Coombs / SITUATIONAL COMMUNICATION JOURNAL OF BUSINESSCRISIS COMMUNICATION THEORY 10.1177/0021943604265607

IMPACT OF PAST CRISES ON CURRENT CRISIS COMMUNICATION
Insights From Situational Crisis Communication Theory
W. Timothy Coombs

Eastern Illinois University

Previous research based on Situational Crisis Communication Theory (SCCT) suggests that an organization’s past crises history affects the reputational threat posed by a current crisis when that crisis results from intentional acts by the organization. The study reported on in this article provides a wider test of crisis history to better assess its role in crisis communication. Results from the present investigation showed that a history of similar crises intensified the reputational threat of a current crisis even when the crisis arose from the victimization of the organization or from an accident, rather than from the organization’s intentional acts. The threat to reputation was primarily direct, rather than indirect, through perceived responsibility for the crisis. There was little difference in the perceptions of organizations identified as having had no history of past similar crises versus those for whom no information about past crises was provided. Perception of an organization’s responsibility was negatively correlated with the perceived impact on reputation. Implications for the practice of crisis communication and further development of SCCT are discussed. Keywords: crisis; communication; management; reputation; theory

On Monday March 27, 2000, a deadly blast ripped through the Phillips Petroleum Company facility in Pasadena, Texas. That day, Phillips’s managers faced not just one crisis, but three. Newspaper reports mentioned a 1989 explosion that killed 23 workers and a 1999 incident that killed two workers and injured four others (Rendon & Watson, 2000). Situational Crisis Communication Theory (SCCT) argues that whether people know about past crises is an important consideration for crisis managers. Indeed, SCCT suggests that to adequately protect an organization’s reputation, management must adjust their communication to account for possible past crises about which relevant publics are aware. SCCT focuses on the use of
W. Timothy Coombs (Ph.D., Purdue University, 1990), is an associate professor in the Department of Speech Communication at Eastern Illinois University. Correspondence concerning this article should be addressed to Dr. W. Timothy Coombs, Speech Communication, Eastern Illinois University, 600 Lincoln Ave., Charleston, IL 61920; e-mail: cfwtc@eiu.edu.
Journal of Business Communication, Volume 41, Number 3, July 2004 265-289 DOI: 10.1177/0021943604265607 © 2004 by the Association for Business Communication

266

JOURNAL OF BUSINESS COMMUNICATION

communication to preserve and protect an organization’s valuable reputation (Barton, 2001; Brown, 2003; Fombrun, 1996; Harris Interactive, 2000; Nakra, 2000). SCCT argues that as the reputational threat increases, the crisis manager should use response strategies that demonstrate acceptance of responsibility for the crisis and that address victim concerns (Coombs, 1995; Coombs & Holladay, 2002). A history of crises is posited to be a critical factor in the determination of the reputational threat (Coombs & Holladay, 2001, 2002). This article presents the results of a study designed to test SCCT’s contention that a history of similar crises intensifies the reputational threat posed by a crisis that results from the victimization of an organization or from an accident, not just from an organization’s intentional misdeeds (Coombs & Holladay, 2001). The first section presents the literature relevant to SCCT, elaborates on the crisis history proposition, and presents the hypotheses. The second section of the article discusses the methods. Results of the study are presented following the discussion of the method. The final two sections of the article present a discussion of the results and examine the investigation’s implications.
LITERATURE ANALYSIS AND HYPOTHESES

The situation facing Phillips Petroleum Company in March 2000 was not unique. News reports about a current crisis often include reminders of past crises. In September 2003, a patron at Disneyland died as a result of an accident on the Big Thunder Mountain roller coaster. The news reports included information about a December 1998 incident that killed a park visitor (e.g., Carter, 2003). What happens when people encounter information about past crises: Is the information background noise or a significant factor considered relevant in the current crisis? SCCT argues that information about past crises is a significant factor that can affect perceptions of a more recent crisis. Specifically, SCCT suggests that the information about past crises can shape perceptions of the current crisis, the reputational threat presented by the current crisis, and, hence, should guide the optimal communication responses for protecting the organizational reputation (Coombs & Holladay, 2002). SCCT evolved from a number of studies that examined how a crisis might shape the selection of crisis response strategies and/or examined the effect of crisis response strategies on organizational reputation (Bradford & Garrett, 1995; Coombs, 1999a; Coombs & Holladay, 1996, 2001; Coombs & Schmidt, 2000; Fediuk, 1999). The idea was to articulate a theory-based system for matching crisis response strategies to the crisis situation to best preserve the organizational reputation. SCCT is designed to follow the use of instructing information. Instructing information tells stakeholders what, if anything, to do to protect themselves physically from a crisis and what the organization is doing to prevent a repeat of the crisis (Bergman, 1994; Sturges, 1994). Attribution theory served as the guide for linking the crisis situations to crisis response strategies (Coombs, 1995, 1998, 1999b).

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

267

Attribution Theory Connection Attribution theory holds that people will make judgments about the causes of events, especially unexpected events with negative outcomes. Attributions are “perceptions of the causality or the perceived reasons for a particular event’s occurrence” (Weiner, 1985b, p. 280). People will attribute the cause of an event to an individual involved in the event (personal causality) or to some outside force (external causality). Attributions indicate if a person believes that the cause of the event was controllable by the people involved (McAuley, Duncan, & Russell, 1992; Weiner, 1985a; Weiner, Perry, & Magnusson, 1988). Causal attributions are important because they affect emotions generated by the event and future interactions with the person involved (Weiner, 1985a). If a person spills red wine on your new carpet and you believe the cause was personal, you are likely to be angry with that person and behave toward him or her differently in the future. Crises are exactly the type of event that will trigger attributions; crises are sudden and negative. It follows that people will make attributions about the cause of the crisis. Was the crisis something the organization could control? Control implies responsibility (Weiner, 1995). If stakeholders believe an organization could control a crisis, they will also hold the organization responsible for the crisis. Marketing studies have applied attribution theory to product recalls, when an organization needs to recall a defective and/or harmful product. The research examined the extent to which people attributed responsibility for the recall to the organization or to the consumer/external factors, and the effect of those attributions on behavioral intentions (e.g., Folkes, 1984; Griffin, Babin, & Attaway, 1991; Griffin, Babin, & Darden, 1992). Crisis attributions do matter as they shape feelings and behaviors toward the organization involved in the crisis (Coombs, 1995). Greater attributions of responsibility lead to stronger feelings of anger and more negative views of people and organizations (Weiner, Amirkhan, Folkes, & Verette, 1987). The messages people develop to explain an event can shape attributions and the feelings created by the attributions (Weiner et al., 1988). Hence, crisis managers can use crisis response strategies in their attempts to shape attributions of the crisis and/or perceptions of the organization itself. People commonly use three causal dimensions when making attributions: stability, external control, and personal control/locus. Stability reflects whether the cause of the event happens frequently (stable) or infrequently (unstable). If someone repeats the same mistake, it is stable; but if the mistake is unique or rare, it is unstable. External control indicates whether the event’s cause was controllable or uncontrollable by some other person. If another actor controls what happens to a person, there is strong external control in the situation. Personal control reflects whether the event’s cause is controllable or uncontrollable by the actor. Personal control addresses the extent to which a person can control his or her own fate in a given situation. Locus reflects the extent to which an event’s cause is located in the actor or in the situation. An internal locus reflects something about the person, whereas external locus reflects on the situation (McAuley et al., 1992; Russell, 1982; Wilson, Cruz, Marshall, & Rao, 1993).

268

JOURNAL OF BUSINESS COMMUNICATION

Research consistently demonstrates a substantial overlap between personal control and locus; thus, it can be argued that they constitute one dimension (Wilson et al., 1993). Both personal control and locus reflect the intentionality of an act. High personal control and a locus in the actor create perceptions of intentional actions by the actor. Yet low personal control and a locus in the situation foster perceptions of unintentional action. These three attribution dimensions help people decide if the cause of an event is the person involved or environmental factors. A person is held more responsible for an event (stronger attributions of personal responsibility) when the event is perceived as stable and when the person has high personal control and/or low external control. Less responsibility is attributed to a person when the event is seen as unstable and when the person has low personal control and/or high external control. Likewise, similar patterns should hold true for attributions of crisis responsibility. Organizational crisis responsibility should be perceived as strongest when the cause is stable (the organization has a history of crises), external control is low (controlled by others outside of the organization), and personal control/locus is internal (the crisis originates from within the organization). When a crisis event is repeated (stable), the public should be more likely to attribute responsibility to the organization. Attributions of low external control indicate that the crisis was not under the control of groups outside of the organization and, thus, the crisis should not be attributed to external agents. Attributions that entail an internal locus/ personal control suggest that the organization could have done something to prevent the crisis. Attribution theory provides the link necessary to match crisis response strategies with a crisis situation. The previously discussed crisis response strategies and crisis situations (e.g., Allen & Caillouet, 1994; Benoit, 1995; Fearn-Banks, 1996: Lerbinger, 1997) were developed independently of one another; thus, there was no apparent logical and consistent connection between them. Indeed, each strategy and approach has its own categorization system, therefore making it challenging to have the situation guide the selection of crisis response strategies. SCCT uses the attributions of responsibility to forge a connection between crisis response strategies and the crisis situation (Coombs, 1995; Coombs & Holladay, 2002, 2004). Crisis Responsibility: The Lynchpin of SCCT Crisis responsibility, the degree to which stakeholders attribute responsibility for a crisis to an organization, is the centerpiece of SCCT. Attributions of crisis responsibility are directly related to the reputational threat posed by a crisis. A crisis becomes a greater threat to an organization’s reputation as attributions of crisis responsibility intensify. The relationship between attributions about crisis responsibility and reputational threat has been documented across a range of crisis types, including product tampering, human-error accidents, organizational misdeeds, and natural disasters (Coombs, 1998, 1999b; Coombs & Holladay, 1996, 2001; Coombs & Schmidt, 2000). As reputational threat increases, crisis management

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

269

teams should utilize strategies that indicate a greater acceptance of responsibility for the crisis and simultaneously demonstrate concern for victims (Bradford & Garrett, 1995; Coombs, 1995, 1998, 1999a, 1999b). However, because accepting responsibility is related to high financial costs for an organization, such responses must be used with caution (Cohen, 1999; Patel & Reinsch, 2003; Tyler, 1997).

SCCT argues that information about past crises is a significant factor that can affect perceptions of a more recent crisis.

The assessment of the crisis threat is a two-step process. In the first step, the crisis team identifies the crisis type. A crisis type is a frame used to guide interpretations of the situation. People can focus on different cues in a crisis. A crisis frame emphasizes which cues should be the focus of consideration when evaluating the situation (Druckman, 2001). SCCT specifies 10 crisis types or frames: natural disaster, rumor, product tampering, workplace violence, challenges, technicalerror product recall, technical-error accident, human-error product recall, humanerror accident, and organizational misdeed. Table 1 provides definitions and key cues for each crisis type. Crisis types vary by how much crisis responsibility stakeholders ascribe to the organization. By understanding how much crisis responsibility a crisis type is likely to generate, a crisis manager can predict the reputational threat posed by the crisis type. Prior research demonstrated that the 10 crisis types will produce some level of crisis responsibility and be a reputational threat. As Weiner (1995) noted, responsibility is not a yes-or-no proposition; instead, it is a matter of degree or magnitude. Moreover, Perrow (1999) found that perceptions of accidents do include some element of organizational fault. Specially, an organization will likely be attributed some level of responsibility simply because it is in a crisis (Coombs & Holladay, 2002). Mitroff (1988) and other crisis management experts recommend collapsing crisis types into clusters, families of similar crises. The contention is that similar crises, those in the same cluster, can be managed in similar ways (Mitroff & Anagnos, 2001; Mitroff, Harrington, & Gai, 1996; Pearson & Mitroff, 1993). Once crises are grouped, crisis management teams can prepare plans for each cluster, instead of generating plans for every possible crisis type an organization might face. Attributions of crisis responsibility have been used to group the various crisis types into three clusters: (a) victim, (b) accidental, and (c) intentional. The victim cluster contains crisis types that produce very low attributions of crisis responsibility (natural disasters, rumors, product tampering, and workplace violence) and represent a mild reputational threat. Organizations are viewed as “victims of the crisis”

270

JOURNAL OF BUSINESS COMMUNICATION

Table 1.

Crisis Types Definitions and Cues

Victim Crisis Cluster Natural disaster: Acts of nature that damage an organization such as an earthquake. Some environmental/weather event impacts the organization. Rumors: False and damaging information about an organization is being circulated. Evidence that the information is false. Workplace violence: Current or former employee attacks current employees onsite. An employee or former employee injures or attempts to injure current employees. Product tampering/malevolence: External agent causes damage to an organization. Some actor outside of the organization has altered the product to make it dangerous. Accidental Crisis Cluster Challenges: Stakeholders claim an organization is operating in an inappropriate manner. There is a public challenge based on moral or ethical, not legal, grounds. Technical error accidents: A technology or equipment failure causes an industrial accident. The cause of the accident is equipment/technology related. Technical error recalls: A technology or equipment failure causes a product to be recalled. A product is deemed harmful to stakeholders. The cause of the recall is equipment or technology related. Intentional Crisis Cluster Human error accidents: Human error causes an industrial accident. The cause of the accident is a person or people not performing job properly. Human error recalls: Human error causes a product to be recalled. A product is deemed harmful to stakeholders. The cause of the recall is a person or people not performing job properly . Organizational misdeed: Laws or regulations are violated by management or stakeholders are placed at risk by management. Members of management knowingly violate laws/regulations or offer a product or service they know could injure stakeholders.

because the crises are seen as driven by external forces that were beyond management’s control (Coombs, Hazleton, Holladay, & Chandler, 1995). The accidental cluster contains crisis types that produce minimal attributions of crisis responsibility (challenges, technical-error accident, and technical-error product recall) and represent a moderate reputational threat. The organization’s management is seen as not meaning for the crisis to happen (lack of volition) and/or could do little to prevent it (limited control). Organizations face some form of risk and sometimes accidents/events happen (Perrow, 1999). Finally, the intentional crisis cluster contains crisis types that produce strong attributions of crisis responsibility (human-error product recalls, human-error accidents, and organizational misdeeds) and represent a severe reputational threat. Organizational misdeeds involve management knowingly violating laws or regulations and/or knowingly placing stakeholders at risk. Violating discrimination laws or allowing a product to go to market with a known defect are examples of intentional acts (Coombs & Holladay, 2001). Human error is perceived as intentional because people believe such mistakes could and should have been preventable (Mitroff & Anagnos, 2001; Morris, Moore, & Sim, 1999; Reason, 1999). In the second step of assessing the crisis threat, the crisis team adjusts the initial reputational threat assessment by considering three intensifiers: (a) crisis history,

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

271

Personal Control

Crisis Responsibility

Organizational Reputation

Performance History -Crisis History -Relationship History
Figure 1. Relationships Between Situational Crisis Communication Theory Variables

whether an organization has had similar crises in the past; (b) relationship history, the organization’s record of good or bad behavior toward stakeholders; and (c) severity, amount of damage done by the accident. Attributions of crisis responsibility are believed to intensify when there is either a history of crises or the relationships with stakeholders have been negative. Crisis and relationship histories have an indirect effect on the reputational threat. Changes in history alter perceptions of crisis responsibility that, in turn, impact the organizational reputation (Coombs, 1995; Coombs & Holladay, 2001). A history of similar crises can indicate that a crisis is stable rather than unstable, thereby increasing attributions of crisis responsibility. The attribution concepts of consistency and distinctiveness are relevant to crisis history and relationship history. People draw on information that can affect their attributions of cause, and this information is labeled causal antecedents. There are three causal antecedents: consistency, behavior is consistent over time; distinctiveness, similar behavior in dissimilar situations; and consensus, others react in the same way (Kelley & Michela, 1980; Weiner, 1985b). Crises history is a form of consistency because it indicates the organization repeats a specific behavior—having crises. A negative performance history is a form of distinctiveness because the organization behaves poorly in a variety of settings. Increasing amounts of damage, a history of past crises, and/or negative relationships with stakeholders can have a direct, negative effect on organizational reputation. Figure 1 uses Coombs and Holladay’s (2002) crisis situation model to illustrate the relationships between the variables in SCCT. Crisis History: Principles, Past Research, and Hypotheses This study focuses on a narrow aspect of SCCT. Specifically, the present investigation examines the viability of past crises as intensifiers of crisis responsibility

272

JOURNAL OF BUSINESS COMMUNICATION

and/or reputational threat across the entire range of crisis clusters. Past crises are a potential indicator of stability because they suggest a particular pattern of behavior. The existence of one or more crises may indicate that the current crisis is part of a pattern (stable) rather than an isolated incident (unstable). Therefore, a history of past crises could lead to stronger attributions of organizational responsibility. Early tests using human-error accidents and organizational misdeeds crisis types compared two conditions: (a) a history of crises and (b) information indicating no previous crises conditions. The history of crises condition produced more crisis responsibility and created a greater reputational threat than the information indicating no crises condition (Coombs, 1998; Coombs & Holladay, 1996). However, these studies failed to determine if the two conditions differed from an unknown/neutral condition where no information about crisis history is reported. A third study, using a human-error accident, did incorporate all three conditions: information indicating no crises, unknown/neutral (no information about crisis history), and a history of crises. The results found a Velcro effect, that is, the history of crises condition was significantly different from the information indicating no crises and the neutral condition as related to attributions of crisis responsibility and organizational reputation. Using the information indicating no crises condition, the study posited that the attributions of crisis responsibility would be lower and organizational reputation scores higher than in the neutral condition. However, the results yielded no significant difference between the information indicating no crises and the neutral conditions for either crisis responsibility or organizational reputation. The study demonstrated that the history of crises condition was the driving force; indeed, this condition created stronger attributions of crisis responsibility and a greater reputational threat. A history of past crises definitely hurt an organization by increasing attributions of crisis responsibility and the reputational threat (Coombs & Holladay, 2001). Information about crisis history seems to matter most when it is unfavorable. A history of crises can have a Velcro-like effect by snagging additional reputational damage (Coombs & Holladay, 2001). Current research and corresponding evidence that would support crisis history acting as an intensifier is surprisingly limited. Human-error accidents and organizational misdeed crises originate from the same intentional crisis cluster. No research has addressed the extent to which the intensifying effect of crisis history holds true in the other two crisis clusters. Crises history acting as an intensifier should help direct which crisis response strategies should be used. SCCT posits that crises in the victim and accidental clusters should be treated like the next higher crisis cluster when an intensifier exists. More precisely, crises in the victim cluster should be treated like those in the accidental crises, and crises in the accidental cluster should be treated like those in the intentional cluster when there is a history of crises or some other intensifier is present. Researchers and, in particular, practitioners would benefit by knowing if crisis history acts as an intensifier in the victim and accidental crisis clusters because such knowledge might influence the design of postcrisis messages.

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

273

Relevant research from marketing has provided inconclusive results about the effect of a crisis history on attributions. Three studies examined the effect of a history of product recalls on a current recall. Although these investigations measured the effects on perceptions of organizations, they did not address crisis responsibility. Two studies found no effect for crisis history on perceptions of an organization (Mowen, 1980; Mowen, Jolly, & Nickell, 1981), whereas one study did find that history of product recalls negatively affected perceptions of the organization

Organizational misdeeds involve management knowingly violating laws or regulations and/or knowingly placing stakeholders at risk. Violating discrimination laws or allowing a product to go to market with a known defect are examples of intentional acts.

(Folkes, 1984). A history of crises alters the effect of a crisis by increasing the reputational threat directly and/or indirectly (Coombs & Holladay, 2001). The direct effect indicates that crises history leads to a greater reputational threat by lowering perceptions of the organization’s reputation. The indirect effect posits that crisis history intensifies attributions of crisis responsibility and that crisis responsibility impacts the organizational reputation. Figure 1 provides a visual depiction of these two effects. The discovery of the importance of crisis history in the intentional cluster research suggests the effect may be found in the other two crisis clusters. This study selected workplace violence and product tampering from the victim cluster, and technical-error recall and technical-error accident from the accidental cluster for further testing. The intensifying effect of crisis history on crisis responsibility should hold across all three crisis clusters. This proposition must be tested before crisis history’s intensifying effect on crisis responsibility becomes an accepted part of SCCT. Hypotheses 1 and 2 examine the proposition:
Hypothesis 1: In the victim cluster, a history of crises will produce stronger attributions of crisis responsibility than either information indicating no past crises or no presentation of information about past crises. Hypothesis 2: In the accident cluster, a history of crises will produce stronger attributions of crisis responsibility than either information indicating no past crises or no presentation of information about past crises.

274

JOURNAL OF BUSINESS COMMUNICATION

Previous research also found that a crisis history intensifies the reputational threat directly by lowering perceptions of the organization’s reputation (Coombs & Holladay, 2001). Again, the intensifying effect of crisis history on organizational reputation should hold across all three crisis clusters. Hypotheses 3 and 4 examine this issue:
Hypothesis 3: In the victim cluster, a history of crises will produce less positive perceptions of organizational reputation than either information indicating no past crises or no presentation of information about past crises. Hypothesis 4: In the accident cluster, a history of crises will produce less positive perceptions of organizational reputation than either information indicating no past crises or no presentation of information about past crises.

The crisis responsibility–organizational reputation relationship is a lynchpin of the situational approach. Accordingly, it is important to determine if the relationship holds across the various crisis clusters (Coombs, 1995; Coombs & Holladay, 2001):
Hypothesis 5: Crisis responsibility and organizational reputation will have a significant, negative correlation in both the victim and accident crisis clusters.

METHOD

Participants The respondents for this study were 321 undergraduate students and community members living in a large Midwestern city in the United States. Crises produce unique forms of publics for organizations, victims, and nonvictims. Victims are those who are directly affected in some way (e.g., evacuated from an area, injured physically, or lost property). The focus of the present study is on nonvictims. Nonvictims are not injured by the crisis but follow the crisis in the news media and therefore may have weakly held perceptions of the organization. Nonvictims are important because these people may currently or in the near future interact with the organization. How the organization handles the crisis, therefore, may affect nonvictims’ future interaction with the organization, that is, it can shape interactions with stakeholders (Newsom, VanSlyke Turk, & Kruckeberg, 2000; Sturges, Carrell, Newsom, & Barrera, 1994). The students and community members fit the parameters of the nonvictim population because they share the characteristics of neither being directly affected by the crisis nor holding strong perceptions of the organization prior to the crisis. A community member was defined as being at least 24 years old and not attending school. Of the respondents, 44.3% (n = 147) were from the community and 55.7% (n = 185) were students. Of the participants, 65.7% were women (n = 218) and 34.3% were men (n = 114). The respondents ranged in age from 18 to 59 years

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

275

old (M = 26, SD = 8.76). The students were drawn from communication classes and general education classes, and thus the sample likely represented a wide range of majors. The community members were drawn from a variety of local organizations, including heavy industry, service, and city government. A preliminary analysis tested for differences in responses to the dependent variables for community/ student and gender. The one-way analyses of variance (ANOVA) found no significant difference between students and community members for either crisis responsibility, F(1, 330) = .21, p = .65, or organizational reputation, F(1, 330) = 2.39, p = .13. ANOVA found no significant difference between women and men for either crisis responsibility, F(1, 330) = 1.50, p = .22, or organizational reputation, F(1, 330) = .56, p = .45. Neither the community/student nor gender differences should affect the results of the study significantly. Design and Materials With the focus on examining the intensifying effect of crisis history across crisis clusters, an array of crisis types were needed for the study. Two crisis types were selected from the victim and accidental crisis clusters identified by Coombs and Holladay (2002): Workplace violence and product tampering were selected from the victim cluster. Rumor was not selected because the cluster creates a unique dynamic with the need to deny the crisis. Natural disaster was not selected because a previous study found that the severity intensifier failed to have an effect on this crisis type (Coombs & Holladay, 2004). A technical-error recall and a technicalerror accident were selected from the accidental cluster. Challenge was not selected because, like rumor, it creates a unique dynamic with the possible need to counter the charges (Hearit, 1996, 2001). Four different crisis scenarios were constructed, one for each crisis type. All four crisis scenarios were based on actual events, and the core elements were derived from news reports about the crisis. The events were all at least 5 years old so they would not likely be familiar to respondents. The corporate names were evaluated with a pretest and Delahye Medilink and Reputation Institute (MRi) survey. A pretest using a similar student population revealed that none of the four companies was identified as eliciting strong positive or negative responses. The MRi survey ranks organizations on the amount and quality (favorable or unfavorable) of media coverage each receives (Brown & Roed, 2001; Calabro, 2003). None of the companies used in the study was in the top 20 of the MRi, and, therefore, community members would know little about them from media coverage. The precautions taken in the study indicate that the respondents would neither have familiarity with the case nor hold strong prestudy perceptions of the companies. The company names were Burroughs-Welcome, Sharp Electronics, Caterpillar, and Radisson. The crisis scenarios featured the key characteristics that defined and differentiated the crisis types. The crises descriptions were balanced so that they were equal in length. Table 2 provides a short description of each crisis scenario used in the study.

276

JOURNAL OF BUSINESS COMMUNICATION

Table 2. Crisis Type

Crisis Scenarios Respondents (n) 29 32 38 20 23 27 20 33 27 19 26 27 Crisis Description A Radisson employee shoots five coworkers. Radisson has had no similar incidents in the past. Radisson has had other incidents of workplace violence in past 3 years. Sudafed tampering in Washington State that killed two people. No previous product tampering cases for this manufacturer and it has strong record of product safety. Three previous product tampering cases for this manufacturer. Sharp recalled televisions that had a switch that would wear out over time and could then cause a fire—poorly built component from a supplier. Sharp has no recent recalls in past 5 years. Three recalls in the past 5 years. Ashland Oil had an oil storage tank burst and send oil into the Ohio River. The cause was faulty steel in the tank walls—manufacturing flaw. Ashland has a strong safety record and no recent accidents. Ashland has a weak safety record and three recent accidents.

Workplace violence Neutral No past crises History of crises Product tampering Neutral No past crises History of crises Technical error: Recall Neutral No past crises History of crises Technical error: Accident Neutral No past crises History of crises

The experimental design involved the manipulation of crisis history for each of the crisis types. Each crisis scenario was created to have three conditions: (a) a history of similar crises, (b) no presentation of information about past crisis history (unknown crisis history/neutral), and (c) information indicating no past crises. An unknown/neutral condition allows for a better judgment about the effect of crisis history and information indicating no past crises conditions. An important aspect of the present study is to determine the extent to which crisis history and information indicating past crises conditions produce effects different than from the results produced by the unknown/neutral condition. In actual crises, stakeholders can easily find themselves in an unknown/neutral condition (i.e., no history information is known). Stakeholders often know little about the specifics of an organization in crisis, including past safety performance, unless the media supply such information. How many people could identify the number of people killed at Disneyland before the September 2003 death on Big Thunder Mountain? Therefore, the unknown/ neutral condition is a valuable aspect of the research design. The 4 (crisis types) × 3 (history of crisis) design required the development of 12 different scenarios. Each of the four crisis types (workplace violence, product tampering, technical-error accident, and technical-error recall) had three variations: a

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

277

history of crises, no information about crisis history, and information indicating no similar crises in the past. Two different crisis history paragraphs reflecting the history of crises and information indicating no past crises conditions for crisis history were developed for each of the four crisis types used in the study. A paragraph about no mention of crisis history was created for the unknown/neutral condition by including generic information about the organization or town. Crisis history was operationalized through the crisis history paragraphs. Table 2 includes a list of the number of respondents for each crisis scenario. The information indicating no past crises condition stated the organization had a strong safety record, no recent recalls, no previous incident of workplace violence, or no previous cases of product tampering. The history of crises condition indicated a number of similar accidents, recent product recalls, previous workplace violence incidents, or past cases of product tampering. Efforts were made to keep all six different crisis scenarios equal in length with 19 lines of text each. Mock newspaper reports of the crisis were written because newspapers are one of the primary sources of business information; indeed, they are the source most people use to retrieve information about corporations (Deephouse, 2000). Furthermore, a nonvictim public will experience a crisis predominantly through the news media, probably the newspaper. Most people receive information about corporations through indirect experiences such as the news media (Carroll & McCombs, 2003). The use of newspaper articles as a stimulus matches well with how nonvictim publics generally encounter a crisis. A reputation is developed through direct and indirect experience with an organization (Brown & Roed, 2001). To study the long-term development of a reputation would require documenting a number of sources that provide people with information about an organization. However, people also rely on and use smaller bits of information that are salient when asked to make reputation judgments (Carroll & McCombs, 2003). The news media, particularly newspapers, are effective at producing an agenda setting effect for organizational reputations. The information reported in the media is given salience when the reputation is formed, that is, media coverage can shape a reputation (Carroll & McCombs, 2003; Deephouse, 2000; Wartick, 1992). This process helps to explain why a crisis and media coverage of a crisis are important to an organizational reputation and why they receive so much attention in writings about crisis management (e.g., Barton, 2001; Ogrizek & Guillery, 1999). What is reported about the crisis can have a profound effect on the organization’s reputation. Measures Crisis responsibility was measured using a 3-item scale based on the work of Griffin et al. (1992). Previous research reported reliabilities ranging between .80 and .91 for the crisis responsibility scale (Coombs, 1998; Coombs & Holladay, 2001, 2002; Coombs & Schmidt, 2000). Organizational reputation was measured using a 5-item version of the Organizational Reputation Scale that is based on

278

JOURNAL OF BUSINESS COMMUNICATION

McCroskey’s (1966) character scale. Extant research reported reliabilities ranging between .82 and .92 for the Organizational Reputation Scale (Coombs, 1998; Coombs & Holladay, 2001, 2002; Coombs & Schmidt, 2000). The instrument included one manipulation check item for crisis history. All of the scale items ranged from 1 (strongly disagree) to 5 (strongly agree). Procedures Each respondent received a packet containing a cover page with directions, a crisis scenario, and a two-page questionnaire. Distribution of the packets was randomized among respondents. In addition to written instructions, respondents were instructed orally to read the crisis scenario carefully and respond to the survey questions based on what they had just read. The instructions also appeared on the first page of the packet. All respondents were told that the study involved perceptions of organizations. Students were administered the materials in class. Five graduate students with full-time jobs were trained to collect data from the community. The graduate students were instructed on the oral directions to provide respondents and not to offer clarification of the materials or to answer questions about the survey items. Graduate students recruited coworkers and provided contact information that could be used to verify a person’s participation in the study. The administration of the materials required 13 to 18 minutes.
RESULTS

Reliabilities The reliability analysis of the items produced an internal consistency of .85 (Cronbach’s alpha) for crisis responsibility and .81 (Cronbach’s alpha) for the Organizational Reputation Scale. All reliability scores fell within the acceptable range (Morgan & Griego, 1998). For each individual factor, the items were summed, resulting in the creation of composite scores for crisis responsibility and organizational reputation. The creation of the composite scores follows the procedures used in past situational approach research (e.g., Coombs & Schmidt, 2000). Manipulation Checks A series of one-way ANOVAs were performed to assess the effectiveness of the crisis history manipulation. To assess perceptions of having had a history of similar crises, the three crisis history conditions were compared on the item, “The organization has a history of similar crises.” The item was the last item on the survey, and was selected in an effort to minimize any potential effects. In addition, including this item from the study provides for a robust test of the manipulation (Perdue & Summers, 1986) and is consistent with procedures recommended for communication

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

279

Table 3.

Manipulation Check for Item History of Similar Crises

Crisis History Condition Unknown Information History/ Indicating No Crisis No Past Crises History Given Crisis Type Workplace violence Product tampering Technical accident Technical-error recall M 1.66a 1.78a 1.50a 1.82a SD 0.94 0.80 0.71 0.85 M 2.48b 1.70a 2.42b 2.15a SD 1.15 0.80 0.84 1.04

History of Past Crises M 4.68c 3.48b 4.48c 4.15b SD 0.81 0.89 0.70 0.91 F 93.63 35.84 111.81 52.42 df 96 67 69 77 p < .001 < .001 < .001 < .001

NOTE: For each test, means indexed by different subscripts are significantly different using Dunette C procedure; p < .01.

research (Frey, Botan, Friedman, & Kreps, 1991). Separate one-way ANOVAs were conducted for the two crisis clusters and four crisis types. Table 3 presents the results of the analyses. All six analyses found a significant difference for the crisis history manipulation item. For the victim cluster, accident cluster, workplace violence, and technical-error accident crisis type, all three conditions differed from one another. Respondents rated the history of crises condition significantly higher than the information indicating no past crisis condition or the unknown/neutral condition. The information indicating no past crisis condition was perceived as having less similar crises than the history of crises or unknown/neutral condition. These differences emerged as expected. When crises involved product tampering and technical-error product recalls, respondents perceived the reputation of the organization with a history of such crises as significantly more negative than the reputation of an organization without past crises or with an unknown history. However, there was no significant difference between the perceived reputation of an organization without past crises and one with an unknown history. The product recall crisis type followed the predicted pattern: Respondents rated the information indicating no past crisis condition lower than the unknown/neutral crisis history condition. Ideally, the manipulation should have produced a difference between the unknown/neutral condition and information indicating no past crisis conditions. However, in the product tampering and technical-error product recall crisis types, the unknown/neutral conditions generated low perceptions of crisis history. These findings seem to suggest a positive assumption about the crisis history: No mention of a crisis history is viewed the same as information indicating no past crisis. The history of past crises manipulation was clearly a success. The information indicating no past crises manipulation was minimally successful and must be considered when interpreting the results. The product tampering and workplace violence crisis scenarios were shown to generate high attributions of external control and low perceptions of personal

280

JOURNAL OF BUSINESS COMMUNICATION

control (Coombs et al., 1995). Such attributions indicate that respondents viewed the organization as a victim, that is, external forces not the organization were responsible for the crisis. In another study, Coombs (2003) examined differences between technical and human-error accidents, and product recalls were examined by asking respondents to list the cause of the crisis. Technical factors were the dominant cause in the technical-error crises scenarios and human actions the dominant cause in the human-error crises scenarios. Respondent perceptions of the four crisis scenarios had been established in previous studies, thereby avoiding the fundamental attribution research error—researchers and respondents perceiving the stimuli differently (McAuley et al., 1992). Tests of Hypotheses A series of MANOVAs were used to test Hypotheses 1 to 4. The victim cluster MANOVA identified a significant effect for crisis history and crisis responsibility, F(2, 136) = 8.11, p < .001, η2 = .11, power = .96. The LSD post hoc analysis discovered the history of past crises condition (M = 2.50) was perceived as producing significantly greater attributions of crisis responsibility than either the information indicating no past crises (M = 2.03) or the unknown/neutral condition (M = 1.81). There was no significant difference between the information indicating no past crises and unknown/neutral conditions for the crisis response scores. Support was found for Hypothesis 1. To test Hypothesis 2, the MANOVA examined the relationship between crisis history and crisis responsibility in the accident cluster. The accident cluster MANOVA identified a significant effect for crisis history and crisis responsibility, F(2, 133) = 3.47, p < .05, η2 = .05, power = .64. The LSD post hoc analysis discovered the history of past crises condition (M = 3.17) was perceived as having significantly greater attributions of crisis responsibility than the information indicating no past crises (M = 2.68) but not the unknown/neutral condition (M = 2.82). There was no significant difference between the information indicating no past crises and unknown/neutral conditions for the crisis responsibility scores. Partial support was found for Hypothesis 2. MANOVA examined the relationship between crisis history and organizational reputation in the victim cluster to test Hypothesis 3. The victim cluster MANOVA revealed a significant effect for crisis history and organizational reputation, F(2, 136) = 20.16, p < .001, η2 = .23, power = 1.00. The LSD post hoc analysis discovered the respondents in the history of past crises condition (M = 3.29) reported significantly lower organizational reputation scores than either the information indicating no past crises (M = 3.98) or the unknown/neutral condition (M = 4.16). There was no significant difference between the information indicating no past crises and unknown/neutral conditions for the organizational reputation scores. Support was found for Hypothesis 3. To test Hypothesis 4, the MANOVA looked for a relationship between crisis history and organizational reputation in the accident cluster. The accident cluster

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

281

MANOVA found a significant effect for crisis history and organizational reputation, F(2, 133) = 8.93, p < .001, η2 = .12, power = .97. The LSD post hoc analysis discovered the respondents in the history of past crises condition (M = 3.07) reported significantly lower organizational reputation scores than either the information indicating no past crises (M = 3.67) or the unknown/neutral condition (M = 3.62). There was no significant difference between the information indicating no past crises and unknown/neutral conditions for the organizational reputation scores. Support was found for Hypothesis 4.

When crises involved product tampering and technical-error product recalls, respondents perceived the reputation of the organization with a history of such crises as significantly more negative than the reputation of an organization without past crises or with an unknown history .

Finally, a series of correlational analyses were conducted for the two crisis clusters. The objective was to determine if the key relationship of crisis responsibility and organizational reputation held across the clusters. Crisis responsibility and organizational reputation correlated at –.49 (r < .01) in the accidental cluster and –.60 (r < .01) in the victim cluster. As expected, the significant, negative crisis responsibility–organizational reputation relationship held across the three crisis clusters, offering support for Hypothesis 5. Follow-Up Analyses A crisis cluster is a macro-level concept because it groups together a number of crisis types. There are unique qualities associated with each crisis type that a macro-level could mask. Therefore, each of the four crisis history hypotheses advanced in the study were examined for microlevel results. As with the hypotheses, MANOVA analyses were used to test for relationships. Table 4 presents the full results of the MANOVA follow-up analyses. For workplace violence and technicalerror accident, the history of past crises lead to significantly stronger attributions of crisis responsibility and lower perceptions of organizational reputation than either the information indicating no past crises or the unknown/neutral condition. In addition, there was no significant difference between the information indicating no past crises and the unknown/neutral condition for these two crisis types.

282

JOURNAL OF BUSINESS COMMUNICATION

TABLE 4. Crisis Cluster and Crisis Type Victim cluster

Summary of Hypothesis and Follow-Up Tests Significant Differences Between Conditions Crisis Responsibility History of crises unknown and information indicating no crisis History of crises unknown and information indicating no crisis Unknown history of crises Information indicating no crisis, history of crises None History of crises unknown and information indicating no crisis Organizational Reputation History of crises unknown and information indicating no crisis History of crises unknown and information indicating no crisis History of crises unknown and information indicating no crisis History of crises unknown and information indicating no crisis None History of crises unknown and information indicating no crisis

Workplace violence Product tampering Accidental cluster Technical-error recall Technical-error accident

For technical-error recall, MANOVA found no relationship between either crisis history and crisis responsibility or crisis history and organizational reputation. For product tampering and crisis responsibility, the history of past crises and unknown/neutral condition were significantly different from one another. There was no difference between the information indicating no crises and either a history of crises or unknown/neutral conditions. For product tampering and organizational reputation, the history of past crises lead to significantly stronger attributions of crisis responsibility and lower perceptions of organizational reputation than either the information indicating no past crises or the unknown/neutral condition. In addition, there was no significant difference between the information indicating no past crises and the unknown/neutral condition.
DISCUSSION

Summary This study examined the extent to which reputational threat of a victimization (mild threat) or accidental (moderate threat) crisis increased when the organization had a history of similar crises. We hypothesized that crisis history increased perceptions of crisis responsibility and intensified negative evaluations of organizational reputation. The results revealed a weak (i.e., effect size) link between crisis history and perceptions of crisis responsibility (Hypotheses 1 and 2) and a strong link between crisis history and organizational reputation (Hypotheses 3 and 4) in the victim and accident clusters. Follow-up analyses found the same pattern of relationships for the four crisis types used in the study. Except for the technical-error product recall, a history of past crises did intensify the attributions of crisis responsibility and lowered perceptions of the organizational reputation.

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

283

No significant benefit was accrued when information indicating no past crises occurred was presented. The conditions with information indicating no past crises were no different than conditions where the crisis history was unknown. One possible explanation is that people expect that organizations will operate safely. We expect products to be safe and for chemical facilities not to explode (Massey, 2001). When news media provide no information about past crises, stakeholders are likely to assume there were no past crises. A premium is still placed on safe and proper operations. An organization wants to maintain a clean crisis record because similar past crises will complicate the management of the current crisis. A history of past crises alters the reputational threat by affecting perceptions of crisis responsibility and organizational reputation. An organization is best served by efforts to maintain strong safety and product quality records, that is, the absence of crises. For instance, when a West Pharmaceutical facility was destroyed by a dust explosion accident in 2003, the lack of prior accidents kept the crisis at a mild level of reputational threat. In contrast, when Ford Motor Company’s Rouge facility had an explosion in 1999, the explosions of 1989 and 1986 were noted and the crisis moved to a severe reputational threat (Maynard, 1999). The results also support the claim that the relationship between crisis responsibility and organizational reputation, a core component of SCCT, held across the victim and accident crisis clusters. The negative relationship between crisis responsibility and organizational reputation proved to be generalizable because both were found in each of the three crisis clusters identified in SCCT. The correlations in this study indicate that the reputational threat of a crisis increases as attributions of crisis responsibility intensify. Crisis history was found to have a direct and an indirect effect on organizational reputation. The direct effect was the connection between crisis history and organizational reputation found in Hypotheses 3 and 4. The indirect effect was a result of intensifying perceptions of crisis responsibility (Hypotheses 1 and 2) that, in turn, increase reputational damage (Hypothesis 5). A crisis originally considered a mild reputational threat moved to the moderate threat level and a crisis originally considered a moderate reputational threat moved to the severe threat level when the organization had a history of crises. Practical Implications In January 1997, an explosion at the Tosco Refinery Co. near Martinez, California, could be felt 20 miles away. The newspaper stories mentioned explosions from 1995 and 1993 (e.g., Hallissy, Lee, & Minton, 1997). A February 1999 blast at Tosco was framed by the 1997 explosion because one of the injured men had been injured by that blast as well (Heredia, Finz, & Schevitz, 1999). In each instance, the Tosco management should have managed the crisis situation as a serious reputational threat. Even if the cause was equipment failure or some other technical error, the crisis history served to intensify the reputational threat. Tosco management would need to use compensation and/or full apology rather than excuse and/or

284

JOURNAL OF BUSINESS COMMUNICATION

justification response strategies, the recommended responses for a technical-error accident (Coombs & Holladay, 1996; Coombs & Schmidt, 2000). The same holds true for victimization crises. The presence of a history of crises means that workplace violence or product tampering should be treated as if it belongs to the accident crisis cluster. Such information has a direct effect on the selection of crisis response strategies. When a second Tylenol tampering occurred in the 1980s, Johnson & Johnson faced a moderate reputational threat (accident cluster) rather than a low reputational threat (victim cluster). A history of product tampering meant that simply presenting instructing information, a viable strategy for victimization crises, would no longer be enough (Coombs & Holladay, 2004). By increasing the reputational threat, crisis history alters what crisis response strategies are appropriate.

The results of this study demonstrated the important effect of crisis history on organizational reputation in victim and accident crises. The results showed a direct, negative relationship between crisis history and organizational reputation.

The results of this study demonstrated the important effect of crisis history on organizational reputation in victim and accident crises. The results showed a direct, negative relationship between crisis history and organizational reputation. A history of similar crises lowered perceptions of an organization’s reputation. The results support an indirect relationship between crisis history and organization reputation, too. First, a history of similar crises was found to increase attributions of crisis responsibility. The relationship was weak but did exist. Second, crisis responsibility demonstrated a negative relationship to organizational reputation in the victim and accident crises. The crisis responsibility–organizational reputation relationship, a central element of SCCT (Coombs, 1995, 1998; Coombs & Holladay, 1996, 2001, 2002), was shown to hold across the three crisis clusters in SCCT: intentional, accident, and victim. We can be reasonably confident that attributions of crisis responsibility are negatively related to perceptions of organizational reputation in crises. By altering perceptions of crisis responsibility, a history of past crises had an indirect effect on organizational reputation as well. Clearly, past crises are an important part of the interpretive framework of present crisis experienced by organizations. The news media often use past crises as frames for current crises. Ford’s River Rouge facility and Tosco Refinery in Avon,

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

285

California, are examples. Each new accident produced news reports that noted the previous crises (e.g., Hallissy et al., 1997; Heredia et al., 1999). Crisis managers should be able to anticipate and calculate the effect of a history of crises when selecting response strategies. Crisis managers should review past crises as a form of crisis history audit to document an accurate picture of the organization’s crisis history. The crisis history audit would provide reliable information on the crisis history that could help to more accurately evaluate the reputational threat posed by the current situation. Crisis managers always have the option of trying to fight stakeholder attributions, using crisis response strategies in an effort to alter attributions of crisis responsibility. For instance, the crisis managers can maintain the current accident was simply an accident (i.e., excuse crisis response strategy). However, such an approach can be risky. Although crisis responsibility can vary from person to person, each crisis type can create a consistent, average crisis responsibility score among nonvictims (Coombs, 1998; Coombs et al., 1995; Coombs & Holladay, 2002). Perceptions of a crisis are a constraint that is difficult to change. People will expect responses that fit the demands of the situation. Both case study and experimental research have found little success when crisis managers offer information that contradict expected responses (e.g., Bradford & Garrett, 1995; Coombs & Holladay, 1996; Hearit, 1996). Fighting the attributions of crisis responsibility can result in a self-serving and ultimately self-defeating victory. The organization may defend its interpretation of the crisis, but such an interpretation is unlikely to have much appeal with its nonvictims, resulting in the further erosion of the organization’s reputation. An organization is served better by accepting and working within the constraints of crisis responsibility attributions. Future Research Future research should consider the difference between crisis types (microlevel) and crisis clusters (macro-level). Although crisis clusters are a useful method for grouping crises and developing crisis portfolios (Mitroff, 1988; Mitroff et al., 1996), they might mask important microlevel differences in SCCT. For instance, crisis history affects technical-error product recalls very little. This finding suggests that technical-error product recalls might possess a unique dynamic that does not fit the general crisis situation model articulated by SCCT. These microlevel differences should be further explored by examining the crisis situation dynamics of individual crisis types. Researchers should carefully map the interaction of SCCT variables for each crisis type to identify and explain any unique patterns. The model of the crisis situation advanced by Coombs and Holladay (2001) deserves greater scrutiny and testing. Thus far, the assessment of their crisis situation model has relied on basic correlations and regression analyses. More advanced statistical analyses, such as path analysis, should be used to assess the appropriateness of the model. As results of this study suggest, some of the connections grew weaker as a larger array of crisis clusters and crisis types were examined. There is

286

JOURNAL OF BUSINESS COMMUNICATION

also the possibility that the exact configuration of the crisis situation model may vary slightly from crisis cluster to crisis cluster, or crisis type to crisis type. These concerns warrant continued study. Limitations Experimental studies are always limited to some extent by the variables and respondents used. To exercise control, a researcher has to select a set of variables to use and exclude (Frey et al., 1991). A purposeful choice was made to limit the present study to crisis history and the effects of such on perceptions of an organization in crisis. As with earlier SCCT-related studies, the crisis history manipulation was given to respondents; participants did not experience the crisis. Although having participants actually experience the crisis might provide for a powerful manipulation, it is also true that most people learn about organizations through media reports and have their perceptions of organizations shaped by that media coverage (Carroll & McCombs, 2003; Deephouse, 2000). We chose to focus on the later phenomenon. The limited success of the information indicating no past crisis history manipulation must be considered. It could be argued that only the history of past crises would produce significant differences because no significant difference was found for the manipulation check item between the information indicating no past crises and unknown/neutral history conditions for two of the four crisis types. A number of factors mitigate that concern. First, the information indicating no past crisis manipulation was successful for the accident and victim crisis clusters, workplace violence crisis type, and the technical-error accident crisis type. The negative effect of crisis history appeared even when the information indicating no crisis history manipulation succeeded, that is, a successful information indicating no crisis history manipulation did not negate the effect of a history of crises. Second, the crisis history scores for the unknown/neutral crisis history conditions were very low for the four crisis types, ranging from 2.42 to 1.70. The two lowest scores were for product tampering (1.70) and technical-error product recall (2.15), the two crisis types where the manipulation check for the information indicating no past crises failed. Such low scores provide little room for creating significant differences between the information, indicating no crisis history and neutral conditions. Overall, the failure of the information indicating no past crises manipulation scores probably had little effect on the results of the study.
CONCLUSION

Despite limitations, the present study offers significant insight into the effect of crisis on perceptions of organizations involved in crisis. Although logically appealing, the intensifying effect of crisis history had to be tested to build reliable social science. Results from this investigation support that part of SCCT which suggests

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

287

that crisis history should result in crisis managers selecting response strategies that accept greater responsibility and that demonstrate increased concern for victims’ needs than would normally be used for a given crisis situation. By accounting for the effects of crisis history, crisis managers can craft messages that more effectively protect the organization’s reputational assets.

REFERENCES Allen, M. W., & Caillouet, R. H. (1994). Legitimate endeavors: Impression management strategies used by an organization in crisis. Communication Monographs, 61, 44-62. Barton, L. (2001). Crisis in organizations II (2nd ed.). Cincinnati, OH: College Divisions South-Western. Benoit, W. L. (1995). Accounts, excuses, and apologies: A theory of image restoration. Albany: State University of New York Press. Bergman, E. (1994). Crisis? What crisis? Communication World, 11(4), 9-13. Bradford, J. L., & Garrett, D. E. (1995). The effectiveness of corporate communicative responses to accusations of unethical behavior. Journal of Business Ethics, 14, 875-892. Brown, K. C. (2003). Laying down the law: Implications of reputation management for the C-suite. The Gauge, 16(3), 1-2. Retrieved December 3, 2003, from http://www.thegauge.com/ v16n3laydownlawprint.htm Brown, K. C., & Roed, B. (2001). Delahaye Medialink’s 2001 media reputation index results. The Gauge, 14(2), 1-2. Retrieved December 3, 2003, from http://www.thegauge.com/SearchFolder/ OldGauges/Vol15No2/mriPart1152.html Calabro, S. (2003, January 27). Microsoft claims top spot in latest reputation survey. PRWEEK, p. 1. Carroll, C. E., & McCombs, M. (2003). Agenda-setting effects of business news on the public’s image and opinions about major corporations. Corporate Reputation Review, 6, 36-46. Carter, C. J. (2003, September 6). One killed, 10 injured on “Thunder Mountain” ride. Associated Press. Cohen, J. R. (1999). Advising clients to apologize. Southern California Law Review, 72, 1009-1131. Coombs, W. T. (1995). Choosing the right words: The development of guidelines for the selection of the “appropriate” crisis response strategies. Management Communication Quarterly, 8, 447-476. Coombs, W. T. (1998). An analytic framework for crisis situations: Better responses from a better understanding of the situation. Journal of Public Relations Research, 10, 177-191. Coombs, W. T. (1999a). Information and compassion in crisis responses: A test of their effects. Journal of Public Relations Research, 11, 125-142. Coombs, W. T. (1999b). Ongoing crisis communication: Planning, managing, and responding. Thousand Oaks, CA: Sage. Coombs, W. T. (2003, November). Technical or human error: Mutability and why that distinction matters in crisis communication. Paper presented at the annual meeting of the National Communication Association, Miami, FL. Coombs, W. T. , Hazleton, V., Holladay, S. J., & Chandler, R. C. (1995). The crisis grid: Theory and application in crisis management. In L. Barton (Ed.), New avenues in risk and crisis management (Vol. 4, pp. 30-39). Las Vegas: UNLV Small Business Development Center. Coombs, W. T., & Holladay, S. J. (1996). Communication and attributions in a crisis: An experimental study of crisis communication. Journal of Public Relations Research, 8, 279-295. Coombs, W. T., & Holladay, S. J. (2001). An extended examination of the crisis situation: A fusion of the relational management and symbolic approaches. Journal of Public Relations Research, 13, 321340. 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, 165186.

288

JOURNAL OF BUSINESS COMMUNICATION

Coombs, W. T., & Holladay, S. J. (2004). Reasoned action in crisis communication: An attribution theory-based approach to crisis management. In D. P. Millar & R. L. Heath (Eds.), Responding to crisis: A rhetorical approach to crisis communication (pp. 95-115). Mahwah, NJ: Lawrence Erlbaum. Coombs, W. T., & Schmidt, L. (2000). An empirical analysis of image restoration: Texaco’s racism crisis. Journal of Public Relations Research, 12, 163-178. Deephouse, D. L. (2000). Media reputation as a strategic resource: An integration of mass communication and resource-based theories. Journal of Management, 26, 1091-1112. Druckman, J. N. (2001). On the limits of framing effects: Who can frame? Journal of Politics, 63, 10411066. Fearn-Banks, K. (1996). Crisis communications: A casebook approach. Mahwah, NJ: Lawrence Erlbaum. Fediuk, T. A. (1999). Corrective action and concern during an accident: An analysis of responses communicated in a crisis situation. Unpublished master’s thesis, Illinois State University, Normal. Folkes, V. S. (1984). Consumer reactions of product failure: An attributional approach. Journal of Consumer Research, 10, 398-409. Fombrun, C. J. (1996). Reputation: Realizing value from the corporate image. Boston: Harvard Business School Press. Frey, L. R., Botan, C. H., Friedman, P. G., & Kreps, G. L. (1991). Investigating communication: An introduction to research methods. Englewood Cliffs, NJ: Prentice Hall. Griffin, M., Babin, B. J., & Attaway, J. S. (1991). An empirical investigation of the impact of negative publicity on consumer attitudes and intentions. Advances in Consumer Research, 18, 334-341. Griffin, M., Babin, B. J., & Darden, W. R. (1992). Consumer assessments of responsibility for productrelated injuries: The impact of regulations, warnings, and promotional policies. Advances in Consumer Research, 19, 870-877. Hallissy, E., Lee, H. K., & Minton, T. (1997, January 22). Refinery explosion kills 1, injures 24. The San Francisco Chronicle, p. A1. Harris Interactive. (2000). 20 attributes of reputation. Retrieved October 27, 2000, from http:// harrisinteractive.com/pop_up/rq/20.asp Hearit, K. M. (1996, Fall). The use of counter-attack in apologetic public relations crises: The case of General Motors vs. Dateline NBC. Public Relations Review, 22(3), 233-248. Hearit, K. M. (2001). Corporate apologia: When an organization speaks in defense of itself. In R. L. Heath (Ed.), Handbook of public relations (pp. 501-511). Thousand Oaks, CA: Sage. Heredia, C., Finz, S., & Shevitz, T. (1999, February 24). Deadly blast at Tosco plant. The San Francisco Chronicle, p. A1. Kelley, H. H., & Michela, J. L. (1980). Attribution theory and research. Annual Review of Psychology, 31, 457-501. Lerbinger, O. (1997). The crisis manager: Facing risk and responsibility. Mahwah, NJ: Lawrence Erlbaum. Massey, J. E. (2001). Managing organizational legitimacy: Communication strategies for organizations in crisis. Journal of Business Communication, 38, 153-182. Maynard, M. (1999, February 2). 1 dead 21 hurt in Ford plant explosion. USA Today, p. 3A. McAuley, E., Duncan, T. E., & Russell, D. W. (1992). Measuring causal attributions: The revised causal dimension scale (CDII). Personality and Social Psychology Bulletin, 18, 566-573. McCroskey, J. C. (1966). An introduction to rhetorical communication. Englewood Cliffs, NJ: Prentice Hall. Mitroff, I. I. (1988, Winter). Crisis management: Cutting through the confusion. Sloan Management Review, 29, 15-20. Mitroff, I. I., & Anagnos, G. (2001). Managing crises before they happen: What every executive and manager needs to know about crisis management. Chicago: AMACOM. Mitroff, I. I., Harrington, K., & Gai, E. (1996, September). Thinking about the unthinkable. Across the Board, 33(8), 44-48.

Coombs / SITUATIONAL CRISIS COMMUNICATION THEORY

289

Morgan, G. A., & Griego, O. V. (1998). Easy use and interpretation of SPSS for windows: Answering research questions with statistics. Mahwah, NJ: Lawrence Erlbaum. Morris, M. W., Moore, P. C., & Sim, D. L. H. (1999). Choosing remedies after accidents: Counterfactual thoughts and the focus on fixing “human error.” Psychonomic Bulletin & Review, 6, 579-585. Mowen, J. C. (1980). Further information on consumer perceptions of product recalls. Advances in Consumer Research, 7, 519-523. Mowen, J. C., Jolly, D., & Nickell, G. S. (1981). Factors influencing consumer responses to product recalls. Advances in Consumer Research, 8, 405-407. Nakra, P. (2000, Summer). Corporate reputation management: “CRM” with a strategic twist? Public Relations Quarterly, 45(2), 35-42. Newsom, D., VanSlyke Turk, J., & Kruckeberg, D. (2000). This is PR: The realities of public relations (7th ed.). Belmont, CA: Wadsworth/Thomson Learning. Ogrizek, M., & Guillery, J. M. (1999). Communicating in crisis: A theoretical and practical guide to crisis management. New York: Aldine de Gruyter. Patel, A., & Reinsch, L. (2003). Companies can apologize: Corporate apologies and legal liability. Business Communication Quarterly, 66, 17-26. Pearson, C. M., & Mitroff, I. I. (1993). From crisis prone to crisis prepared: A framework for crisis management. The Executive, 7, 48-59. Perdue, B. C., & Summers, J. O. (1986). Checking the success of manipulations in marketing experiments. Journal of Marketing Research, 23, 317-326. Perrow, C. (1999). Normal accidents: Living with high-risk technologies. Princeton, NJ: Princeton University Press. Reason, J. (1999). Human error. New York: Cambridge University Press. Rendon, R., & Watson, B. (2000). Twist of fate. The Houston Chronicle, p. A-1. Russell, D. (1982). The causal dimension scale: A measure of how individuals perceive causes. Journal of Personality and Social Psychology, 42, 1137-1145. Sturges, D. L. (1994). Communicating through crisis: A strategy for organizational survival. Management Communication Quarterly, 7, 297-316. Sturges, D. L., Carrell, B., Jr., Newsom, D., & Barrera, M. (1994). Crisis communication: Knowing how is good; knowing why is essential. In M. B. Goodman (Ed.), Corporate communication: Theory and practice (pp. 339-353). Albany: State University of New York Press. Tyler, L. (1997). Liability means never being able to say you’re sorry: Corporate guilt, legal constraints, and defensiveness in corporate communication. Management Communication Quarterly, 11, 51-73. Wartick, S. L. (1992). The relationship between intense media exposure and change in corporate reputation. Business & Society, 31, 33-49. Weiner, B. (1985a). An attributional theory of achievement motivation and emotion. Psychology Review, 92, 548-573. Weiner, B. (1985b). Human motivation. New York: Springer-Verlag. Weiner, B. (1995). Judgments of responsibility: A foundation for a theory of social conduct. New York: Guilford. Weiner, B., Amirkhan, J., Folkes, V. S., & Verette, J. A. (1987). An attribution analysis of excuse giving: Studies of a naive theory of emotion. Journal of Personality and Social Psychology, 53, 316-324. Weiner, B., Perry, R. P., & Magnusson, J. (1988). An attribution analysis of reactions to stigmas. Journal of Personality and Social Psychology, 55, 738-748. Wilson, S. R., Cruz, M. G., Marshall, L. J., & Rao, N. (1993). An attribution analysis of compliancegaining interactions. Communication Monographs, 60, 352-372.