A “Buffer” or “Boomerang?”— The Role of Corporate Reputation in Bad Times

Abstract

Is a favorable prior reputation antibiotics or a hemlock cup in times of organizational crisis? To answer this question, the current study casts light on the contextual cues of crises by applying Brown and Dacin’s (1997) concepts of CA (corporate ability) and CSR (corporate social responsibility) and examines how the cues work in different crisis situations and affect the valence of reputation effects. Drawing on the expectancy violations (EV) theory and the cognitive dissonance perspectives, this study opens the door to reconciling contradicting research findings in literature and provides clues to why and when a good reputation yields buffering or boomerang effects in bad times.

Keywords

corporate reputation reputation management crisis management crisis communication corporate ability crisis corporate social responsibility crisis confirmatory biases the expectancy violations theory cognitive dissonance theory

Arguably one of the prime casualties of a corporate crisis is reputation of the firm confronting the crisis. Needless to say, tangible harms caused by a crisis such as property destruction and financial losses pose great threats to the organization. However, even without these inflicted tangible damages, the crisis can still create a severe blow by changing key stakeholders’ perceptions about the organization. In this sense, Fombrun and van Riel (2004) argue that “the value of a corporate reputation is magnified” during crises (pp. 34-35).

Not only does corporate reputation fall apart in crises but it also plays a role of moderating the extent to which crises inflict damaging effects on the firm. For instance, Jones, Jones, and Little’s (2000) analysis of share price fluctuations following the stock market crash in 1989 revealed that companies with good reputations suffered significantly less declines in market value compared to those without positive reputational standing. The value of a good reputation in times of crisis has also been empirically replicated in laboratories (e.g., Ahluwalia, Burnkrant, & Unnava, 2000; Coombs & Holladay, 2006; Dawar & Pillutla, 2000; Grunwald & Hempelmann, 2011; Klein & Dawar, 2004; Siomkos & Kurzbard, 1994; Siomkos & Shrivastava, 1993).

While reputation is hailed as organizations’ “white knight” during bad times, some research raises concern about the possibility that a good reputation will backfire and inflict even more severe damage to firms, a consequence we call a “boomerang effect.” Researchers in this standing argue that, due to elevated public expectations, firms of good repute may shoulder an extra burden of paying higher costs when confronted with a crisis than those that have a poor reputation or no name (e.g., Dean, 2004; Grunwald & Hempelmann, 2011; Lyon & Cameron, 2004; Rhee & Haunschild, 2006).

Thus the question is “Will a good reputation function as a buffer or boomerang in times of crisis?” Or will reputation have both consequences? If reputation truly has a Janus’ face, yielding both buffer and boomerang effects, then which face will show up in what condition? Tackling the solving of this puzzle, the current study explores conditional cues that may reconcile the conflicting propositions and research findings. In particular, this study sheds light on the type of crisis as a contextual cue that can determine which side of the Janus’ face will stand out. To accomplish this, the current article puts forth a set of hypotheses based on literature, followed by presentation of empirical evidence and interpretation of findings.

Literature Review

Motivational Approaches to Reputation Effects

The concept of corporate reputation, or “a cognitive representation of a company’s actions and results that crystallizes the firm’s ability to deliver valued outcomes to its stakeholders” (Fombrun, Gardberg, & Sever, 2000, p. 87), garners keen attention from both academia and industry because of its value as a firm’s competitive strategic resource (Hall, 1992). The value of reputation shines not only as a strategic resource in the market but also as a protector of firms during bad times. As Bennett and Gabriel (2001) stated, “a well-managed and carefully nurtured corporate reputation can be stored over time to the extent that banked goodwill cushions the adverse consequences of bad publicity” (p. 390). As mentioned above, however, reputation is also viewed as a “double-edged sword,” which may burden firms of fame with liability to pay extra costs in order to keep public’s favors.

One stream of approaches to the buffering or boomerang effects of reputation focuses on stakeholders’ motivations to engage in cognitive processing of negative information during a crisis. In particular, the buffering effect of reputation is explained by consumers’ motivations to keep internal cognitive consistency. Taking its root in Festinger’s (1957) cognitive dissonance theory, this cognitive approach suggests that inconsistent pieces of information make people uneasy, therefore motivating them to reduce the cognitive inconsistency—that is, cognitive dissonance. The reduction of dissonance is accomplished by selectively paying attention to information that is consistent with previously held beliefs and weighing unequal values on different pieces of information. This mental strategy that people adopt to eliminate their cognitive dissonance and keep mental equilibrium is called confirmatory bias (Dawar & Pillutla, 2000; Dean, 2004; Dowling, 2004; Grunwald & Hempelmann, 2011).

According to the perspective of confirmatory bias, when a crisis imparts negative information about the firm of high repute, consumers experience cognitive dissonances due to the new information conflicting with their prior beliefs. Then, in order to relieve cognitive dissonance, they go through a biased cognitive processing in such a way to confirm their prior beliefs and expectations. As a result, negative information related to the crisis will likely be discounted or dismissed (Dawar & Pillutla, 2000, pp. 217, 218; Dowling, 2004, pp. 23-24). On the other hand, when a firm of no name or poor reputation is confronted with a crisis, the process of cognitive dissonance may not occur, therefore generating no confirmatory bias effect (Grunwald & Hempelmann, 2011, p. 267).

In contrast to the cognitive dissonance-based perspectives, which maintain that people see what they want to see, the expectancy violations (EV) theory argues that people actually do not completely ignore or discount the disconfirming information. Rather, preinteraction expectancy held by individuals is likely to be juxtaposed with behaviors of target objects (Burgoon & LePoire, 1993, pp. 90-91). According to EV theory, the target’s violation of the expectancy functions as a motivational trigger for cognitive processing, thus influencing the postinteraction evaluation of the target in such a way that “positive and negative violations (disconfirmation) lead to more positive and negative interaction outcomes respectively than does conformity to expectations” (Burgoon & LePoire, 1993, p. 69).

The EV theory originally was developed to explain interpersonal communication phenomena. Nevertheless, this theory is applicable to stakeholders’ interactions with organizations because people tend to anthropomorphize organizations and view them as conscious social actors or “whole” rather than social aggregates or collectivities (Davies, Chun, da Silva, & Roper, 2001, 2004; Dowling, 2001; Love & Kraatz, 2009). Hence stakeholders view organizations as exchange partners that possess character traits such as trustworthiness and reliability, and evaluate them based on such underlying character elements (Love & Kraatz, 2009, p. 316). Simply put, EV theory predicts that negative information about an organization with good repute will be punished even more harshly for its violation of stakeholders’ high-level expectancies, compared to firms with a neutral or unfavorable reputation that arouse only lower level expectancies for stakeholders.

As reviewed above, both the cognitive dissonance and EV theories provide positive intuitive insight into the mechanism of how prior reputation plays a role in stakeholders’ evaluations of firms in times of crisis. Nonetheless, their apparently contradictory predictions lead researchers to a conundrum: Which theory provides a better explanation of reality?

Cue-Diagnosticity Approach and Cognitive Cue-Processing

Attempting to compare cognitive dissonance and EV theories on the same plane and find which one has better predictive power may be too simplistic and monolithic a view from the beginning. Rather, isn’t it possible that people take different cues from differing pieces of information and trigger divergent mechanisms of cognitive processing? If this is the case, how and why is it that one mechanism—cognitive dissonance or EV—is activated, whereas the other remains dormant?

With regard to this question, Love and Kraatz (2009) provided one clue by contending that people may use different logic systems when evaluating firms depending on the context of assessment, which includes a firm’s organizational characteristics (e.g., trustworthiness and reliability), its symbolic conformity with cultural expectations; or its technical efficacy (e.g., financial performance; p. 318). These propositions imply that stakeholders will employ different cognitive mechanisms contingent on the context in which the evaluation task is deemed more relevant and meaningful. Pertinent to this notion, it is worthwhile to refer to the diagnosticity approach to cognitive biases. Skowronski and Carston (1987) presented a good summary of this approach:

The diagnosticity approach . . . views impression formation as essentially a categorization process. People use available cues, such as an actor’s behaviors, to assign that actor to one or more trait categories. Individual cues generally contribute to categorization probabilistically: For example, someone who once stole money probably belongs in the dishonest category . . . Cues that strongly suggest one categorization (e.g., honest) over alternative categorizations (e.g., dishonest) are said to be diagnostic . . . When an actor is described by multiple cues, categorization should be determined primarily by those cues that lead to the most confident categorizations, that is, by the more diagnostic cues. (p. 689)

According to this diagnosticity approach, diagnostic cues that are perceived as less ambiguous or more informative receive more weight—thus have greater impact—in impression formation than nondiagnostic cues (Skowronski & Carlston, 1987, pp. 689-690). A consequence of the categorization process of diagnostic cues is judgmental bias. Specifically, in a situation where the valence of information is mixed (i.e., both positive and negative information is available), negativity biases occur when negative cues are deemed more diagnostic and positive cues are discounted or ignored. Likewise, positivity biases occur when positive cues are viewed as more diagnostic (Skowronski & Carlston, 1987, p. 690).

Further extending the notion of cue diagnosticity and consequent biases, Skowronski and Carlston (1987) differentiated ability and morality categorizations and connected them with different biases:

With morality categories, positive behaviors may be attributed to many factors, including conformity (Jones & Davis, 1965) and ingratiation (Jones, 1964), whereas negative behaviors are more clearly indicative of morality. Consequently, morality categories are defined more in terms of negative than positive performances, and negative performances should be perceived as more diagnostic. On the other hand, with ability categories, negative performance (failure) may be attributed to many factors, including fatigue and lack of motivation (Anderson & Butzin, 1974; Kun, 1977; Surber, 1984), whereas positive performance (success) is more clearly indicative of ability (cf. Heider, 1958). Consequently, ability categories are defined more in terms of positive than negative performances, and positive performances should be perceived as more diagnostic. (p. 690)

The ability and morality categorizations were later adopted by Wojciszke (1994, 1997, 2005), who successfully provided empirical evidence for the connection between categorizations and their consequent biases. Using the labels of M-trait (for morality categorization) and C-trait (for ability categorization), Wojciszke (2005) demonstrated that a negativity bias is typical for situations where the information on target persons pertains solely or partially to M-trait, therefore negative information is more decisive than positive; the opposite will be true when the information is pertinent solely to C-trait, thereby yielding a positivity bias (p. 61). In the realm of organizational reputation, Pfarrer, Pollock, and Rindova (2010) supported a positivity bias in stakeholders’ evaluation of firms’ ability trait. The authors claimed that “high reputations provide positive analytical frames about firms’ demonstrated ability to deliver value,” thus reducing the diagnosticity of negative information (p. 1136).¹

Based on the literature reviewed, it is assumable that, confronted with ambivalent information in a crisis situation, stakeholders will take the crisis context as a cue and undergo a biased cognitive processing that selectively focuses on either positive or negative information and weigh one more than the other. Specifically, if the crisis calls into question a firm’s ability to perform, the firm’s good performance in the past may serve as a “firm-specific interpretive frame” associated with positive expectations of the firm’s ability to generate values for the future (Pfarrer et al., 2010, p. 1136). Within this frame, stakeholders will likely ignore or discount seemingly less diagnostic negative information in order to keep cognitive consistency for peace of mind, therefore generating the so-called buffering effect of reputation.

On the other hand, if a crisis involves issues of morality or integrity, or violation of other socially approved norms, negative information, which appears more diagnostic for understanding the situation, will hardly be ignored. Rather, the negative information will send an alert to stakeholders whose expectancy about norms is violated. As a consequence, people will likely “punish” the violator by giving the culprit a bad evaluation, as the EV theory predicts. The higher the expectancy in which the target firm was held by stakeholders, the greater the deterioration in their attitudes toward the firm is expected, resulting in backfiring, the so-called boomerang effect of reputation.

Contextual Cues: Corporate Ability Versus Corporate Social Responsibility Cues

To summarize the literature reviewed above, it is suggested that the categorization of ability and morality domains will provide diagnostic contextual cues, which in turn should initiate divergent cognitive processing among stakeholders, thereby resulting in different reputation effects (i.e., buffering or boomerang effects). Extending this notion, the current study proposes to examine reputation effects in time of crisis by using the two types of contextual cues: corporate ability (CA) and corporate social responsibility (CSR). CA and CSR were introduced in the reputation literature by Brown and Dacin (1997). Proposing two types of corporate associations (i.e., what a person knows about a company), the authors defined corporate ability as “expertise in producing and delivering product and/or service offerings” and corporate social responsibility as the “character of the company, usually with regard to important societal issues” (Brown & Dacin, 1997, p. 70).

Extending Brown and Dacin’s (1997) notion of the two-dimensional system to the crisis communication field, the present study suggests two types of reputational crisis: corporate ability (CA) crisis and corporate social responsibility (CSR) crisis. The current study defines CA crisis as a critical event that primarily affects corporate reputation associated with expertise of product and service, technological innovation, and industry leadership; and CSR crisis as a major event that poses a threat to reputation associated with norms and values cherished by society, and socially expected obligations. The social obligations may involve environmental friendliness, commitment to diversity in employment, community involvement, and corporate philanthropic activities.

Table 1 summarizes the empirical studies in literature on reputational effects using the frame of CA and CSR reputation/crisis. As shown in the summary table, studies that manipulated CA crises dominantly supported the buffering effects of a favorable prior reputation, whereas those that supported the boomerang effect of reputation do not clearly show a consistent pattern.

Table 1.

Summary of Empirical Research on the Role of Reputation in Crisis.

	Author(s)	Manipulated crisis type	Manipulated reputation type	Target stakeholders	Outcome measures
Supporting a buffering effect of reputation	Ahluwalia, Burnkrant, and Unnava (2000)	CA	Global (Commitment)	Consumers	Attitude to brand
	Coombs and Holladay (2006)	CA	Global (Overall reputation)	Consumers	Organizational reputation
	Dawar and Pillutla (2000)	CA	CA	Consumers	Perceived brand quality; attitude to brand, brand trust, purchase likelihood
	Grunwald and Hempelmann (2011)	CA	CA	Consumers	Perceived crisis severity & responsibility
	Klein and Dawar (2004)	CA	CSR	Consumers	Blame, brand evaluations, buying intention
	Siomkos and Kurzbard (1994)	CA	Global	Consumers	Perceived danger, purchase intention
Supporting a boomerang effect of reputation	Rhee and Haunschild (2006) ^a	CA	Global	(Investors)	Stock market response
	Dean (2004) ^b	CA	CSR	Consumers	Attitude to brand
	Lyon and Cameron (2004)**	CSR	CSR	Consumers	Attitude to company, behavioral intention, source credibility

This study is based on market data analysis, not on lab experiment.

It should be noted that the studies of Dean (2004), and Lyon and Cameron (2004) did not test the direct impact of prior reputation on stakeholders’ evaluation; they examined the interplay between prior reputation and crisis responses of firms. Their findings demonstrated that a firm with a good reputation was inflicted more serious damage than a firm with a poor reputation when both firms responded “inappropriately” to crises (i.e., denying guilt). Therefore, their studies suggested the liability that organizations of good reputation take in responding to crises.

Based on the literature reviewed, the current study proposes a set of hypotheses and research questions:

Hypothesis 1a (1a): In CA crises, favorable prior reputation will result in a buffering effect.

Hypothesis 1b (1b): In CSR crises, favorable prior reputation will result in a boomerang effect.

Research Question 1(RQ1): Will the contextual cues of crisis (i.e., CA or CSR crisis) interact with reputation cues (i.e., CA or CSR reputation)?

Meanwhile, the construct of reputation is often conceived as an attribute that is built or accumulated over time (e.g., Caruana, 1997; Hall, 1993). For instance, Fombrun and Van Riel (1997) define reputation as a “subjective, collective assessment of its trustworthiness and reliability which . . . summarizes assessments of past performance by diverse evaluators” (p. 10). Therefore, the impact of the time dimension of reputation will also be tapped:

Research Question 2 (RQ2): Will the length of formation of prior reputation interact with reputation effects?

Method

Study Overview

The current study utilized a 2 (CA vs. CSR crisis cues) × 2 (CA vs. CSR reputation cues) × 3 (time point of exposure to crisis information) mixed design, where the responses of every participant randomly assigned to one of the12 experimental conditions were measured over four sessions with a week of time lag between sessions. The time point of exposure to information about a crisis was manipulated in a way that participants read crisis news either in the first (the Negative-Positive-Positive-Positive cue condition), second (the Positive-Negative-Positive-Positive cue condition), or final session (the Positive-Positive-Positive-Negative cue condition), depending on their experimental membership. All participants were randomly assigned to each experimental condition.

Participants and the Stimulus Product Category

Research participants were initially recruited from a population of undergraduate students at a large public university. In a variant of snowball sampling, each participant was asked to recruit a friend. In addition to extra credits, incentives were provided by random drawings of 10 names from the participant pool, each of whom received US$30. IRB approval was obtained prior to data collection. Ages of the 550 participants who completed all sessions ranged from 18 to 35, with a mean age of 21 years (SD = 2.6). In order to prevent potential introduction of an error stemming from the wide range of ages of participants, the test for detecting outliers was conducted by calculating Mahalanobis distance (D²) statistic. As the result, a total of 23 outliers (2 participants aged over 25, and 21 participants aged under 24) with an excessive Mahalanobis D² value were deleted from the sample pool. Of the total 527 participants retained after deleting outliers, 380 (72.1%) were female and 140 (26.6%) were male. Gender of 7 participants was unverified. Majority of participants were White (85%), with a mix of African Americans (6.1%), Hispanics (1.3%), Asians (4.2%), and others (2.5%).

In order to strengthen the external validity of the study, the current experiments used online casual game business as the stimulus manipulation. According to the Nielsen Television Index for the fourth quarter of 2006, two thirds of all men aged from 18 to 34 played video games in their homes, while about 60% of women in the same age range accessed video games (“The State of the Console,” 2007). In particular, in the online casual game business, the majority (63%) are female gamers (Tinney, 2005). Therefore, the gaming industry as the topic of manipulation has relevance to participants in the experiment.

Manipulations

For the experimental treatments, a fictitious online casual game developing company (GameNetworks, Inc.) was used. By using an imaginary company, potential compounding effects stemming from prior attitudes toward and relationships with the target objects were controlled. In order to increase believability of the manipulations, participants were instructed that all information was real and extracted from online news articles. Fombrun and Shanley (1990) claimed that publics use information about firms’ activities, which could come either directly from firms themselves or indirectly from media, and assess the firms’ successes or failures from their individual interpretations of the information (p. 234). Also, from the signaling perspective, communication on information about a firm’s market actions and characteristics is thought to shape the firm’s reputation as it help stakeholders form opinions about its ability to create value for them (Basdeo, Smith, Grimm, Rindova, & Derfus, 2006, p. 1206). Therefore, it is assumed respondents will use the treatments as a signal of reputation of the target firm of the experiment, thereby forming their own individual opinions and beliefs about the target.

In experimental studies researchers often manipulate reputation by introducing scenarios that emphasize the organizations’ outstanding performances, prizewinning records and/or rankings in the relevant fields (e.g., Brown & Dacin, 1997; Grunwald & Hempelmann, 2011; Klein & Dawar, 2004); or donations and other philanthropic activities (e.g., Dean, 2004). Following suit, this study developed stories that showed a history of positive activities for manipulation of reputation. For the CA reputation, news stories covered business-wise recognition for the successful career path of the firm’s CEO, his outstanding entrepreneurship, leadership, and prizewinning history for managerial excellence. For the CSR reputation, news articles about the CEO’s philanthropic activities were presented. Corporate philanthropy fits with Carroll’s (1979) discretionary category of social responsibilities and is viewed as an outlet for top managers’ concern for the welfare of others in community (Choi & Wang, 2007, pp. 347-348).

Utilizing reputation of leadership as the means of promoting corporate reputation has been justified by the notion that a CEO literally and symbolically becomes the organization itself to stakeholders (Bromley, 2001; Grunig, 1993; Pincus, Rayfield, & Cozzens, 1991). In addition, manipulating leadership reputation has a benefit of avoiding a potential confounding effect stemming from interaction between crisis and organizational reputation types. That is, if the core value of a firm’s reputation is product quality, a product-harm crisis could cause more devastating consequences than other types of crises (See Dawar & Lei, 2009; Rhee & Valdez, 2009), the effect which is outside the scope of the present study. Three stories for each type of reputation were developed. In order to avoid presentation order effects, participants of an experimental condition were divided into three subgroups, to which different stories were randomly assigned.

For manipulation of crises, two crisis scenarios were developed. In the CA-crisis scenario, the website of GameNetworks was breached by hackers and individual information of game users (including credit card information) was stolen. Also, the firm’s server system was disrupted twice due to system failure a month before the outbreak. In this story, the firm was described as a victim, implying the firm’s lack of malicious intentionality, thus restricting potential for negative attribution to the firm’s integrity.² In the CSR-crisis scenario, the firm was accused of selling its customers’ private information without permission to other marketing companies. The firm was also charged with manipulating its accounting books to conceal numerous illegal sales, thus evading taxes. To minimize inference of the firm’s performance based on its poor CSR status, the firm’s superior performance history was emphasized by using such words as “a rising star” and “unprecedented growth.”

All of the stories used for manipulation were written based on actual media reports with a view to obtain ecological validity and increase believability of the stories. In addition, prior to the main test, a series of pilot tests were conducted with 58 undergraduate/graduate students, who were not part of the sampling pool for the main study. The pilot test results showed that the treatment stories were manipulated successfully as the between-condition differences were statistically significant, while the within-condition differences were negligible.

Experiment Procedures

All four experimental sessions were conducted online. All instruments including stimulus stories were uploaded to the server with different links determined by the experimental design. All survey sites were open only 3 days a week to control the flow of participation. After being randomly assigned, each participant received a private invitation e-mail, which contained the link to the survey site. The instrument started with an instruction that read,

The following information is about a real, well-known online game developing company. For the purposes of this study we call the company GameNetworks, and the company’s CEO Bob Glase. Your answers are completely confidential so be as frank as you wish. This is not a test—your opinion is the only right answer.

And from the second session, another paragraph, which was aimed to remind respondents of the previous story, was added following the instruction. After reading treatment materials, participants responded to a set of questions. After completion of all experimental sessions, participants were debriefed.

Dependent Measures

The scope of the current study is consumer stakeholders. Their responses to the treatment materials were tapped by their overall attitudes toward the firm and other evaluative measures including satisfaction, trust, commitment, and loyalty. The measures of satisfaction, trust, and commitment were adopted from Hon and Grunig’s (1999) organization-public relationships (OPR) measures. Trust is conceptualized as “one party’s level of confidence in and willingness to open oneself to the other party,” while satisfaction is defined as “the extent to which one party feels favorably toward the other because positive expectations about the relationships are reinforced” and commitment as “extent to which one party believes and feels that the relationship is worth spending energy to maintain and promote” (Hon & Grunig, 1999, pp. 19-20). Participants’ overall attitudes toward the firm were measured with two items. All outcome measures were quantified on 7-point Likert-type scale with 1 being very strongly disagree and 7 being very strongly agree. Several items were flipped in scale before being collapsed into each of the outcome variables, therefore rearranging the scales created from 1 (unfavorable) to 7 (favorable). Cronbach’s alpha ranged from .816 to .760 for satisfaction; from .869 to .769 for trust; from .891 to .919 for commitment; and from .917 to .923 for attitudes to firm. For loyalty, which is defined as “a feeling of attachment to or affection for a company’s people, products, or services” (T. O. Jones & Sasser, 1995, p. 94), a scale was created by adapting De Ruyter, Wetzels and Bloemer’s (1998) scale. Four items were used and each of them was quantified on a 7-point bipolar scale, ranging from 1 (unfavorable) to 7 (favorable). Cronbach’s alpha for the scale ranged from .917 to .925.

As the internal locus of crisis was manipulated as a determinant of crisis type in the current study, perceived crisis responsibility and controllability were also measured in order to control potential confounding effects. Respondents’ perceived crisis controllability was measured by asking straightforwardly (i.e., “With regard to intention and power to control the situation, how likely do you think it was that this company could have prevented this incident from occurring?”). The measures of crisis responsibility were developed by adopting Klein and Dawar’s (2004) items. Each item was measured on 7-point Likert-type scale. Cronbach’s alpha for the responsibility scale ranged from .900 for the CA crisis condition and .956 for the CSR crisis condition. All measures are presented in Table A1 of the Appendix.

Manipulation and Random Assignment Check

In order to see whether manipulations were successfully accomplished, scales to measure CA and CSR corporate reputation were developed by adopting Fombrun et al.’s (2000) Reputation Quotient (RQ). All items were quantified on a 7-point Likert-type scale with 1 being very strongly disagree and 7 being very strongly agree. Cronbach’s alpha for CA reputation ranged from .843 to .886; and from .907 to .908 for CSR reputation. These items are shown in Table A1 of the Appendix.

For the manipulation check, a series of tests was run using SPSS v. 18. For the tests, 159 subjects who were exposed only to positive reputational information were used. The mean comparison between CA and CSR reputation conditions showed that, as the manipulations intended, the CA-reputation group yielded a higher evaluation for the firm’s CA reputation (n = 76, M = 5.0, SD = .85), compared to the CSR-reputation group (n = 83, M = 4.5, SD = .68), and this between-group difference was significant at the .01 level, t(157) = 4.236, p = .000. The mean for the CSR-reputation condition was also higher (n = 81, M = 4.7, SD = .87) than for the CA-reputation condition (n = 76, M = 4.3, SD = .71), the test result being significant at the .01 level, t(155) = 2.959, p = .004. When comparing to the neutral point on the scale (M = 4.0) by using a one-sample t test, the mean of the CA-reputation group (M = 5.0) was significantly higher, t(75) = 10.034, p = .000; and the mean of the CSR-reputation condition (M = 4.7) was also higher, t(80) = 7.218, p = .000. To inspect the crisis manipulations, 197 sample participants who were exposed only to the negative information of crisis were used. The independent t test showed that the CA crisis condition had a lower CA reputation score (n = 99, M = 3.66, SD = 1.048), compared to the CSR reputation score (n = 98, M = 3.79, SD = .781), while the CSR crisis condition caused more damage to the CSR reputation (n = 99, M = 3.07, SD = .895) than to the CA reputation (n = 94, M = 4.01, SD = 1.025). These differences were statistically significant at the .05 level, t(191) = −2.347, p = .020 for CA reputation; t(195) = 5.956, p = .000 for CSR reputation.

The tests of random assignment were conducted by comparing mean differences between the experimental groups that belong to the same type of manipulation cluster. A series of ANOVA tests was run and none of the tests were statistically significant, suggesting successful accomplishment of random assignments.

Results

For the tests for hypotheses and research questions proposed, SPSS v.18 was used for analysis. To test the effects of prior reputation (H1a and H1b), as well as the interaction between reputation and crisis types (RQ1), subjects were selected from the NPPP (Negative [crisis cues]-Positive [reputation cues]-Positive-Positive) and the PPPN (Positive-Positive-Positive-Negative) conditions. Their responses were measured immediately following their exposure to the crisis news (i.e., responses at the N-time point). In this way, outcome measures of the PPPN condition were supposed to reflect prior reputation effects, whereas the NPPP condition served as a control group as its subjects were exposed to only crisis information without any reputational information at the time point of measure. For the test, a series of 2 (NPPP-control group vs. PPPN-effect group) × 2 (CA vs. CSR reputation) factorial analyses was conducted respectively for the CA and CSR crisis situations.

The tests detected no sign of interactions between the crisis and reputation types, nor did reputation types have any main effects. However, the mean comparisons between the different time points conditions (i.e., NPPP vs. PPPN groups) yielded significant main effects for some outcome variables. In the CA crisis situation, the mean scores of the effect group (PPPN group) were higher than those of the control group (NPPP group) on satisfaction, commitment, and loyalty. These mean differences were statistically significant at the .05 level, F(1, 176) = 4.678, p = .032, effect size η²_p = .026 for satisfaction; F(1, 175) = 9.154, p = .003, η²_p = .05 for commitment; and F(1, 176) = 5.952, p = .016, η²_p = .033 for loyalty. Therefore, H1a, which posited a buffering effect of good prior reputation in the CA crisis context, was partially supported. In the CSR crisis context, only overall attitude to the firm had a main effect, where the mean of the control group was significantly higher than that of the effect group at the .01 level, F(1, 177) =7.243, p = .008, η²_p = .039. Therefore, H1b, which hypothesized a boomerang effect of favorable prior reputation in the CSR crisis context, was marginally supported. After controlling crisis controllability and responsibility as covariates, the test result did not differ considerably from the previous results except that the grouping effect in the CA crisis condition was also significant on trust, F(1, 171) =4.393, p = .038, η²_p = .025. Meanwhile, evidence that supports RQ1 was not found. The descriptive statistics are presented in Table 2.

Table 2.

Prior-Reputation Effects.

		CA Crisis						CSR Crisis
		N (Control group)			P-N (Effect group)			N (Control group)			P-N (Effect group)
DVs	Reputation type	n	M	SD	n	M	SD	n	M	SD	n	M	SD
Attitude to Company	CA	49	3.4	.99	41	3.3	.80	50	3.0	1.04	39	2.4	.88
	CSR	50	3.2	1.02	41	3.5	.88	49	3.0	1.01	43	2.7	1.03
	Total	99	3.3	1.0	82	3.4	.84	99	3.0	1.02	82	2.6**	.97
Satisfaction	CA	49	3.4	.99	41	3.6	.79	50	3.5	.77	39	3.1	.82
	CSR	49	3.3	.81	41	3.6	.73	50	3.2	.93	43	3.3	.92
	Total	98	3.3	.90	82	3.6*	.76	100	3.4	.87	82	3.2	.88
Trust	CA	49	3.7	.96	40	3.8	.62	49	3.2	.92	39	2.7	.85
	CSR	50	3.6	.77	40	3.8	.75	50	3.0	.98	43	3.0	.90
	Total	99	3.7	.87	80	3.8	.69	99	3.1	.95	82	2.8	.89
Commitment	CA	49	2.6	1.11	41	2.9	.92	50	2.8	1.01	39	2.7	.96
	CSR	49	2.5	1.01	40	3.1	.92	50	2.6	1.01	43	2.7	1.01
	Total	98	2.6	1.06	81	3.6**	.92	100	2.7	1.01	82	2.7	.98
Loyalty	CA	48	2.9	1.2	41	3.2	.93	50	2.9	1.06	39	2.5	1.04
	CSR	50	2.8	.99	41	3.3	.81	49	2.6	1.14	43	2.8	1.03
	Total	98	2.9	1.1	82	3.2*	.87	99	2.8	1.1	82	2.6	1.04

Significant at the .05 level.

Significant at the .01 level.

As seen in Table 2, standard deviation numbers appear somewhat large for the 7-point scale. Intuitively, it was deemed plausible that differences in individuals’ scores in a single conditional group might offset each other in the process of averaging the measures. Therefore, in order to further test H1a and H1b, subjects were selected from the PPPN condition and were divided into high- and low-prior reputation subgroups based on their initial attitude-to-firm scores that were measured at the first time point. The mean score of 4 was used as the cut point and subjects whose scores were between 4 and 5 were discarded in order to maximize differences between the two subgroups. As a result, the low-prior reputation group had 54 participants, whose average attitude-to-firm score was 3.7 (SD = .47); the high-prior reputation group had 78 participants, whose average attitude score was 5.3 (SD = .52). Then their responses to crisis news measured at the fourth time point were compared by utilizing 2 (high- vs. low-prior reputation groups) × 2 (CA vs. CSR reputation) factorial analyses.

The test result showed no significant interaction effects between the factors or the significant effect of reputation type. However, the mean comparison between high- and low-reputation groups yielded significant main effects of prior reputation. In the CSR crisis situation, all outcome measures indicated significant grouping effects, F(1, 61) = 4.021, p = .049, effect size η²_p = .062 for satisfaction; F(1, 61) = 9.410, p = .003, η²_p = .134 for trust; F(1, 61) = 7.183, p = .009, η²_p = .105 for commitment; and F(1, 61) = 10.021, p = .002, η²_p = .141 for loyalty. By contrast, two groups did not show any significant differences on the outcome measures in the CA crisis context. For all outcome measures, the mean scores of the high-reputation group were significantly lower than those of the low-reputation group, meaning that participants of the high-reputation group were more negatively influenced by the crisis than those in the low-reputation group. This result clearly supported the boomerang effect of a favorable prior reputation in the CSR crisis context, therefore supporting H1b. The results did not change after controlling for the crisis factors. The summary of descriptive statistics is presented in Table 3.

Table 3.

Prior-Reputation Effects Between High- and Low-Reputation Groups in the CSR Crisis.

	Low-prior reputation group			High-prior reputation group
	n	M	SD	n	M	SD
Satisfaction*	27	3.5	.54	38	3.0	.99
Trust**	27	3.3	.92	38	2.6	.78
Commitment**	27	3.1	.96	38	2.5	.88
Loyalty**	27	3.1	.99	38	2.3	1.01

Significant at the .05 level.

Significant at the .01 level.

Finally, in order to test the long-term effect of prior reputation (RQ2), the PNPP and the PPPN conditions were compared based on the participants’ responses to crisis news measured at the second (for the PNPP condition) and fourth (for the PPPN condition) sessions respectively. The PPPN condition, where participants were exposed to positive reputational information 3 weeks before their exposure to crisis news, is designed to reflect reputation of a long history, while the PNPP condition represents reputation of a short history. For this test, 2 (PNPP-short history group vs. PPPN-long history condition) × 2 (CA vs. CSR reputation) factorial analyses were conducted respectively for the CA and CSR crisis situations.

As the result of the test, no significant main effects or interaction effects were found in the CA crisis context. In the CSR crisis situation, the outcome variables of trust and loyalty showed significant main effects, F(1, 157) = 8.617, p = .004, effect size η²_p = .052 for trust; and F(1, 153) = 4.883, p = .029, η²_p = .031 for loyalty. For both outcome measures, the mean scores for the PNPP (short history) condition (n = 79, M = 3.2, SD = .83 for trust; n = 75, M = 3.0, SD = .92 for loyalty) were higher than those for the PPPN (long history) condition (n = 82, M = 2.8, SD = .88 for trust; n = 82, M = 2.6, SD = 1.04 for loyalty). However, after suppressing the effects of crisis responsibility and controllability, the main effect was significant only on trust, F(1, 155) =6.471, p = .012, η²_p = .040, but not on loyalty. This test result partially supported the boomerang effect of prior reputation. That is, the longer history a prior reputation had, the greater the damage that was inflicted to individuals’ evaluations of the firm in terms of trust. The summary of hypothesis tests is presented in Table 4. The interpretations of these findings, along with implications and limitations, are presented in the following section.

Table 4.

Summary of Hypothesis Test Results.

	Test 1 [N vs. PPPN]		Test 2 [High vs. Low prior reputation (PPPN only)]		Test 3 [PN vs. PPPN]
Outcome measures	CA crisis	CSR Crisis	CA crisis	CSR Crisis	CA crisis	CSR Crisis
Overall attitude	-	M_N > M_PPPN	-	-	-	-
Satisfaction	M_N < M_PPPN	-	-	M_Lo > M_Hi	-	-
Trust	M_N < M_PPPN^a	-	-	M_Lo > M_Hi		M_PN > M_PPPN
Commitment	M_N < M_PPPN	-	-	M_Lo > M_Hi	-	-
Loyalty	M_N < M_PPPN	-	-	M_Lo > M_Hi	-	M_PN > M_PPPN^b
Reputation effects
Buffering effects	Yes	-	-	-	-	-
Boomerang effects	-	Yes (marginally)	-	Yes	-	Yes

Notation: M = group mean; N = Negative (crisis) information; P = Positive (reputation) information; PN = Positive-Negative time order; PPPN = Positive-Positive-Positive-Negative time order.

^aSignificant after controlling over the crisis factors.

***

^bNonsignificant after controlling over the crisis factors.

Discussion

Is a favorable prior reputation antibiotics or a hemlock cup in times of organizational crisis? Amidst the majority of literature that advances the benefits of favorable reputation during bad times, another research stream argues against such an optimistic view; there are opposite empirical findings although the optimistic voice is dominant. Motivated to resolve this disagreement, the current study drew attention to the contextual cues of crises and examined, employing an experimental method, how the cues work in different crisis situations and change the valence of reputation effects.

Toward this purpose, this study applied Brown and Dacin’s (1997) concepts of CA (corporate ability)-and CSR (corporate social responsibility) to the organizational crisis domain and demonstrated that the CA-versus-CSR categorization of crisis cues presents a boundary condition that determines divergent reputation effects. That is, in the CA crisis context, positive information of a favorable prior reputation led participants to discount negative information about the crisis. This result indicates that, in the CA crisis context, where positive information becomes more diagnostic and salient, cognitive dissonance theories provide a better explanation and predictive power than any other. By contrast, in the CSR crisis, where morality or integrity of a firm was challenged, its favorable prior reputation backfired as participants lowered their evaluations of the firm with a good reputation. This finding demonstrates that, in the CSR crisis context, the expectancy violations (EV) theory may provide a mechanism that better explains the role of reputation during bad times.

It is not surprising that the EV theory has superior predictive power in the CSR crisis context. In fact, EV theory was developed to explain the role of cultural norms in changing communication behaviors (Burgoon, Le Poire, & Rosenthal, 1995) and postinteraction impressions (Burgoon & LePoire, 1993). It was later applied to other social norm contexts (e.g., Campo, Cameron, Brossard, & Frazer, 2004; Kalman & Rafaeli, 2011; Levine et al., 2000). Levine et al. (2000) that asserted that norms dictate the range of behaviors that are socially acceptable and create expectancies and violations of such expectations, “raise suspicion, create doubts, and ultimately lead to judgments of deceitful content and intent” (p. 123). In the CSR crisis context, socially approved norms are violated, therefore funneling perceivers’ attention to negative cues. These contextual cues then will trigger the EV mechanism. In this situation, favorable prior reputation that has built up higher levels of expectancy for the target firm is likely to backfire such that stakeholders will increase their scrutiny of the firm’s negatively deviant behaviors.

In addition to the major finding of the role of crisis contextual cues in moderating reputation effects, the current study provides intriguing results that warrant further discussion and exploration. One noteworthy finding is the “history effect” of reputation. In literature, reputation is often regarded as the result of a history of positive performance or good relations with stakeholders. For instance, de Castro, Lopez, and Saez (2006) argued that a firm’s present reputation is “accumulated in a historical context and in unique circumstances that surely cannot be repeated” (p. 362). Interestingly, the history effect of reputation backfired in the CSR crisis domain. That is, participants’ longer exposures to favorable reputational information caused greater damage to the target firm. This result can be explained by the EV theory, which posits that higher expectancies may lead to more severe punishment. It is also presumable that the more individuals are exposed to positive information about a firm, the more likely they will be to heighten their expectancies of the firm over time. Therefore, the findings of this study suggest that the EV mechanism includes a trait that builds up over time—even though it ironically functions in an unfavorable way. This finding offers a significant implication: there may be no fortress in prior reputation that can protect a firm with good repute. Even one instance of an unethically appearing tragic event may be enough to destroy years of efforts to build repute as a good citizen.

In the CA crisis context, on the other hand, this study found a significant buffering effect only when the treatment groups (i.e., the experimental groups that were exposed to positive information about the target firm) were compared with the control group (i.e., the experimental group that was exposed only to crisis news). Whether participants were exposed to positive reputational information once or several times made no difference in generating a buffering effect from prior reputation. Based on this finding, it is assumable that once the cognitive dissonance mechanism is activated individuals may compare only the relevance or salience of positive or negative information available at the moment of judgment, and the history factor of reputation may not wield much influence. This finding implies that, in a CA crisis situation, relative new comers are likely to enjoy as much benefit accrued to a good name as old players, once they’ve established a good name as a capable player in the market.

However, it may be a rash conclusion to declare a weak or no history effect of reputation based on a single empirical study. A favorable prior reputation related to corporate ability may influence individuals’ judgment in a different way, for instance, by affecting the degree of acceptance of negative information. That is, if a firm has built up positive reputation as a good value-creator for stakeholders in the market over a long period, people could be more tolerant of the firm’s failure, compared to failures of other firms with a shorter reputational history. Validating this postulation requires further investigation in the future by varying the severity of a crisis as well as manipulating the history factor.

Another intriguing finding of this study is the marginal support found for the boomerang effect in the CSR crisis domain when comparing the treatment group (i.e., the experimental condition where participants were exposed to positive corporate information three times before being exposed to crisis news) to the control group (i.e., the group that was exposed only to crisis news). But after dividing the treatment group into two subgroups based on their attitudes toward the firm (i.e., high-versus-low reputation groups), the test result revealed obvious boomerang effects of prior reputation. That is, reputation effects were marginal when the averaged responses of the entire treatment group were measured, whereas measurements taken after assessment of personal differences clearly demonstrated the existence of a boomerang effect of reputation. This finding suggests that individuals’ personal intrinsic values or beliefs about social norms may play a significant role in evaluating CSR-crisis information. This correlation between personal values and CSR reputation has been previously reported in literature (e.g., Basil & Weber, 2006; Golob, Lah, & Jančič, 2008; Siltaoja, 2006). For instance, Basil and Weber (2006) demonstrated that individuals motivated by egoistic enhancement and personal values made purchases in support of corporate philanthropy, whereas those who were not motivated by such personal values did not view corporate social responsibility as a normative requirement. Silatoja’s (2006) study also suggested that people who have concern about welfare interests expected firms to possess responsibilities other than making money. In sum, this study’s finding draws attention to the importance of considering personal values in examining CSR-related reputation.

Implications

The implication of the present study for the CA crisis domain is straightforward: a well-nurtured organizational reputation will pay off by buttressing firms of good names at bad times. If a firm has not yet developed a good reputation and is confronted with a crisis that casts doubt on its ability to perform in the market, the negative impression will reside in the stakeholders’ minds with no resistance; there’s nothing that can cushion the adverse impact of crisis. However, interpreting this study’s findings needs some caveats. The buffering effect of a good reputation in times of crisis was significant in relative terms—that is, only in comparison with a condition where no reputation has been built. Although not reported here, the same data set showed that the CA crisis inflicted damage to participants’ evaluation of the target firm on all outcome measures in absolute terms. That is, the participants’ attitudes to the target firm were not the same before and after the crisis. Therefore, a favorable reputation cannot be a cure-all for crises. The message is clear: Crises should be prevented at all times and costs.

In the CSR crisis domain, on the other hand, the situation gets complicated as the findings imply that carrying a favorable prior reputation can be a burden for the firm. Does it mean that companies will benefit from keeping a low profile? In fact, Burgoon, Dunbar, and Segrin (2002) noted that research in the sales and marketing arena has urged setting expectations low so that a delivered product or service “positively violate those expectations” (p. 463). In a somewhat different context, Coombs and Holladay’s (2002) study also indicated no benefit from a favorable prior reputation over a neutral reputation but only disadvantage from an unfavorable prior reputation, which the authors called the Velcro effect. However, considering the buffering effect of a favorable reputation in the CA crisis context, keeping a firm from building up a good name does not appear to be a good idea. Considering all sides, more practical advice for organizations will be as follows: Avoid groundless hype that will only create unduly high expectations.

Future Research Agenda

By introducing contextual cues of crisis to research on reputation effects, the current study opened the door to reconciling contradictory research findings and successfully provided clues to why, of the cognitive dissonance and EV mechanisms, one perspective explains certain reputation effects better than the other. This study, therefore, highlights the necessity to specify the context in research involving the effects of reputation and/or crises. Despite its value of contributing to a better understanding of reputation and crisis phenomena, this study has some limitations that deserve further investigation in the future.

First, the type of reputation (i.e., CA-vs.-CSR reputation) did not play a significant role in this study. It can hardly be concluded whether this test result reflects a true reality or not, based on a single study. It may be because the strength of the manipulation might have failed to arouse expected responses from participants. However, all experimental studies employing scenario-based manipulations are not free from the effect of strength of messages or overall context manipulated. That is, if the messages are too weak, the manipulations will not detect any effects; if the messages are too strong, the manipulations will not have any ability to differentiate the effects. Therefore, more studies that adopt different manipulations with a variety of message strength, contextual cues, and other considerations should be conducted until the insight found in the present study is validated with a certain degree of confidence.

Another limitation of this study stems from the nature of experimental methods generally. Since many variables are not included in experimental research in order to isolate the variables of study, its generalizability to the real world is sometimes called into question. For instance, this study’s findings are contradictory to that of Rhee and Haunschild (2006), which showed a boomerang effect in the CA crisis context (i.e., product recalls of automakers). However, their study analyzed real-world stock market data. Therefore, there is a chance that extraneous variables intervened in their study. The authors actually explained that recalls by firms with a good reputation attracted more media attention, therefore imparting quantitatively more negative information than recalls of those that did not have established reputations. This difference warrants further research by considering a variety of potentially moderating factors. Rhee and Haunschild’s (2006) study also differed from other studies presented in Table 1 as it targeted responses of stock investors, whereas most other studies targeted consumer responses. Intuitively, it is questionably argued that consumers would have the same mindset as investors. Empirically, Berens, van Riel, and van Rekom (2007) demonstrated that, depending on the tasks given to participants in the research (e.g., job offer or investing stocks), a firm’s CSR reputation had significantly different influence on participants’ responses. Therefore, future research should also consider variation of target audience and the context of the tasks.

In manipulating crisis scenarios, the present study considered crisis responsibility as one element determining the CA-CSR crisis type. This strategy was justified by prior research, which empirically showed that higher responsibility raised the odds of a crisis being perceived as a CSR crisis, instead of a CA crisis (Sohn and Lariscy, 2014). All the hypotheses in the present study were tested with and without suppressing the impact of the crisis factors, only finding minor differences between the results. However, it is precarious to conclude that high responsibility and internal locus of crisis are the key elements defining the nature of CSR crises. Further research delving into the natures of CA and CSR crises is warranted.

Certainly more questions await future investigations. There are doubtless other conditions not mentioned in this study, which will yield significant differences in results. Considering all other conditions is beyond the scope and capacity of any single study. However, by bringing identifying the boundary conditions that play key roles in determining types of reputation effects, this study fulfilled its task.

Footnotes

Appendix

Table A1

Measures of Outcome Variables

Crisis controllability

1. “With regard to intention and power to control the situation, how likely do you think this company could have prevented this incident from occurring?”

Crisis responsibility

1. “In my opinion, this company is responsible for this incident.”

2. “In my opinion, this company should be held accountable for this incident.”

3. “In my opinion, this incident is the fault of this company.”

CA reputation

1. “This company would offer high quality products and services.”

2. “This company would develop innovative products and services.”

3. “This company seems competent and effective in providing its products and services.”

CSR reputation

1. “This company would support good causes.”

2. “This company would be an environmentally responsible company.”

3. “This company would be honest.”

4. “This company would be sincere and genuine.”

5. “This company would behave ethically.”

Overall attitudes to firm

1. “I have a good feeling about this company.”

2. “I admire and respect this company.”)

Satisfaction

1. “My experience with this company would be excellent.”

2. “Most consumers like me would be unhappy in their interactions with this company.”

3. “Both this company and consumers like me would benefit from the relationship.”

4. “I would feel that this company falls to satisfy the needs of consumers like me.”

Trust

1. “I would feel that this company treats consumers like me fairly and justly.”

2. “This company cannot be relied on to keep its promises to consumers like me.”

3. “I would feel that sound principle guide this company’s behavior.”

4. “I would feel that this company misleads consumers like me.”

Commitment

1. “I would feel a strong sense of belonging to this company.”

2. “I would feel emotionally attached to this company.”

3. “My relationship with this company would be important to me.”

4. “If this company no longer exists, this would be a significant loss for me.”

5. “I would feel a strong sense of identification with this company.”

Loyalty

1. “I would definitely recommend this company to someone who seeks my advice.”

2. “I will consider this company my first choice to buy the products/services I need.”

3. “I will intend to do more business with this company in the next few years.”

4. “I would continue to do business with this company even if its prices increased somewhat.”

Declaration of Conflicting Interests

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

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

YJ Sohn teaches strategic communication and research courses at Florida Tech. Her research interests include reputation management, crisis communication, and organization-public relationships management. She received doctorate degree from the University of Georgia.

Ruthann Weaver Lariscy teaches public relations principles, research, and public opinion at UGA. Her research focuses on political advertising and campaign management, public relations standards and health campaigns. She earned doctorate degree from the University of Missouri.

References

Ahluwalia

Burnkrant

R. E.

Unnava

H. R.

(2000). Consumer response to negative publicity: The moderating role of commitment. Journal of Marketing Research (JMR), 37, 203-214.

Basdeo

D. K.

Smith

K. G.

Grimm

C. M.

Rindova

V. P.

Derfus

P. J.

(2006). The impact of market actions on firm reputation. Strategic Management Journal, 27, 1205-1219.

Basil

D. Z.

Weber

(2006). Values motivation and concern for appearances: The effect of personality traits on responses to corporate social responsibility. International Journal of Nonprofit & Voluntary Sector Marketing, 11(1), 61-72. doi:10.1002/nvsm.38

Bennett

Gabriel

(2001). Corporate reputation, trait covariation and the averaging principle: The case of the UK pensions mis-selling scandal. European Journal of Marketing, 35, 387-413.

Berens

van Riel

C. B. M.

van Rekom

(2007). The CSR-quality trade-off: When can corporate social responsibility and corporate ability compensate each other? Journal of Business Ethics, 74, 233-252.

Bromley

D. B.

(2001). Relationships between personal and corporate reputation. European Journal of Marketing, 35, 316-334.

Brown

T. J.

Dacin

P. A.

(1997). The company and the product: Corporate associations and consumer product responses. Journal of Marketing, 61(1), 68-84.

Burgoon

J. K.

Dunbar

N. E.

Segrin

(2002). Nonverbal influence. In Dillard

J. P.

Pfau

(Eds.), The persuasion handbook: Developments in theory and practice (pp. 445-473). Thousand Oaks, CA: SAGE.

Burgoon

J. K.

LePoire

B. A.

(1993). Effects of communication expectancies, actual communication, and expectancy disconfirmation on evaluations of communicators and their communication behavior. Human Communication Research, 20(1), 67-96. doi:10.1111/j.1468-2958.1993.tb00316.x

10.

Burgoon

J. K.

Le Poire

B. A.

Rosenthal

(1995). Effects of preinteraction expectancies and target communication on perceiver reciprocity and compensation in dyadic interaction. Journal of Experimental Social Psychology, 31, 287-321. doi:10.1006/jesp.1995.1014

11.

Campo

Cameron

K. A.

Brossard

Frazer

M. S.

(2004). Social norms and expectancy violation theories: Assessing the effectiveness of health communication campaigns. Communication Monographs, 71, 448-470. doi:10.1080/0363452042000307498

12.

Carroll

A. B.

(1979). A three-dimensional conceptual model of corporate performance. Academy of Management Review, 4, 497-505.

13.

Caruana

(1997). Corporate reputation: Concept and measurement. Journal of Product & Brand Management, 6(2), 109-118.

14.

Choi

Wang

(2007). The promise of a managerial values approach to corporate philanthropy. Journal of Business Ethics, 75, 345-359. doi:10.1007/s10551-006-9257-4

15.

Coombs

W. T.

Holladay

S. J.

(2002). Helping crisis managers protect reputational assets. Management Communication Quarterly, 16, 165-186.

16.

Coombs

W. T.

Holladay

S. J.

(2006). Unpacking the halo effect: Reputation and crisis management. Journal of Communication Management, 10(2), 123-137. doi:10.1108/13632540610664698

17.

Davies

Chun

da Silva

R. V.

Roper

(2001). The personification metaphor as a measurement approach for corporate reputation. Corporate Reputation Review, 4(2), 113-127.

18.

Davies

Chun

da Silva

R. V.

Roper

(2004). A corporate character scale to assess employee and customer views of organization reputation. Corporate Reputation Review, 7(2), 125-146.

19.

Dawar

Lei

(2009). Brand crises: The roles of brand familiarity and crisis relevance in determining the impact on brand evaluations. Journal of Business Research, 62, 509-516. doi:10.1016/j.jbusres.2008.02.001

20.

Dawar

Pillutla

M. M.

(2000). Impact of product-harm crises on brand equity: The moderating role of consumer expectations. Journal of Marketing Research (JMR), 37, 215-226.

21.

de Castro

G. M.

Lopez

J. E. N.

Saez

P. L.

(2006). Business and social reputation: Exploring the concept and main dimensions of corporate reputation. Journal of Business Ethics, 63, 361-370.

22.

de Ruyter

Wetzels

Bloemer

(1998). On the relationship between perceived service quality, service loyalty and switching costs. International Journal of Service Industry Management, 9, 436-453. doi:10.1108/09564239810238848

23.

Dean

D. H.

(2004). Consumer reaction to negative publicity: Effects of corporate reputation, response, and responsibility for a crisis event. Journal of Business Communication, 41, 192-211.

24.

Dowling

G. R.

(2001). Creating corporate reputations: Identity, image, and performance. New York, NY: Oxford University Press.

25.

Dowling

G. R.

(2004). Corporate reputations: Should you compete on yours? California Management Review, 46(3), 19-36.

26.

Ferrin

D. L.

Kim

P. H.

Cooper

C. D.

Dirks

K. T.

(2007). Silence speaks volumes: The effectiveness of reticence in comparison to apology and denial for responding to integrity- and competence-based trust violations. Journal of Applied Psychology, 92, 893-908. doi:10.1037/0021-9010.92.4.893

27.

Festinger

(1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.

28.

Fombrun

C. J.

Gardberg

N. A.

Sever

J. M.

(2000). The reputation quotient: A multiple stakeholder measure of corporate reputation. Journal of Brand Management, 7, 241-255.

29.

Fombrun

C. J.

Shanley

(1990). What’s in a name? Reputation building and corporate strategy. Academy of Management Journal, 33, 233-258.

30.

Fombrun

C. J.

van Riel

C. B. M.

(1997). The reputational landscape. Corporate Reputation Review, 1(1/2), 5-13.

31.

Fombrun

C. J.

van Riel

C. B. M.

(2004). Fame & fortune: How successful companies build winning reputations. Upper Saddle River, NJ: Prentice Hall.

32.

Golob

Lah

Jančič

(2008). Value orientations and consumer expectations of Corporate Social Responsibility. Journal of Marketing Communications, 14(2), 83-96. doi:10.1080/13527260701856525

33.

Grunig

L. A.

(1993). Image and symbolic leadership: Using focus group research to bridge the gaps. Journal of Public Relations Research, 5(2), 95.

34.

Grunwald

Hempelmann

(2011). Impacts of reputation for quality on perceptions of company responsibility and product-related dangers in times of product-recall and public complaints crises: Results from an empirical investigation. Corporate Reputation Review, 13, 264-283. doi:10.1057/crr.2010.23

35.

Hall

(1992). The strategic analysis of intangible resources. Strategic Management Journal, 13(2), 135-144.

36.

Hall

(1993). A framework linking intangible resources and capabilities to sustainable competitive advantage. Strategic Management Journal, 14, 607-618.

37.

Hon

L. C.

Grunig

J. E.

(1999). Guidelines for measuring relationships in public relations. Retrieved from http://www.instituteforpr.org/files/uploads/1999_MeasuringRelations.pdf

38.

Jones

G. H.

Jones

B. H.

Little

(2000). Reputation as reservoir: Buffering against loss in times of economic crisis. Corporate Reputation Review, 3(1), 21-29.

39.

Jones

T. O.

Sasser

W. E.

(1995). Why satisfied customers defect. Harvard Business Review, 73(6), 88-99.

40.

Kalman

Y. M.

Rafaeli

(2011). Online pauses and silence: Chronemic expectancy violations in written computer-mediated communication. Communication Research, 38(1), 54-69. doi:10.1177/0093650210378229

41.

Kim

P. H.

Dirks

K. T.

Cooper

C. D.

Ferrin

D. L.

(2006). When more blame is better than less: The implications of internal vs. external attributions for the repair of trust after a competence- vs. integrity-based trust violation. Organizational Behavior & Human Decision Processes, 99(1), 49-65. doi:10.1016/j.obhdp.2005.07.002

42.

Kim

P. H.

Ferrin

D. L.

Cooper

C. D.

Dirks

K. T.

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

43.

Klein

J. G.

Dawar

(2004). Corporate social responsibility and consumers’ attributions and brand evaluations in a product–harm crisis. International Journal of Research in Marketing, 21, 203-217. doi:10.1016/j.ijresmar.2003.12.003

44.

Lee

B. K.

(2004). Audience-oriented approach to crisis communication: A study of Hong Kong consumers’ evaluation of an organizational crisis. Communication Research, 31, 600-618.

45.

Levine

T. R.

Anders

L. N.

Banas

Baum

K. L.

Endo

A. D. S.

Wong

N. C. H.

(2000). Norms, expectations, and deception: A norm violation model of veracity judgments. Communication Monographs, 67(2), 123-137. doi:10.1080/03637750009376500

46.

Love

E. G.

Kraatz

(2009). Character, conformity, or the bottom line? How and why downsizing affected corporate reputation. Academy of Management Journal, 52, 314-335.

47.

Lyon

Cameron

G. T.

(2004). A relational approach examining the interplay of prior reputation and immediate response to a crisis. Journal of Public Relations Research, 16, 213-241.

48.

Pfarrer

M. D.

Pollock

T. G.

Rindova

V. P.

(2010). A tale of two assets: The effects of firm reputation and celebrity on earnings surprises and investors’ reactions. Academy of Management Journal, 53, 1131-1152.

49.

Pincus

J. D.

Rayfield

R. E.

Cozzens

M. D.

(1991). The chief executive officer’s internal communication role: A benchmark program of research. In Grunig

J. E.

Grunig

L. A.

(Eds.), Public Relations Research Annual (Vol. 3, pp. 1-35). Hillsdale, NJ: Lawrence Erlbaum.

50.

Rhee

Haunschild

P. R.

(2006). The liability of good reputation: A study of product recalls in the U.S. automobile industry. Organization Science, 17(1), 101-117. doi:10.1287/orsc.1050.0175

51.

Rhee

Valdez

M. E.

(2009). Contextual factors surrounding reputation damage with potential implications for reputation repair. Academy of Management Review, 34, 146-168.

52.

Siltaoja

(2006). Value priorities as combining core factors between CSR and reputation—A qualitative study. Journal of Business Ethics, 68(1), 91-111.

53.

Siomkos

Kurzbard

(1994). The hidden crisis in product-harm crisis management. European Journal of Marketing, 28(2), 30-41.

54.

Siomkos

Shrivastava

(1993). Responding to product liability crises. Long Range Planning, 26(5), 72-79.

55.

Skowronski

J. J.

Carlston

D. E.

(1987). Social judgment and social memory: The role of cue diagnosticity in negativity, positivity, and extremity biases. Journal of Personality and Social Psychology, 52, 689-699. doi:10.1037/0022-3514.52.4.689

56.

Sohn

Lariscy

R. W.

(2014). Understanding reputational crisis: Definition, properties, and consequences. Journal of Public Relations Research, 2014, 26, 23-43.

57.

The state of the console: Video game console usage. (2007). Retrieved from http://www.nielsenmedia.com/nc/nmr_static/docs/Nielsen_Report_State_Console_03507.pdf

58.

Tinney

(2005). Understanding the casual gamer. Demogrphics Issue, 1(2). Retrieved from http://www.igda.org/online/quarterly/1_2/casual.php

59.

Wojciszke

(1994). Multiple meanings of behavior: Construing actions in terms of competence or morality. Journal of Personality & Social Psychology, 67, 222-231.

60.

Wojciszke

(1997). Parallels between competence- versus morality-related traits and individualistic versus collectivistic values. European Journal of Social Psychology, 27, 245-256.

61.

Wojciszke

(2005). Affective concomitants of information on morality and competence. European Psychologist, 10(1), 60-70.