networked society. International Journal of Communication, 1, 238–266. Coombs, W. T., & Holladay, S. J. (1996). Communication and attributions in a crisis: an experimental study in crisis communication. Journal of Public Relations Research, 8(4), 279–295. http://dx.doi.org/10.1207/s1532754xjprr0804 04 Coombs, W. T., & Holladay, S. J. (2008). Comparing apology to equivalent crisis response strategies: Clarifying apology’s role and value in crisis communication. Public Relations Review, 34, 252–257. http://dx.doi.org/10.1016/j.pubrev.208.04.001 Coombs, W. T., & Holladay, S. J. (2012). Amazon.com’s Orwellian nightmare: exploring apology in an online environment. Journal of Communication Management, 16(3), 280–295. Coombs, W. T., Frandsen, F., Holladay, S. J., & Johansen, W. (2010). Why a concern for apologia and crisis communication? Corporate Communications: An International Journal, 15(4), 337–349. http://dx.doi.org/10.1108/13563281011085466 Coombs, W. T. (2007). Ongoing crisis communication: Planning, managing, and responding (2nd ed.). Thousand Oaks, CA: Sage. Donaldson, T., & Preston, L. E. (1995). The stakeholder theory of the corporation: Concepts evidence, and implications. The Academy of Management Review, 20(1), 65–91. Doreian, P., & Mrvar, A. (1996). A partitioning approach to structural balance. Social Networks, 18, 149–168. Doreian, P. (2002). Event sequences as generators of social network evolution. Social Networks, 24, 93–119. Festinger, L. A. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Freeman, I. L. (1984). Strategic management: A stakeholder approach. MA, Pitman: Marshall. Friedman, A. L., & Miles, S. (2002). Developing stakeholder theory. Journal of Management Studies, 39(1), 1–21. Goffman, E. (1971). Relations in public: Microstudies of the public order. New York, NY: Basic Books. Gruber, M. C. (2011). Pseudo-apologies in the news. In E. Yuasa, T. Bagchi, & K. Beals (Eds.), Pragmatics and autolexical grammar (pp. 93–106). Philadelphia, PA: John Benjamins. Habisch, A., Patelli, L., Pedrini, M., & Schwartz, C. (2011). Different talks with different folks: a comparative survey of stakeholder dialog in Germany, Italy, and the US. Journal of Business Ethics, 100(3), 381–404. Hallahan, K. (2001). Inactive publics: the forgotten publics in public relations*. Public Relations Review, 26(4), 499–515. Hargie, O., Stapleton, K., & Tourish, D. (2010). Interpretations of CEO public apologies for the banking crisis: attributions of blame and avoidance of responsibility. Organization, 17(6), 721–742.

Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012

G Model PUBREL-1581; No. of Pages 11

ARTICLE IN PRESS A. Yang, J. Bentley / Public Relations Review xxx (2017) xxx–xxx

11

Hearit, K. M. (1994). Apologies and public relations crises at Chrysler, Toshiba, and Volvo. Public Relations Review, 20(2), 113–125. http://dx.doi.org/10.1016/0363-8111(94)90053-1 Hearit, K. M. (2006). Crisis management by apology: Corporate response to allegations of wrong. Mahwah, NJ: Erlbaum. Heath, R. L. (1994). Management of corporate communication: From interpersonal contacts to external affairs. Hillsdale, NJ: Lawrence Erlbaum. Heider, F. (1946). Attitudes and cognitive organization. Journal of Psychology, 21, 107–112. http://dx.doi.org/10.1080/00223980.1946.9917275 Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hummon, N. P., & Doreian, P. (2003). Some dynamics of social balance processes: Bringing Heider back into balance theory. Social Networks, 25(1), 17–49. Ihlen, Ø. (2010). The cursed sisters: Public relations and rhetoric. In R. L. Heath (Ed.), The SAGE handbook of public relations (pp. 59–70). Thousand Oaks, CA: Sage. Illia, L., Lurati, F., & Casalaz, R. (2013). Situational theory of publics: Exploring a cultural ethnocentric bias. Journal of Public Relations Research, 25(2), 93–122. Jones, E., Watson, B., Gardner, J., & Gallois, C. (2004). Organizational challenges for the new century. Journal of Communication, 54, 722–750. Kampf, Z. (2009). Public (non-) apologies: The discourse of minimizing responsibility. Journal of Pragmatics, 41, 2257–2270. http://dx.doi.org/10.1016/j.pragma.2008.11.007 Kleinnijenhuis, J., & de Nooy, W. (2013). Adjustment of issue positions based on network strategies in an election campaign: A two-mode network autoregression model with cross-nested random effects. Social Networks, 35(2), 168–177. Lazare, A. (2004). On apology. New York, NY: Oxford University Press. Leskovec, J., Huttenlocher, D., & Kleinberg, J. (2010). Signed networks in social media. In Proceedings of the SIGCHI conference on human factors in computing systems (pp. 1361–1370). Lewis, L. K., Hamel, S. A., & Richardson, B. K. (2001). Communicating change to nonprofit stakeholders. Management Communication Quarterly, 15(1), 5–41. Lewis, L. K. (2007). An organizational stakeholder model of change implementation communication. Communication Theory, 17, 176–204. Mitchell, R., Agle, B., & Wood, D. (1997). Toward a theory of stakeholder identification and salience: Defining the principle of who and what really counts. Academy of Management Review, 22, 853–886. Newcomb, T. M. (1953). An approach to the study of communicative acts. Psychological Review, 60(6), 393–404. http://dx.doi.org/10.1037/h0063098 Norman, R. Z., & Roberts, F. S. (1972). A measure of relative balance for social structures. In J. Berger, M. Zelditch, & A. Bo (Eds.), Sociological theories in progress (Vol. 2) (pp. 358–391). New York: Houghton Mifllin. Phillips, R., Freeman, R. E., & Wicks, A. C. (2003). What stakeholder theory is not. Business Ethics Quarterly, 1, 479–502. Publican Quality Meats (2014). We eat meat [Facebook post]. Retrieved from https://www.facebook.com/photo.php?fbid=567260226693142&set=a.501545036597995.1073741827.223802047705630&type=1&theater. Rawlins, B. (2008). Give the emperor a mirror: Toward developing a stakeholder measurement of organizational transparency. Journal of Public Relations Research, 21(1), 71–99. Rowley, T. J. (1997). Moving beyond dyadic ties: A network theory of stakeholder influences. Academy of Management Review, 22, 887–910. Schlenker, B. R., & Darby, B. W. (1981). The use of apologies in social predicaments. Social Psychology Quarterly, 44(3), 271–278. Stokman, F. N., & Doreian, P. (1997). Evolution of social networks: Processes and principles. In P. Doreian, & F. N. Stokman (Eds.), Evolution of social networks (pp. 233–250). Netherlands: Gordon and Breach Science Publishers. Stolberg, S. G. (2002, Decemberr 10). Under fire, Lott apologizes for his comments at Thurmond’s Party. The New York Times. Retrieved from http://www.nytimes.com/2002/12/10/politics/10LOTT.html. Tavris, C., & Aronson, E. (2007). Mistakes were made (but not by me). Orlando, FL: Harcourt. Tavuchis, N. (1991). Mea culpa: A sociology of apology and reconciliation. Stanford, CA: University Press. Ulmer, R. R. (2001). Effective crisis management through established stakeholder relationships Malden Mills as a case study. Management Communication Quarterly, 14(4), 590–615. Wood, D., & Jones, R. (1995). Stakeholder mismatching: A theoretical problem in empirical research on corporate social performance. International Journal of Organizational Analysis, 3, 225–332. Yang, A., & Taylor, M. (2015). Looking Over, looking out: and moving forward: A network ecology framework to position public relations in communication theory. Communication Theory, 25, 91–115.

Please cite this article in press as: Yang, A., & Bentley, J. A balance theory approach to stakeholder network and apology strategy. Public Relations Review (2017), http://dx.doi.org/10.1016/j.pubrev.2017.02.012