MANAGING THE MESSAGE: THE EFFECTS OF … · THE EFFECTS OF FIRM ACTIONS AND INDUSTRY SPILLOVERS ON...

MANAGING THE MESSAGE:THE EFFECTS OF FIRM ACTIONS AND INDUSTRY

SPILLOVERS ON MEDIA COVERAGEFOLLOWING WRONGDOING

ANASTASIYA ZAVYALOVARice University

MICHAEL D. PFARRERUniversity of Georgia

RHONDA K. REGERDEBRA L. SHAPIRO

University of Maryland, College Park

We contribute to research on the management of social perceptions by considering therelative effectiveness of a firm’s technical and ceremonial actions in managing mediacoverage after its own or its competitors’ wrongdoing. We examine these relationshipsin the context of product recalls by U.S. toy companies over the ten-year period1998–2007. As hypothesized, firms with higher levels of wrongdoing experience lesspositive media coverage; however, this decline is mitigated during periods of higherindustry wrongdoing. Additionally, we find support for a negative spillover effect: thetenor of media coverage about a focal firm is less positive if others in its industryrecall products. Further, technical actions help firms attenuate the negative effectof their own wrongdoing on the tenor of media coverage, whereas ceremonialactions amplify this effect. In contrast, ceremonial actions are more effective inattenuating the negative effect of industry wrongdoing on the tenor of mediacoverage about a focal firm.

Information intermediaries—third parties suchas the media, financial analysts, regulators, andconsumer organizations—disseminate information,frame issues, and assist stakeholders in making

sense of firm actions. By influencing stakeholders’perceptions about a firm, these infomediaries(Deephouse & Heugens, 2009; Pfarrer, Pollock, &Rindova, 2010; Pollock & Rindova, 2003) play anactive role in the formation of a firm’s social ap-proval (Fombrun, 1996; Fombrun & Shanley, 1990;Pollock & Rindova, 2003; Rao, 1994). In turn, thefirm can leverage these favorable perceptions todevelop intangible assets—such as reputation andlegitimacy—achieve competitive advantage, andaccrue performance benefits (e.g., Deephouse,2000; Deephouse & Carter, 2005; Kennedy, 2008;Martins, 2005; Pollock & Rindova, 2003; Pollock,Rindova, & Maggitti, 2008; Rao, 1994; Zuckerman,1999). Although recent organizational research hasenriched scholars’ understanding of the role of in-fomediaries in shaping stakeholder perceptions, weknow less about how firms can influence the wayinfomediaries portray them, particularly afterwrongdoing. Thus, increasing understanding aboutthis topic is theoretically and practically important(cf. Desai, 2011; Lamertz & Baum, 1998; Pollock &Rindova, 2003; Rindova, Pollock, & Hayward, 2006;Westphal & Deephouse, 2011).

We thank Associate Editor Tim Pollock and the threeanonymous reviewers for many suggestions and supportthat substantively improved our article throughout thereview process. We also thank Wilbur Chung, RafaelCorredoira, Brent Goldfarb, Brad Greenwood, Ben Hal-len, Pam Haunschild, Freek Vermeulen, and DavidWaguespack for their help and comments on earlier ver-sions. We also thank the participants, discussants, andreviewers of the University of Maryland CAP SeminarSeries, Academy of Management Meetings, StrategicManagement Society Meetings, Mid-Atlantic StrategyColloquium, Atlanta Competitive Advantage Conference,University of Texas at Austin, University of Texas atDallas, Texas A&M University, and Texas Christian Uni-versity for their suggestions and encouragement. Finally,we acknowledge Chuck LaHaie and Steven Lawrence fortheir generous assistance with data collection and anal-ysis. Portions of this research were funded by the RobertH. Smith School of Business Summer Research GrantProgram, the Smith Fellows Research Assistance Pro-gram, and the University of Texas at Austin.

� Academy of Management Journal2012, Vol. 55, No. 5, 1079–1101.http://dx.doi.org/10.5465/amj.2010.0608

1079

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download, or email articles for individual use only.

The purpose of this study is to answer two re-lated questions: What are the effects of wrongdoingby a firm and an industry on the tenor of mediacoverage about the focal firm? How do the firm’sactions influence these effects? In this article, wefocus on one particularly influential infomediary,the media—both the printed press and blogs—be-cause of stakeholders’ reliance on these sources asdisseminators of information about firms (Alvesson,1990; Einwiller, Carroll, & Korn, 2010; Jin & Liu,2010). We define wrongdoing as firm behaviors thatplace a firm’s stakeholders at risk and violate stake-holders’ expectations of societal norms and generalstandards of conduct (Coombs, 1995; Pfarrer, De-Celles, Smith, & Taylor, 2008). By examining theinterdependencies among firms, we provide a moredetailed picture of the dynamics occurring betweenthe firm and the industry in which it operates, aswell as the media’s coverage of these dynamics.

We develop our arguments drawing from twostreams of research on (1) the media’s influence onsocial perceptions of firms and (2) crisis and im-pression management. The first research streamsuggests that media shape public opinion (Mc-Combs & Shaw, 1972) and influence stakeholders’impressions of a firm’s behavior (Deephouse, 2000;Pollock et al., 2008; Pollock & Rindova, 2003) bychoosing which issues to cover (McCombs, 1981)and how to frame them (Entman, 2007; Scheufele,1999). Stakeholders use media-disseminated infor-mation to decide if a firm’s behavior is consistentwith their expectations about social and industrynorms (Suchman, 1995) and to decide whether totransact with the firm. However, when a firm vio-lates stakeholder expectations by engaging inwrongdoing, its level of social approval drops, andstakeholders may withdraw transactions (Jonsson,Greve, & Fujiwara-Greve, 2009), thus threateningthe firm’s reputation, legitimacy, and survival(Suchman, 1995).

The second research stream, on crisis and im-pression management, suggests that amid viola-tions of stakeholder expectations, firms must learnto manage stakeholder perceptions of firm behavior(Benoit, 1995; Coombs, 2007; Dukerich & Carter,2000; Elsbach, 2003). However, different respon-sive actions may either attenuate or amplify thenegative effects of wrongdoing on stakeholder per-ceptions. Because the media can shape social per-ceptions about a firm, it is practically and theoret-ically important to understand which firm actionsmay better restore the positive tenor of media cov-erage and increase the firm’s social approval fol-lowing wrongdoing.

Thus, the first contribution of our study lies inthe extension of current organizational research on

social perceptions with the first large-scale empir-ical study that examines how firms can activelyinfluence media coverage, and ultimately, socialapproval. Specifically, we test the relative effec-tiveness of firms’ information subsidies—prepack-aged written pieces of information about firms’ ac-tivities disseminated to the media (Rindova et al.,2006: 62)—on influencing the tenor of media cov-erage about the firms and how this influence mayvary depending on the source of wrongdoing. Wefocus on the relative effectiveness of the announce-ment of two types of firm actions, technical andceremonial. We define technical actions as onesthat have the potential to address the causes ofwrongdoing and thus attract the media and stake-holders’ attention to the internal processes of a firm(cf. Godfrey, Merrill, & Hansen, 2009). In contrast,ceremonial actions are ones that have the potentialto positively alter impressions about the firm anddeflect media and stakeholders’ attention awayfrom the causes of wrongdoing (cf. Kirsch, Gold-farb, & Gera, 2009). Through the theoretical devel-opment and empirical testing of these constructs,we respond to calls for examining how firms influ-ence the infomediaries covering them as well as theprocess of the loss and recovery of a firm’s socialapproval, both of which have been identified asunderstudied areas in organizational research (e.g.,Jonsson et al., 2009; Pollock et al., 2008; Westphal& Deephouse, 2011).

We also contribute to research on impression andcrisis management that has examined how firmaccounts generate responses among stakeholders(Elsbach, 1994; Elsbach & Kramer, 1996) but hasnot investigated the circumstances under whichthe effectiveness of the accounts may vary. Specif-ically, drawing on research from social psychologythat investigates stakeholders’ formation of judg-ments about firms (cf. Mishina, Block, & Mannor,2012), we provide novel theoretical arguments thatexplore the sociocognitive processes that explainhow technical and ceremonial actions affect themedia coverage about a firm after its own or indus-try wrongdoing.

We tested our hypotheses in the context of prod-uct recalls by U.S. toy firms from 1998 to 2007. Ourempirical results indicate that although firms thatrecall more toys are covered less positively in themedia, this effect is mitigated when others in theindustry also engage in recalls. Additionally, wefind support for a negative spillover effect: thetenor of media coverage about a focal firm is lesspositive if others in the industry recall products.We also find that the effectiveness of technical andceremonial actions in restoring the positive tenor ofmedia coverage depends on the source of the recall.

1080 OctoberAcademy of Management Journal

In particular, announcements of technical actionsare helpful in attenuating the negative effect of afocal firm’s recalls on the tenor of media coverage,but announcements of ceremonial actions amplifythis effect. Ceremonial actions, however, help re-duce the negative effect of recalls by other firms inthe industry on the tenor of media coverage aboutthe focal firm.

The remainder of this article proceeds as follows:We first provide an overview of the process ofinfomediaries’ influence on social perceptions.Next, we develop theoretical arguments that ex-plain (1) the effect of firm and industry wrongdoingon the tenor of media coverage about a focal firm,(2) the moderating effect of industry wrongdoing onthe relationship between focal firm wrongdoingand the tenor of media coverage, and (3) the relativeeffectiveness of technical and ceremonial actions inmoderating the negative effect of firm and industrywrongdoing on the tenor of media coverage. Wethen empirically test our hypotheses and discussthe results. We conclude with a discussion of thestudy’s contributions to organization theory andimplications for future research.

THEORY AND HYPOTHESES

Infomediaries, Social Perceptions, andInformation Subsidies

Infomediaries, such as the business press, finan-cial analysts, consumer groups, regulatory agen-cies, and industry experts, influence and shapestakeholders’ perceptions about firms and their ac-tions (Deephouse, 2000; Desai, 2011; Kennedy,2008; Pollock & Rindova, 2003; Zuckerman, 1999).By selecting which issues to cover and how toframe them, infomediaries set the agenda for stake-holder discourse about firm behavior and can in-fluence collective social perceptions (McCombs,1981; Pollock & Rindova, 2003; Rogers, Dearing, &Bregman, 1993). Whereas stakeholder impressionsare formed in many ways, either through direct orindirect experiences (Bandura, 2001; Fiske & Tay-lor, 2008), infomediaries play a major role as third-party actors in firm-stakeholder relationships byproviding opportunities for stakeholders to learnvicariously about firm actions. In this study, wefocus on the media—both the printed press andblogs—because of stakeholders’ reliance on them asprimary sources of information about firms (Alves-son, 1990; Einwiller et al., 2010; Jin & Liu, 2010).

The ability of the media to influence the socialapproval of firm actions has important ramifica-tions for firms. Stakeholders are more likely totransact with, and approve of, firms that meet and

exceed their expectations (Floyd, Ramirez, & Bur-goon, 1999; Scott, 1995). As a result, firms thatobtain stakeholder approval are more likely tobuild intangible assets, such as legitimacy and rep-utation, and they have greater chances of survivaland economic success (Pfarrer et al., 2010). In theearly stages of the U.S. auto industry, for instance,firms obtained legitimacy and built reputationthrough the media coverage of their contests (Rao,1994). Similarly, commercial banks covered posi-tively in the media had higher financial returnsthan banks that received less favorable coverage.The media “provided a forum of what constitutes agood firm,” and positive coverage represents favor-able public opinion about firms (Deephouse, 2000:1097). Similarly, investors often form their impres-sions about new firms and evaluate them finan-cially by relying on information provided by themedia as well as by other types of infomediaries(Pollock et al., 2008; Zuckerman, 1999). Thus, afirm’s ability to influence media coverage can beconsequential for how it manages stakeholder im-pressions and their approval of firm actions, as wellas for the development and maintenance of intan-gible assets.

Although organizational researchers have ac-knowledged that a firm’s information subsidiescan influence the media, which in turn can affectstakeholder perceptions (e.g., Rao, 1994; Rhee &Haunschild, 2006; Rindova et al., 2006; Westphal &Deephouse, 2011), what is less clear is how firmscan influence the tenor of media coverage, espe-cially following wrongdoing. In particular, when afirm or its peers are engaged in wrongdoing, whatactions should the firm take to alter negative mediacoverage and consequently alter stakeholders’ per-ceptions? Additionally, does the effectiveness ofthe announced firm actions depend on whether thefirm itself was involved in wrongdoing or whetherit was instead only “guilty by association”? Inves-tigating the effectiveness of firm actions requiresfirst gaining better understanding of both wrongdo-ing as a violation of stakeholders’ expectations anddirect effects of the magnitude of firm and industrywrongdoing on the tenor of media coverage ofa firm.

Wrongdoing and Expectancy Violations

Research in social psychology suggests that indi-viduals generally expect positive outcomes and areoptimistic about their futures (Fiske & Taylor, 2008;Parducci, 1968; Pezzo, Pezzo, & Stone, 2006). Ac-cording to expectancy violations theory, conform-ing behavior remains largely unnoticed, but viola-tions attract attention because of their salience

2012 1081Zavyalova, Pfarrer, Reger, and Shapiro

and deviance from commonly held expectations(Floyd et al., 1999). Specifically, negative viola-tions of expectations generate negative emotionalresponses and create cognitive dissonance by alter-ing individuals’ views of the way things should be(Burgoon, Denning, & Roberts, 2002; Festinger,1957; Greifeneder, Bless, & Pham, 2011). Thus,when a firm or its peers engage in wrongdoing, themedia and stakeholders actively seek new informa-tion about the firm (Floyd et al., 1999; Planalp &Fitness, 1999) and recalibrate their impressions ofit (Taylor, 1991).

Product recall announcements are an example ofnegative violations of social expectations. Sales ofdefective products breach the implied social con-tract between a firm and its stakeholders, whomaintain “a set of values, beliefs, and norms” ac-cording to which their exchange with the firmshould take place (Morrison & Robinson, 1997:246). Thus, product recalls are likely to be inter-preted as wrongdoing because they violate socialexpectations about the firm’s ability to act accord-ing to an implied promise of appropriate behavior.In the following section, we propose two baselinehypotheses that investigate two types of wrongdo-ing—those committed by a firm and those commit-ted by its peers—and their effects on the tenor ofmedia coverage about the firm.

Firm wrongdoing. Firms engaged in wrongdoingare likely to generate negative media coverage. Inturn, they may lose stakeholder approval becausesuch conduct violates stakeholder perceptions ofacceptable firm behavior and potentially placesstakeholders at risk (Pfarrer, et al., 2008, Pfarrer,Smith, Bartol, Khanin, & Zhang, 2008; Suchman,1995). For example, financial wrongdoing leads tomore negative perceptions by stakeholders about afirm’s ability to create value over time (Kang, 2008).Layoffs and downsizing can violate stakeholders’expectations of the firm’s credibility and commit-ment to them and may be interpreted by the af-fected stakeholders as wrongdoing (Love & Kraatz,2009). Further, the amount of damage caused bycorporate wrongdoing should be expected to affectthe tenor of media coverage. Research in socialpsychology and management suggests that thegreater the scale of a wrongdoing event, the moreattention the media will pay to the event (Deep-house, 2000; Fiske & Taylor, 2008; Rindova et al.,2006), thus lowering the social approval of the in-volved firm. For example, severe industrial acci-dents (Zyglidopoulos, 2001) and severe auto recalls(Rhee & Haunschild, 2006) are associated withlower reputation because such incidents receivehigher levels of media attention and scrutiny. Al-though past organizational research has proposed

theoretical arguments associating firm wrongdoingwith greater media coverage, the relationship be-tween wrongdoing and the tenor of coverage hasnot been studied empirically. Thus, we hypothesize:

Hypothesis 1. The greater the magnitude of afirm’s wrongdoing, the less positive the tenor ofmedia coverage about the firm.

Industry wrongdoing. Recent organizational re-search suggests that firms belonging to industrieswhose members engage in wrongdoing can sufferfrom negative spillovers (Barnett & Hoffman, 2008;Barnett & King, 2008; Kostova & Zaheer, 1999).Because stakeholders stratify objects, includingfirms, into categories (Porac & Thomas, 1990; Reger& Palmer, 1996), firms from the same industry as awrongdoer could be perceived as being “guilty byassociation” (Lange, Lee, & Dai, 2011: 181). Sa-lience of negative social stimuli is usually greaterthan salience of positive stimuli when social expec-tations are violated (Skowronski & Carlston, 1989);thus, negative events, such as product recalls, at-tract more attention than positive ones. The mediaand stakeholders are likely to focus on negativestimuli associated with a specific firm engaged inwrongdoing when forming judgments of otherfirms in the same industry. Specifically, the mediaand stakeholders may assume that if one firm inthe industry is engaged in wrongdoing, other firmsmay be experiencing similar problems. Further-more, stakeholders’ memories about the name ofthe firm engaged in wrongdoing may be incorrect,creating “category confusion” (Fiske & Taylor,2008: 268). Such generalizations may contribute toa mistaken attribution of guilt to an otherwise in-nocent firm in the same industry as a wrongdoer.

In organizational research, generalizing the mis-conduct of one firm to others has been described as“categorical delegitimization” (Greve, Palmer, &Pozner, 2010: 89; Jonsson et al., 2009) or “negativespillover” (Barnett & King, 2008: 1160). A few re-cent empirical studies show support for these ef-fects but have not investigated the mechanismsthrough which negative spillovers occur. For exam-ple, Jonsson and colleagues (Jonsson et al., 2009)theorized that when stakeholders generalized theeffects of organizational wrongdoing by one firm toother firms in the mutual fund industry, the gener-alization led stakeholders to withdraw transactionsfrom firms similar to the violating firm. Barnett andKing (2008) likewise found that the wrongdoing ofone firm in the U.S. chemical industry spilled overto other firms, thus harming firms that were notdirectly at fault. Finally, in a study of the effects offinancial mismanagement on the spillover of repu-tational penalties, Kang demonstrated that firms


unsuspected of wrongdoing but retaining a boardmember from an accused firm were also likely to bepenalized by their stakeholders (2008).

This research suggests that one firm’s wrongdo-ing can negatively impact other firms in the sameindustry, even if those other firms are innocent.However, the mechanism underlying the effect ofnegative spillovers on the tenor of media coveragehas not been studied empirically. In the context ofthis study, some firms in the U.S. toy industry mayhave become guilty by association because of me-dia reports that did not always name the companieswith recalls, referring instead to the recalls as anindustry-wide problem and using more generalphrases such as “toy recalls” or “children’s productrecalls.” When media coverage about an industry isnegative, it can result in a negative evaluation of aspecific firm, even if the firm itself is not “the targetof the negative tone” (Carroll, 2009: 5). Such gen-eralizations and uncertainty about a firm’s involve-ment in wrongdoing may cause stakeholders to per-ceive the firm to be guilty of wrongdoing because itis associated with a guilty industry (Fiske & Taylor,2008). We therefore hypothesize:

Hypothesis 2. The greater the magnitude ofwrongdoing in a firm’s industry, the less posi-tive the tenor of media coverage about the firm.

In sum, our theoretical framework suggests thatchanges in the tenor of media coverage depend notonly on a focal firm’s violation of stakeholder ex-pectations (Hypothesis 1), but also on the preva-lence of wrongdoing by other firms in its industry(Hypothesis 2). If, as we argue above, the wrongdo-ing of some firms can spill over to tarnish otherfirms in an industry (e.g., Kostova & Zaheer, 1999;Yu, Sengul, & Lester, 2008), does a high prevalenceof wrongdoing in the industry attenuate or amplifythe negative effect of firm wrongdoing on the tenorof media coverage about the focal firm?

The Safety-in-Numbers Effect

Social psychology research on attention suggeststhat a particular object is more salient—or standsout more relative to others in the environment—if itis novel or unusual for its category (Jones & McGil-lis, 1976). Increased salience of a particular object,in turn, attracts a perceiver’s attention and en-hances perceptions of prominence. For instance,being the only unpleasant person in a room willgarner disproportionate condemnation (Fiske &Taylor, 2008). We apply similar reasoning to thefirm level of analysis. When a single firm from anindustry engages in wrongdoing, this action is sa-lient because it is novel and unusual in the indus-

try. In such a case, the firm is more likely to attracta disproportionate share of negative publicity andattention and suffer more negative media coverage.However, if several other firms also engage in sim-ilar negative actions, the act loses its novelty andsalience, decreasing the amount of attention paid toany firm in particular (Ahmadjian & Robinson,2001; Pfarrer, et al., 2008). Thus, during times ofwrongdoing in an industry, a focal firm may expe-rience a safety-in-numbers effect: the direct nega-tive effect wrongdoing has on the tenor of mediacoverage will be weakened. This suggests that toaccurately predict the tenor of media coverage as-sociated with different levels of a firm’s wrongdo-ing, one has to account for the level of wrongdoingin its entire industry. We therefore hypothesize:

Hypothesis 3. The magnitude of wrongdoing ina firm’s industry attenuates the negative effectof a firm’s wrongdoing on the tenor of mediacoverage about the firm.

Firm Actions following Wrongdoing

Research in crisis and impression managementsuggests that when a firm’s name is tarnished by itsown or competitors’ wrongdoing, managers can at-tempt to influence media coverage of that firm byproviding a public response (Desai, 2011; Elsbach &Kramer, 1996; Elsbach & Sutton, 1992; Pfarrer, etal., 2008; Westphal & Bednar, 2008; Westphal &Deephouse, 2011). Firms often issue verbal ac-counts about their activities through the media,thus simplifying stakeholders’ search for informa-tion (Rindova et al., 2006) and influencing stake-holder perceptions about the appropriateness offirm actions. However, there is little empirical ev-idence about the effectiveness of different informa-tion subsidies in influencing the tenor of mediacoverage about firms.

We propose that the effectiveness of informationsubsidies in influencing media coverage about afirm after wrongdoing depends on the actions thefirm announces. We categorize firm actions as ei-ther technical or ceremonial, depending on the ac-tions’ capacity to address the cause of wrongdoing,to focus on changes in internal processes or exter-nal evaluations about the firm, and to attract ordeflect the media and stakeholders’ attention towrongdoing.1 The technical category includes ac-

1 These action categories are similar to other classifi-cations found in the organizational literature, such as“technical” and “institutional” (e.g., Godfrey et al., 2009;Lamertz & Baum, 1998; Scott, 2003; Thompson, 1967),“core” and “peripheral” (e.g., Hannan & Freeman, 1984),


tions that are perceived as having the potential toaddress the cause of wrongdoing and thus attractmedia and stakeholder attention to it. Examples oftechnical actions in our sample include Mattel’smonitoring of manufacturing facilities by involvingan “independent monitoring council” (Newswire,2002) and LeapFrog’s public announcement afterrecalling the musical activity center Learn-AroundPlayground in 2006 that “consumers should imme-diately take the recalled activity center away fromchildren and contact LeapFrog for a free repair kit”(Newswire, 2006). In contrast to the technical cat-egory, the ceremonial category includes a firm’sactions taken in the days after recalls that are per-ceived as having the potential to alter stakeholderperceptions of the firm by emphasizing its positive,alternative character traits (Ashforth & Gibbs, 1990)and deflecting media and stakeholder attentionaway from the wrongdoing. In the context of ourstudy, examples of ceremonial actions include Has-bro’s announcement that “to . . . commemorate theheroes that served during the Pearl Harbor attack,[the firm is making] a donation to the USS ArizonaMemorial Fund” (Business Wire, 2001). Anotherexample of a ceremonial action is Mattel’s organi-zation of the children’s contest “How Looney CanYou Be?” (Business Wire, 2003) or FAO Schwarz’sannouncement that they “are inviting kids acrossthe country to help make a difference by participat-ing in the first-ever Play-A-Thon™ . . . by offeringthem a chance to raise money for a charity of theirchoice” (Business Wire, 2001). Using theory fromsocial psychology, we hypothesize how and whythe two types of actions may attenuate or amplifythe negative effect that firm or industry wrongdoinghas on the tenor of media coverage about firms.

Actions after firm wrongdoing. As we hypothe-sized above, the magnitude of a focal firm’s wrong-doing decreases the positive tenor of media cover-age. More specifically, a product recall will inducenegative emotions among interested stakeholderswho view the recall as a violation of their expecta-tions of appropriate firm behavior (Floyd et al.,1999; Greifeneder et al., 2011; Mishina et al., 2012).Research in social psychology suggests that indi-viduals in emotional states are more likely to avoidcognition that is inconsistent with an overall per-ception they hold (e.g., Chaiken, Liberman, &Eagly, 1989; Festinger, 1957; Forgas, 1995). That is,information that reinforces social perceptions

about a firm’s recall is more likely to be utilizedwhen subsequent judgments are made (cf. Mishinaet al., 2012). Thus, the media may be more likely toattend to a firm’s technical actions because they areinterpreted as related to the recall and are consis-tent with the already-established impressions thatthe media and stakeholders have of the wrongdo-ing firm.

In line with these arguments, the crisis and im-pression management literature suggests that tak-ing actions that signal a firm is in control andaddressing the problem is effective in helping thefirm recover from wrongdoing (Coombs, 2007;Mishina et al., 2012; Pfarrer, et al., 2008). Conse-quently, theory suggests, and empirical studieshave shown, that the focal firm can effectively re-store its social approval after wrongdoing by di-rectly identifying the cause of the problem, dedi-cating resources to minimizing the effects of thenegative event, and addressing the perceived vio-lation as soon as possible (Elsbach, 1994; Mercer,2005; Mishina et al., 2012; Pfarrer, et al., 2008a).The firm can do this by providing information sub-sidies that direct the media’s attention to thechanges made to the firm’s internal processes andreinforce the ways the firm is perceived to be rec-tifying the problem.

In contrast, the media may view ceremonial ac-tions as inconsistent with their perceptions of thefocal firm subsequent to wrongdoing. Becausewrongdoing is perceived as a negative violation ofsocial expectations, it is a more salient and domi-nant cue than positive actions in which the firmmay engage (Lingle, Geva, Ostrom, Leippe, &Baumgardner, 1979). Moreover, negative behaviorsare perceived as more diagnostic than positive be-haviors after social violations (Mishina et al., 2012).They also tend to be “stickier” (Skowronski & Carl-ston, 1987), making it more difficult for the firm torestore its social approval after wrongdoing simplyby engaging in ceremonial actions that are per-ceived as positive. As a result, attempts to deflect oralter current stakeholder perceptions by emphasiz-ing highly visible actions that are consistent withsocial expectations (Ashforth & Gibbs, 1990) maybackfire. If these actions are viewed as superficialor hypocritical, the media and stakeholders couldperceive them with suspicion, distrust, and skepti-cism (Desai, 2011; Elsbach, Sutton, & Principe,1998; Lyon & Maxwell, 2011; Mishina et al., 2012).In sum, the negative impressions generated by afirm engaged in wrongdoing will dominate the im-pression formation process. In turn, the media maysuspect concealed motivation behind ceremonialactions, which would hinder the firm’s attempts at

“analytical” and “affective” (e.g., Epstein, 1994; Hastie &Dawes, 2001), “substantive” and “symbolic” (Westphal &Zajac, 1998), and “threat-addressing” and “attention-deflecting” (Elsbach & Kramer, 1996).


restoring its positive media coverage. We thereforehypothesize:

Hypothesis 4a. A firm’s announcements oftechnical actions attenuate the negative effectof the firm’s wrongdoing on the tenor of mediacoverage about the firm.

Hypothesis 4b. A firm’s nnouncements of cer-emonial actions amplify the negative effect ofthe firm’s wrongdoing on the tenor of mediacoverage about the firm.

Actions after industry wrongdoing. We theorizethat the relationships hypothesized above will bereversed when others in a firm’s industry, ratherthan the firm itself, have engaged in wrongdoing.We derive our logic from the increased levels ofuncertainty the media and stakeholders face whenattributing wrongdoing to a specific firm. In thiscase, an innocent firm may be able to distance itselffrom wrongdoers in its industry by deflecting theattention of the media away from the transgressionsof peers and toward the focal firm’s positive actionsand attributes (Elsbach & Kramer, 1996). If the me-dia portray the focal firm as the perpetrator ofwrongdoing, there is likely little uncertainty amongstakeholders about its culpability. In contrast, justas stereotypes are formed about people, when somefirms in the industry are engaged in wrongdoing,observers may think that other industry membersare guilty of wrongdoing as well (Jonsson et al.,2009; Yu et al., 2008). In this case, the beliefs abouta specific firm’s culpability, however, are less cer-tain. Thus, the media may interpret actions takenby a firm that is perceived to be guilty by associa-tion differently than actions taken by the perpetrat-ing firm.

After a firm’s competitors have engaged inwrongdoing, the firm’s announcements of technicalactions may draw attention to its operations andreinforce negative impressions about the firm stem-ming from its membership in a guilty industry.Additionally, stakeholders may interpret the firm’sannouncements of technical actions in the contextof industry wrongdoing as an indication that thefirm is encountering problems similar to those ofits competitors, but that its wrongdoing simplyhas not been exposed yet. Thus, in this instance,taking technical actions may signal that a firm isalso culpable.

Ceremonial actions, in contrast, are highly visi-ble actions that “are consistent with social expec-tations [but leave] the essential machinery of theorganization intact” (Ashforth & Gibbs, 1990: 181).The purpose of ceremonial actions is to appear toadhere to the norms and expectations of the larger

social context in which a firm operates and to pro-mote the belief that the firm’s activities are congru-ent with the values and expectations of its stake-holders (Kirsch et al., 2009). Taking such actionsamid industry wrongdoing may “invoke alternatecategorization schemes” (Elsbach & Kramer, 1996:466) by deflecting the media’s attention away fromindustry wrongdoing and emphasizing the positivedimensions of the firm. These actions help firmsthat are not directly involved in product recallsavoid association with the wrongdoers (Elsbach &Kramer, 1996; Elsbach & Sutton, 1992) and helpthem differentiate themselves from the guilty in-dustry by focusing media attention on socially de-sirable actions (Elsbach & Sutton, 1992). By doingso, firms may be able to recategorize themselves inthe media by amplifying their positive attributesnot associated with the wrongdoing of other firms.

Although the effect of ceremonial actions on thetenor of media coverage has not been studied em-pirically, recent research has shown that organiza-tions that the media and stakeholders link to aguilty organization will avoid being “draggeddown” (Greve et al., 2010: 89) by cutting off allties to the wrongdoer (cf. Jensen, 2006; Sullivan,Haunschild, & Page, 2007). Similarly, executiveswho leave a firm prior to its failure are less likely tosuffer demotions in their new jobs than managerswho stayed (Semadeni, Cannella, Fraser, & Lee,2008). Announcements of ceremonial actions mayhelp reduce uncertainty about the character of aspecific firm that stakeholders may otherwise havecategorized as being guilty by association. Suchannouncements help a firm distance itself from thecategory of wrongdoers by deflecting the media’sattention away from the transgressions of theircompetitors and toward the focal firm’s positiveattributes (Greve et al., 2010; Jonsson et al., 2009).In turn, the media are more likely to report posi-tively about such a firm. Thus, we hypothesize:

Hypothesis 5a. A firm’s announcements oftechnical actions amplify the negative effect ofindustry wrongdoing on the tenor of mediacoverage about the firm.

Hypothesis 5b. A firm’s announcements of cer-emonial actions attenuate the negative effect ofindustry wrongdoing on the tenor of mediacoverage about the firm.

METHODS

Sample

To test the hypotheses derived from our theoret-ical model, we examined media coverage of public


U.S. toy firms in each quarter from 1998 to 2007.2

We focused on publicly traded firms because theyare required to report financial statements and arealso more likely than private firms to make publicannouncements regarding their actions due to theirvisibility in the media and the pressure of meetingstakeholders’ expectations (e.g., Chen & Meindl,1991; Marcus & Goodman, 1991; Salancik &Meindl, 1984; Sutton & Galunic, 1996). Our focuson one industry allowed us to examine media cov-erage within a clearly bounded area where firmsaffect, and are affected by, each other’s actions. Ourchoice to examine the toy industry was also due tothe industry’s high level of concentration (www.hoovers.com), which provided a good context forexamining instances in which firms were prone toexperience negative spillover effects from compet-itors’ actions (Yu et al., 2008).

To construct the sample, we used the CRSPSift2and CPSC toy recalls databases to identify all com-panies whose primary business was listed withinthe toy industry from 1998 to 2007. Because severalfirms in the databases did not explicitly producechildren’s toys (e.g., casinos, golf courses, and golfequipment makers), we screened the pool in fouradditional steps to ensure that firms in our samplewere relevant members of the toy industry. Thefirst author and two trained research assistants(1) tracked each firm’s ownership type (public orprivate) for the entire period of the study by usingthe SDC Platinum database; (2) evaluated eachfirm’s (or in the case of subsidiaries, each parentfirm’s) SIC code; if the description of the SIC codeincluded the word “toy” (SIC codes 3942, 3940,3944, 5945, and 5092), the firm was retained in thesample; (3) used Hoover’s database (www.hoovers.com) to conduct a company name search, retainingin the sample firms whose industry was listed inHoover’s as “Toys and Games” or “Toys andGames, Retail”; and (4) examined each firm’s web-site to determine whether toys were its primaryproduct. This search yielded 45 firms, 21 of whichexperienced toy recalls during the period of thestudy. Because some firms were not publicly tradedthroughout the entire period, the final sample con-sists of 940 firm-quarter observations for which allthe variables of interest were observed.

Variables

Tenor of media coverage. Our dependent vari-able is the tenor of media coverage about a firm.The media frame stories and influence stakehold-ers’ perceptions about firms through their use ofpositive or negative language (Deephouse, 2000;McCombs, Llamas, Lopez-Escobar, & Rey, 1997;Pollock & Rindova, 2003). To measure the tenor ofmedia coverage of a firm, we used the Lexis-Nexisdatabase to identify articles published in the 50largest U.S. newspapers by circulation and blogs3

from the first quarter of 1998 through the first quar-ter of 2008. This search resulted in approximately37,500 articles and blog postings. We then analyzedthe content of the articles and blogs about each firmin a given quarter using Linguistic Inquiry andWord Count (LIWC), text analysis software de-signed to determine the rate at which authors orspeakers use words connoting positive or negativeemotion in a given text (Pennebaker, Booth, & Fran-cis, 2007).4 Like the authors of previous studies(e.g., Deephouse, 2000; Pew Research Center, 2008;Pfarrer et al., 2010; Pollock & Rindova, 2003), wecalculated the percentage of positive and negativecontent in each article and coded it as positive if itstotal affective content was at least 66 percent pos-itive and as negative if its total content was at least66 percent negative.5

One issue that may arise from coding the affec-tive content of entire articles and blogs throughLIWC is confusion resulting from mentions of mul-tiple firms in a text (cf. Carroll, 2009; Lamertz &Baum, 1998; Pollock et al., 2008). For example, anarticle may have predominantly negative affectivecontent, but it may describe a specific firm in apositive light. To ensure that LIWC’s coding pro-gram accurately reflected the tenor of coverageabout a given firm and to follow established prac-tices in content analysis methodology, the first andthird authors recoded 100 randomly selected arti-cles and blogs (cf. Desai, 2011; McCarthy, McPhail,& Smith, 1996; Pollock et al., 2008). A discrepancybetween the tenor of an entire article and the tenorof coverage of a specific firm arose in only threecases: in each occurrence, the overall tenor of the

2 The Consumer Product Safety Commission (CPSC)began tracking toy recalls in 1974. As of 2007, the major-ity of recall announcements (51%) and the majority ofrecalled toys (59%) had occurred during the ten-yearperiod of our study.

3 The “Web Blogs” category of Lexis-Nexis sourcesprovides information published in 35 different blogs thatcontain postings of news discussions and public opinionon various issues (www.lexisnexis.com).

4 For an in-depth description of the reliability andexternal validity of LIWC results of text analysis, pleasesee http://liwc.net/liwcdescription.php.

5 We also recoded each article’s “positivity” at the 60and 75 percent levels. Our results were unchanged.


article was negative, whereas the tenor associatedwith a specific firm was positive. As we have the-orized above, however, stakeholders are likely togeneralize the overall negative tenor of an article toany firm mentioned in it due to the high diagnos-ticity and likely spillover of negative informationresulting from social violations (Carroll, 2009;Skowronski & Carlston, 1987). Thus, we concludedthat in the rare event that the overall tenor of anarticle was negative, while the tenor associatedwith a specific firm was positive, the negative affectwas likely to spill over to the firm mentionedpositively.6

Researchers in management (Deephouse, 2000;Pfarrer et al., 2010; Pollock & Rindova, 2003), com-munications (Carroll, 2009), economics (Houser &Wooders, 2006), and finance (Tetlock, Saar-Tsechansky, & Macskassy, 2008) have measuredthe tenor of media coverage about a firm as therelative prevalence of positive over negative arti-cles, feedback, or words. Following this research,we measure the tenor of media coverage about afocal firm as the difference between the number ofpositive and the number of negative articles andblog posts published about a firm.7 We measure the

tenor of media coverage in the quarter of each recalland the following quarter to account for the effecton media coverage of the recall as well as the firm’sinformation subsidies.

Firm wrongdoing. We measure firm wrongdoingas the sum of the number of toys recalled in a givenquarter. To gain insight about U.S. firms’ toy recalls,we turned to the Consumer Product Safety Com-mission (http://cpsc.gov/cpscpub/prerel/category/toy.html), an independent federal regulatoryagency created in 1972 by the U.S. Congress as partof the Consumer Product Safety Act. It “is chargedwith protecting the public from unreasonable risksof serious injury or death from thousands of typesof consumer products under the agency’s jurisdic-tion” (www.cpsc.gov/about). Because sales of de-fective products may cause health hazards such assuffocation, laceration, poisoning, and in extremecases, death (Beamish & Bapuji, 2008; www.cpsc.gov), product recalls violate stakeholder expecta-tions about proper firm conduct. They are also ac-companied by a loss of consumer confidence (Luo,2008) and an increase in negative perceptionsabout the recalling firm (Beamish & Bapuji, 2008).In addition, product recalls can be interpreted aspublic admission of a specific type of wrongdo-ing—wrongdoing that is due to mismanagement(Bromiley & Marcus, 1989; Near, Rehg, Van Scotter,& Miceli, 2004)—because they are consequent toproblems and errors in manufacturing and design(Beamish & Bapuji, 2008; Rhee & Haunschild, 2006)and occur after the sale of defective toys. We col-lected data on all recalls from 1998 to 2007. The 21firms that recalled toys collectively experiencedrecalls of over 56 million toys associated with thou-sands of reported injuries and incidents.

Industry wrongdoing. We measure industrywrongdoing as the sum of all toys recalled in agiven quarter, excluding a focal firm’s toys.

Firm actions. To examine what types of actionsare described in the information subsidies thatfirms issue, the first author collected and codeddata from over 5,500 firm press releases dissemi-nated by the Business Wire and PR Newswiredatabases, two leading sources for press releasesthat companies use to disseminate information

6 It is also possible that the negative tenor of coverageof a given firm could overshadow the generally positivetone of an article (cf. Richey, Koenigs, Richey, & Fortin,1975). Although we did not find any occurrences in oursubsample, to the extent that they do exist, their omis-sions should result in more conservative tests of ourhypothesized relationships.

7 An established measure of the tenor of media cover-age is the Janis-Fadner (JF) coefficient of imbalance(Deephouse, 2000; Janis & Fadner, 1943). However, thisand other ratio-based measures are less suitable in oursample for two reasons. First, high variance in theamount of coverage among firms may contribute to lowcriterion validity of the JF coefficient and other ratio-based measures. For instance, a firm with one positivearticle and no negative articles and a firm with 100 pos-itive articles and no negative articles will both receive ascore of 100 percent. Additionally, a firm with one pos-itive article and no negative articles will receive a higherscore than a firm with 100 positive articles and onenegative article. Therefore, employing the JF coefficientand other ratio-based measures may restrict samples tohighly publicized firms. For example, management stud-ies that have used the JF measure focus on highly visibleand prominent firms in multiple industries (e.g., Pfarreret al., 2010; Pollock & Rindova, 2003). However, intrain-dustry studies similar to ours that employed the JF coef-ficient lost half of their sample size (e.g. Deephouse,1996, 2000; Deephouse & Carter, 2005). Because it wasessential for our study to uncover intraindustry dynam-ics and interdependencies between the actions of a focalfirm and its competitors, the JF and other ratio-based

measures appeared not suitable. Second, an assumptionunderlying such measures is that arguments for oragainst a specific issue are equally salient. However,research in social psychology suggests that people per-ceive negative information to be as much as five timesmore salient than positive or neutral information(Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001;Fiske & Taylor, 1998; Richey et al., 1975), an issue weaddress in the robustness checks section below.


about their actions (www.businesswire.com; www.prnewswire.com). We chose press releases as asource of information on firms’ actions becausethey are reports that firms themselves view as im-portant and because the media use them as infor-mation sources about firm actions (Carroll &McCombs, 2003; Kennedy, 2008). All public com-panies issue press releases at least quarterly. Somestudies have used industry publications to measurefirm actions (Basdeo, Smith, Grimm, Rindova, &Derfus, 2006; Smith, Grimm, Gannon, & Chen,1991), but in our context smaller firms are lesslikely to be mentioned in these media outlets. Us-ing firms’ press releases thus allows us to minimizethis potential bias. By doing so, we also control forthe alternative explanation that the visibility of afirm’s actions through media coverage is drivingour results, as information on both technical andceremonial actions from press releases is equallyvisible to the media and external stakeholders. Toensure the reliability of the initial coding, a trainedresearch assistant recoded a subsample of pressreleases about firm actions in 40 randomly selectedfirm-quarters. The values of Krippendorff’s alphacoefficient for interrater reliability were 0.90 and0.86 for technical and ceremonial actions, respec-tively, indicating high reliability (Krippendorff,2004; Short, Broberg, Cogliser, & Brigham, 2010).

To code firms’ actions, we employed a structuredcontent analysis method (cf. Duriau, Reger, & Pfar-rer, 2007) widely used in the competitive dynamicsliterature (e.g., Basdeo et al., 2006; Smith et al.,1991). Because in a given quarter, some companiesmade several announcements about the same ac-tion, while others made only one announcement,we used a binary measure in which each categoryof actions was coded 1 for a quarter in which a firmannounced the action and 0 otherwise. To createthe categories of technical and ceremonial actionsa firm announced in a given quarter, we summedthe binary measures of action types that belong toeach of the two categories. We measure technicalactions—actions that are perceived as addressingthe problem of manufacturing and selling defectivetoys—as the sum of the following: changes inoperations and distribution channels, changes inproduction management, improvements in manu-facturing processes, start of investigations, discon-tinuance of product shipments, compensation fordefective products, and cooperation with regula-tory agencies investigating the recalls. We measureceremonial actions—actions that do not directlyaddress the cause of a recall but instead highlightpositive characteristics of a firm—as the sum of thefollowing: firm name changes, celebrity endorse-ments, charitable donations, promotions and

sweepstakes, acts of corporate citizenship (e.g.,sponsoring children’s talent shows), and an-nouncements of company awards.

In the context of this study, for instance, follow-ing a recall of ten million Power Wheels® by Fish-er-Price in October 1998, the parent company, Mat-tel, issued a number of press releases updating thepublic on the progress of the work of the toy repaircenters it opened, in which qualified electriciansreceived and repaired shipments of defective toys.Similarly, after toy recalls in summer 2007, RC2announced suspended shipments of defective toyswith the following explanation: “During the quarterwe suspended shipment of products that are sub-ject to the wooden toy voluntary recall announcedon June 13, 2007” (Business Wire, 2007). Such ac-tions were categorized as technical in our sample.

In contrast, after Hasbro recalled nearly nine mil-lion toys in June 2000, no company press releasesmentioned the recall; rather, the day following therecall Hasbro issued a press release announcing thewinners of Hasbro Teens With The Courage ToGive Awards™. As another example, in 2005 Ac-tion Products International Inc. provided “freelearning materials to teachers at over 8,000 publicand private schools across the country” (BusinessWire, 2006). Such actions were categorized as cer-emonial. To ensure that the coded firm actionsoccurred after wrongdoing, we measure technicaland ceremonial actions in the quarter following themeasure of wrongdoing.

Control variables. An alternative explanation forour findings is a size effect: large firms in our samplerecall more toys, take more actions, and may enjoymore positive coverage in the press. To account forthis possibility, we include firm size as a controlvariable measured by assets of a focal firm in a givenquarter. This information was obtained from theCRSPSift2 database. To delineate the effect of currentwrongdoing on subsequent levels of media coverage,we include social memory of firm wrongdoing andsocial memory of industry wrongdoing as two controlvariables. We measure social memory of firm wrong-doing as the sum of the number of toys recalled by afirm and a decay measure that assigns a weight of 1/nfor each quarter prior to the quarter of the recall. Wemeasure social memory of industry wrongdoing asthe sum of the number of toys recalled by the firm’scompetitors and the decay measure. Employing a de-cayed measure accounts for the residual effect of so-cial memory about the firm’s or competitors’ pastactions and thus reflects the media’s and stakehold-ers’ cumulative perceptions about the firm (cf. Darr,Argote, & Epple, 1995; Pfarrer et al., 2010). Both vari-ables are measured in the quarter prior to a recall.Additionally, we control for the number of injuries


and also for incidents, the official CPSC term forconsumer-reported toy defects that did not result inany physical harm to children. We also control for theaverage price of toys recalled in a given quarter aswell as the amount of media coverage. We measurethe latter by the number of articles published about afirm in a given quarter. To delineate the tenor ofmedia coverage of the firm as distinct from that ofcoverage of the toy industry in general, we control forthe tenor of media coverage about the industry, mea-sured as the difference between the number of posi-tive and negative articles published about a firm’scompetitors in the quarter of a recall and the follow-ing quarter. As the operations of the U.S. toy industryare seasonal, in that more toys are sold during theholiday season (Freedman, Kearney, & Lederman,2012; Johnson, 2001), we also include a fourth quar-ter dummy in our analysis. Finally, the largest num-ber of toy recalls occurred in 1998 and 2007. Wetherefore include dummies for both of those years.

Data Analysis

To test our hypotheses, we selected an estimationprocedure that was appropriate for our theoreticalarguments and was robust to typical issues that arisewith analyses of panel data. Because we controlledfor previous levels of firm and industry wrongdoing,we were cognizant that autocorrelation could affectour results. Using the Wooldridge test of autocorrela-tion (via the Stata command “xtserial”), we rejectedthe null hypothesis of absence of autocorrelation. Toaddress this issue, we implemented the Arellano-Bond model and the generalized method of moments(GMM) procedure (Arellano & Bond, 1991; Halaby,2004; Wade, Porac, Pollock, & Graffin, 2006), whichincludes a lagged dependent variable as an instru-ment. Additionally, this estimation reports first dif-ferences, thus accounting for the change in the tenorof media coverage over each period and making itfunctionally similar to a fixed-effects model (Halaby,2004). The Arellano-Bond model is also robust toheteroskedasticity and violations of normality, mak-ing it suitable for the panel structure of our sample(Arellano & Bond, 1991; Halaby, 2004). Because threeof our hypotheses predict a moderation effect, westandardized all variables to account for possiblemulticollinearity between the main effects and inter-action effects (Aiken & West, 1991).

RESULTS

Table 1 shows descriptive statistics of unstan-dardized variables and correlations for the stan-dardized variables of interest. Using the full model,we tested all the variables for the presence of mul-

ticollinearity. All variance inflation factors werebelow 6, with an average of 2.44; thus, multicol-linearity was not a concern (Chatterjee & Price,1991). Table 2 presents the results of the hypothe-ses’ tests. Model 1 includes control variables;model 2 illustrates the results of the main effectsregression; and model 3 is the full model that testsall main and interaction effects. The chi-square dif-ference test indicates a marginally significant, pos-itive difference between model 1 and model 2 (�2[4]� 8.38, p � .10) and a significant positive differ-ence between model 2 and model 3 (�2[5] � 110.25,p � .001), suggesting that model 2 fits the databetter than model 1 and that the full model, model3, is the best fit for the data, demonstrating jointsignificance of all five interaction terms (�2[5] �82.63, p � .001). We utilize model 2 to interpret theresults of the main effects (Hypotheses 1 and 2[Aiken & West, 1991]) and model 3, the full model,to interpret the interaction effects (Hypotheses 3, 4aand 4b, and 5a and 5b).

To test Hypotheses 1 and 2, we regressed themain effects of firm and industry wrongdoing onthe tenor of media coverage and a vector of controlvariables as well as technical and ceremonial ac-tions. As model 2 shows, consistently with Hypoth-esis 1, firm wrongdoing had a significant negativeeffect on the tenor of media coverage (� � �0.03, p� 0.05). The coefficient indicates that if a firmrecalled 635,000 toys (one standard deviationabove the mean), the number of negative articlespublished about it would increase by slightly morethan 2 (�0.03 � 72). Hypothesis 2 predicts thathigher levels of industry wrongdoing are associatedwith a more negative tenor of media coverage. In-dustry wrongdoing had a marginally significant,negative effect on the tenor of media coverage (� �–0.02, p � .10). Thus, Hypothesis 2 receives mar-ginal support and indicates that if the number oftoys recalled by a firm’s competitors increased by3,489,000 (one standard deviation above the mean),the media would publish approximately 1.5 morenegative articles about the focal firm (�0.02 � 72).

Although the effects of firm and industry wrong-doing on the tenor of media coverage may seemsmall, when interpreting these results one shouldconsider the widely observed positivity bias in thebusiness press (Deephouse, 2000; Fombrun & Shan-ley, 1990; Pollock & Rindova, 2003) and the per-ceived salience and diagnosticity of negative infor-mation for social perceptions (Baumeister et al.,2001; Richey et al., 1975). Thus, stakeholders mayperceive a small increase in neutral or negativearticles as a definite negative signal, much likefinancial analysts’ “hold” recommendations (Abar-banell & Lehavy, 2003; Brown, 2001; Skinner &


TA

BL

E1

Des

crip

tive

Sta

tist

ics

and

Cor

rela

tion

sa

Var

iabl

eM

ean

s.d

.M

inim

um

Max

imu

m1

23

45

67

89

1011

1213

1415

1.T

enor

ofm

edia

cove

rage

offo

cal

firm

t�

(t�

1)

30.9

971

.93

�4.

0044

3.00

2.F

irm

wro

ngd

oin

gb0.

606.

350.

0011

3.22

.27*

3.In

du

stry

wro

ngd

oin

gb14

.65

34.8

90.

0015

4.00

�.0

1.0

34.

Tec

hn

ical

acti

ons t

�1

0.26

0.53

0.00

5.00

.17*

.07*

.00

5.C

erem

onia

lac

tion

s t�

10.

350.

720.

004.

00.4

0*.0

8*�

.08*

.12*

6.S

ocia

lm

emor

yof

pri

orfi

rmw

ron

gdoi

ng t

–1

1.79

8.14

0.00

127.

10.4

3*.0

8�

.03

.15*

.30*

7.S

ocia

lm

emor

yof

pri

orin

du

stry

wro

ngd

oin

g t–

1

44.1

437

.65

0.59

160.

01�

.07*

�.0

4�

.11*

.01

�.1

2*�

.04

8.A

sset

sc79

9.80

1,77

3.32

0.00

10,6

95.0

0.8

7*.1

5*�

.04

.16*

.38*

.32*

�.0

8*9.

Art

icle

s28

.99

74.2

10.

0084

1.00

.92*

.38*

.01

.16*

.35*

.42*

�.0

6.8

3*10

.In

juri

es0.

708.

830.

0019

7.00

.17*

.44*

.01

.03

.06

.07*

.01

.12*

.15*

11.

Inci

den

ts3.

7541

.55

0.00

705.

00.2

4*.6

9*.0

2.0

6.0

8*.0

8*�

.04

.13*

.30*

.56*

12.

Ave

rage

pri

ceof

reca

lled

toys

1.31

7.66

0.00

140.

27.2

5*.5

4*.0

1.1

6*.1

0*.2

3*�

.05

.17*

.27*

.26*

.61*

13.

Ten

orof

ind

ust

rym

edia

cove

rage

t�

(t�

1)

908.

9128

4.58

408.

001,

573.

00�

.15*

.01

.39*

�.0

2�

.19*

�.0

7*.2

7*�

.24*

�.1

4*�

.01

�.0

1�

.04

14.

Yea

r19

980.

110.

310.

001.

00.0

0.0

5.3

6*.0

1�

.09*

�.0

7*�

.39*

�.0

4�

.01

�.0

1.0

9*.0

4.2

4*15

.Y

ear

2007

0.06

0.25

0.00

1.00

.04

.09*

.17*

.10*

.03

.10*

.18*

.02

.08*

.02

.06

.12*

.03

�.0

9*16

.F

ourt

hqu

arte

r0.

250.

430.

001.

00.0

4.0

2.1

6*�

.01

.00

.02

.03

.02

.09*

�.0

4.0

0.0

0.3

4*.0

7*.0

0

an

�94

0.b

Hu

nd

red

sof

thou

san

ds.

cM

illi

ons.

*p

�.0

5

Sloan, 2002) or a lukewarm recommendation letterfrom a previous employer (Range Menyhert, Walsh,Hardin, Ellis, & Craddick, 1991; Siskind, 1966).

Hypothesis 3 predicts that industry wrongdoingwill attenuate the negative effect of firm wrongdo-ing on the tenor of media coverage about a focalfirm. Model 3 presents the test of this hypothesis.The interaction coefficient of firm wrongdoing andindustry wrongdoing was positive and significant(� � 0.02, p �.05). To interpret this result in moredetail, we graphed the moderating relationship (Ai-ken & West, 1991; Hoetker, 2007). Figures 1–4 aredrawn using the respective coefficients from model3, with “low” indicating levels of the standardizedvariables one standard deviation below the meanand “high,” one standard deviation above themean. Concerning Hypothesis 3, Figure 1 illus-trates that the negative effect of firm wrongdoing onthe tenor of media coverage about a focal firm wasattenuated with increasing levels of industrywrongdoing. This indicates that it is more damag-ing for firms to recall products when very few otherfirms engage in recalls. Conversely, recalling prod-ucts during periods when others engage in recallsseems to have a less adverse effect on the tenor of a

focal firm’s coverage. This result suggests that highindustry wrongdoing provides a safety-in-numberseffect for firms engaged in product recalls. Thus,Hypothesis 3 is supported.

Hypothesis 4a predicts that technical actions willattenuate the negative effect of firm wrongdoing onthe tenor of media coverage, and Hypothesis 4b pre-dicts that ceremonial actions will amplify this effect.Model 3 confirms that the interaction term of firmwrongdoing and technical actions was positive andsignificant (� � 0.05, p �.001), but the interactionterm of firm wrongdoing and ceremonial actions wasnegative and significant (� � �0.09, p � .001). Thus,Hypotheses 4a and 4b are supported. Figures 2 and 3illustrate these relationships. Figure 2 shows thattechnical actions are more effective with increasedlevels of firm wrongdoing. Figure 3, however, showsthe opposite relationship: ceremonial actions amplifythe negative effect of firm wrongdoing on the tenor ofmedia coverage.

Hypothesis 5a predicts that technical actions willamplify the negative effect of industry wrongdoingon the tenor of media coverage about a firm, andHypothesis 5b predicts that ceremonial actions willattenuate this effect. As can be seen from model 3

TABLE 2Results of Arellano-Bond Analyses Predicting Tenor of Media Coveragea

Variables Model 1 Model 2 Model 3

Firm wrongdoing (H1) �0.03* 0.00Industry wrongdoing (H2) �0.02† �0.01Firm wrongdoing � industry wrongdoing (H3) 0.02*Firm wrongdoing � technical actions (H4a) 0.05***Firm wrongdoing � ceremonial actions (H4b) �0.09***Industry wrongdoing � technical actions (H5a) �0.01Industry wrongdoing � ceremonial actions (H5b) 0.03*Technical actionst � 1 �0.01 �0.01Ceremonial actionst � 1 0.00 0.01Tenor of media coverage of focal firmt – 1 0.30*** 0.29*** 0.19***Social memory of prior firm wrongdoingt – 1 �0.01 �0.01 0.00Social memory of prior industry wrongdoingt – 1 0.02† 0.02 0.02Assets 0.52*** 0.51*** 0.46***Articles 0.30*** 0.33*** 0.42***Injuries 0.00 0.00 �0.01Incidents 0.02 0.03* 0.05**Average price of recalled toys �0.01 0.00 �0.04***Tenor of media coverage of industryt � (t � 1) 0.08*** 0.08*** 0.09***Year 1998 0.22*** 0.24*** 0.21***Year 2007 �0.09† �0.07 �0.07Fourth quarter �0.09*** �0.09*** �0.11***Constant 0.08*** 0.08*** 0.10***�2 1,114.75 1,123.13 1,233.38

a The Arellano-Bond model uses a lagged dependent variable as an instrument. Thus, the number of observations (n � 897) is smallerthan in the original sample.

† p � .10* p � .05

** p � .01*** p � .001


in Table 2, the interaction term of industry wrong-doing and technical actions was not significant,indicating a lack of support for Hypothesis 5a. Incontrast, the interaction coefficient of ceremonialactions and industry wrongdoing was positive andsignificant (� � 0.03, p � .05), supporting Hypoth-esis 5b. To better understand the significant rela-tionship in Hypothesis 5b, we graphed it. Figure 4illustrates that the relationship between industrywrongdoing and the tenor of media coverage abouta focal firm becomes less negative with an increasein the number of the firm’s ceremonial actions.

Robustness Checks

Alternative operationalizations of tenor of me-dia coverage. In addition to the alternative speci-fications of our dependent variable (described infootnote 5) several other robustness checks wereconducted. First, to account for the larger perceivedsalience and impact of negative information onmedia coverage and stakeholders’ perceptions(Baumeister et al., 2001; Fiske & Taylor, 2008;Wartick, 1992), we weighted the number of nega-tive articles in our composite measure by 2, 3, 4,

FIGURE 1Moderation Effect of Industry Wrongdoing on the Relationship between

Firm Wrongdoing and Tenor of Media Coverage

FIGURE 2Moderation Effect of Technical Actions on the Relationship between



and 5 times. Our results remained substantivelyunchanged. Second, following research in socialpsychology that suggests that positive and negativeemotions lie on two separate continua rather than asingle continuum (Brief & Motowidlo, 1986; Ca-cioppo & Gardner, 1999; Rindova et al., 2006), andrelying on prior empirical work that separates pos-itive and negative information into two distinctconstructs (Houser & Wooders, 2006; Tetlock et al.,2008), we reran our analyses using (1) the sum ofpositive articles and (2) the sum of nonpositive (i.e.,neutral and negative) articles as two alternative

dependent variables. Our latter measure is consis-tent with the above-mentioned prevalence of a pos-itivity bias in the business press, which may lead tointerpretations of statements neutral (i.e., nonposi-tive) in tenor as negative (Abarbanell & Lehavy,2003; Houser & Wooders, 2006; Siskind, 1966; Tet-lock et al., 2008). For the former measure, our re-sults remained unchanged, signaling that toy re-calls and subsequent firm actions affect theabsolute levels of positive media coverage in thesame way that they affect our primary dependentvariable. Similarly, for the latter measure, our re-

FIGURE 4Moderation Effect of Ceremonial Actions on the Relationship between

Industry Wrongdoing and Tenor of Media Coverage

FIGURE 3Moderation Effect of Ceremonial Actions on the Relationship between



sults were mostly supported, but, as expected, inthe opposite direction, indicating that toy recallsincrease the number of nonpositive articles andthat firm actions can amplify or attenuate theseeffects. Given the highly positive skew of coveragein the business press that we mention above, both adecrease in positive tenor and an increase in non-positive tenor lend credence to our arguments.

Endogeneity of recalls and actions. Because ofthe dynamic interaction between a firm, its com-petitors, and media coverage of their actions, wealso investigated whether endogeneity due to un-observed variables may have affected our analyses.

First, we conducted an interview with a CPSCofficial to learn more about the toy recall process.The official stated that recalls are not usually initi-ated by toy manufacturers; rather, after the CPSCbecomes aware of a defective toy, a representativecontacts the manufacturer to request a recall. Theofficial also indicated that toy companies alwayscomply, which makes toy recalls involuntary exog-enous shocks to the firms’ operations. We also ob-tained evidence for the exogeneity of product re-calls from the CPSC’s website (www.cpsc.gov) andrecent research on toy recalls (Beamish & Bapuji,2008; Freedman et al., 2012).

Further, to empirically support our theoreticaland qualitative evidence, we ran a two-stage Heck-man correction model. In the first stage, followingrecommendations for the selection of appropriatevariables (Bascle, 2008; Hamilton & Nickerson), weregressed prior firm wrongdoing, year dummies for1998 and 2007, and the number of toys recalled twoquarters after a focal quarter on the likelihood of afirm’s recalling toys. We selected the latter predic-tor variable, the number of recalled toys in t � 2, asour instrument in the first stage because it wasindicative of the firm’s propensity to recall toys inthe focal quarter, but it was not indicative of oursecond-stage dependent variable, the tenor of me-dia coverage (Bascle, 2008; Hamilton & Nickerson,2003). The selected instrument as well as priorlevels of firm wrongdoing significantly predictedthe probability of recalling a toy, but the year dum-mies did not. The inverse Mills ratio was not sig-nificant in the second stage. Taken together, ourfindings suggest that toy recalls are an exogenousvariable (Bascle, 2008; Mesquita & Brush, 2008;Tong, Reuer, & Peng, 2008).

Second, to test for the potential endogeneity oftechnical and ceremonial actions, we ran two two-stage Heckman correction models and two David-son-MacKinnon tests of exogeneity (using the Statacommand “dmexogxt”), the latter of which fol-lowed two-stage least squares estimation for eachtype of actions. In the first stage of each Heckman

estimation, we selected the number of technical(ceremonial) actions taken by a firm two quartersafter a recall as our instrument, as it was theoreti-cally associated with the firm’s propensity to taketechnical (ceremonial) actions but not with our de-pendent variable in the second stage, the tenor ofmedia coverage. We also included firm wrongdo-ing, firm assets, and a dummy for the fourth quarterin our first-stage model, as theoretically theyshould influence a firm’s propensity to engage intechnical or ceremonial actions. In both iterations,our selected instruments and firm assets signifi-cantly predicted propensity to engage in eithertechnical or ceremonial actions, but firm wrongdo-ing and the fourth-quarter dummy did not. InverseMills ratios in both estimations were not signifi-cant, indicating that both types of actions wereexogenous. With these robustness checks, we be-lieve we have ruled out the most likely explana-tions of endogeneity in our hypothesized relation-ships. However, as in all social science research,we cannot conclusively claim causality in ourrelationships or definitively rule out alternativeexplanations (Bascle, 2008; Hamilton & Nicker-son, 2003).

DISCUSSION

Theoretical Contributions

With this study, we make several contributionsto organizational theory. First, we extend recentresearch that has examined the influence of info-mediaries on firm-level and industry-level out-comes but has paid little attention to how infome-diaries themselves can be influenced by the firmsthey cover (Pollock et al., 2008; Westphal & Deep-house, 2011). We contribute to this research streamby explicating the strategies firms take to influencemedia reporting following their own or industrywrongdoing. Drawing from research that aims tounderstand how individuals make sense of a firm’sactions relative to its peers and industry (e.g., Greveet al., 2010; Porac & Thomas, 1990; Reger & Palmer,1996), we also provide insight into the processthrough which firms become associated with ordisassociated from the cognitive categories thatstakeholders construct from media coverage, de-pending on the source of wrongdoing.

Second, our focus on a specific industry allowedus to generate novel theoretical arguments thatdelve into the cognitive categorizations of technicaland ceremonial actions a firm may take afterwrongdoing. We base our categories of firm actionson how they are likely to be perceived by the mediaand stakeholders. Additionally, we drew on re-search in social psychology to explain how the two


types of actions affect the media and stakeholdersdifferently under different circumstances. By doingso, we extend crisis and impression managementresearch that has investigated how firm accountsaffect stakeholder responses but has not providedsociocognitive explanations for why specific ac-counts are more or less effective under differentcircumstances. By taking technical actions afterwrongdoing in an industry, a firm may remindstakeholders about its membership in the industry,and this may reinforce already-established negativeimpressions of the firm. Taking ceremonial actionsmay also backfire if the firm engaged in wrongdo-ing, because such tactics could be interpreted asself-serving and cynical. We believe that the theo-retical categories of actions developed in this studywill be useful in guiding future research beyond thedomain of social approval. For instance, classifyingfirm actions by how they are likely to reinforce ordeflect the attention of competitors may yield in-teresting results in competitive dynamics research(Smith et al., 1991). We also recognize that there areopportunities to develop more nuanced categoriesthat may more fully encompass the range of infor-mation subsidies available to firms when respond-ing to wrongdoing. We encourage future studies toexamine these possibilities in more depth.

Additionally, we extend research on the forma-tion of stakeholders’ judgments about firm’s actionsthat violate social expectations (e.g., Mishina et al.,2012; Skowronski & Carlston, 1987) by theoreti-cally investigating and providing support for therelative effectiveness of a firm’s actions in restoringthe positive tenor of media coverage after a viola-tion takes place. However, although our focus ontoy recalls allowed us to test previously unexam-ined dynamics in a specific industry, our studydoes not account for potential variance in the waydifferent stakeholder groups interpret firm actions.It is possible that while the media and consumersmay view product recalls in the toy industry asnegative actions that violate social norms and putcustomers at risk, other stakeholder groups, such asinvestors, may interpret such actions differently,perhaps as necessary quality control practices. Fur-thermore, given the sensitivities associated withtoy recalls and recalled toys’ potential to harm chil-dren, product recalls in other industries may not beviewed similarly, and they may not lead to thesame levels of decline in social approval. For ex-ample, it would be informative to know if recalls ofother consumer products, such as appliances orelectronics, would generate the same kind of mediacoverage and firm impression management strate-gies as we have seen here. Finally, different kindsof wrongdoing due to mismanagement, such as

earnings restatements, financial fraud, and envi-ronmental compliance failures (e.g., Mishina,Dykes, Block, & Pollock, 2010; Pfarrer, Smith,2008), may affect the media and stakeholder groupsdifferently and thus alter the effectiveness of firmactions in ways other than those we saw in thisstudy. For example, research has investigated theeffects of wrongdoing associated with competenceor capability and those associated with character orintegrity (e.g., Mishina et al., 2012; Skowronski &Carlston, 1987, 1989). It appears that in our case,the media and stakeholders perceive toy recalls asmore of an integrity violation than a competenceviolation. That is, they consider a toy recall a vio-lation of acceptable values and norms. Thus, wewould expect, and have theorized above, that neg-ative information would have greater weight andhigher diagnosticity for media and stakeholder per-ceptions in this context. Therefore, the effective-ness of a firm’s actions depends on how the firmrefocuses attention to a recall. Of course, the effectsof technical and ceremonial actions on the tenor ofmedia coverage and stakeholders’ perceptions maybe different for competence- and performance-based violations, such as a negative earnings sur-prise or a sudden drop in sales (Pfarrer et al., 2010).We encourage researchers interested in the linksamong firm wrongdoing, firm actions, and mediacoverage to examine the relationships tested herein other industries and among multiple stakeholdergroups.

Finally, by focusing on the dynamics occurringbetween a firm and its peers in a specific industry,we extend nascent organizational research on neg-ative spillovers that has begun to examine theserelationships (e.g., Barnett & Hoffman, 2008; Kang,2008; Kostova & Zaheer, 1999). Specifically, by uti-lizing sociocognitive arguments, we add explana-tory mechanisms to recent findings of direct nega-tive spillover effects. Additionally, our finding of asafety-in-numbers effect for a focal firm after a re-call provides new sociocognitive insights into in-dustry dynamics that can influence media coverageafter a negative event.

Empirical Contributions

Our first empirical contribution is in conductinga ten-year analysis of over 37,500 articles that re-flect changes in media coverage of firms and acontent analysis of over 5,500 press releases aboutfirm actions. Previous research in crisis and im-pression management has been largely qualitativeand case-based (e.g., Coombs, 2007; Elsbach, 2003),and it has not explicitly tested the interdependen-cies among an industry engaged in wrongdoing, a


firm’s actions following its own or its competitors’wrongdoing, and the media’s coverage of these dy-namics. To our knowledge, this is the first large-scale empirical investigation that tests how a focalfirm’s actions influence media coverage of wrong-doing and how this influence may vary dependingon the source of the wrongdoing. A second empir-ical contribution of our study is our use of firms’press releases as sources of information aboutfirms’ actions as well as our use of blogs as one ofthe media sources that influence and reflect socialperceptions. Our use of press releases allowed us toinvestigate intraindustry dynamics by observingannounced actions of large and small firms alike.Our use of blogs extends previous research on in-fomediaries that focuses only on the official pressand allows us to obtain a measure more closelyrelated to perceptions of certain stakeholder groups(Jin & Liu, 2010).

Finally, our use of advanced content analysistechniques to code affective content of articles andblog posts continues to extend recent organiza-tional research on social perceptions managementthat recognizes the importance of trying to open theblack box that is often present in strategy research(e.g., Pfarrer et al., 2010; Pollock & Rindova, 2003;Pollock et al., 2008; Short et al., 2010). We recog-nize the coarseness of our firm-quarter observationsand the potential limitations of our computer-aidedcoding processes, including the difficulty in recti-fying dissonant descriptions within a text (cf.Lamertz & Baum, 1998; Pollock et al., 2008). How-ever, continuous advances in content analysisshould permit scholars to gather increasingly largeramounts of text and code them with increasinglyfiner-grained techniques (Duriau et al., 2007).

Practical Implications

Our study also has practical implications formanagers in industries prone to actions that violatesocietal norms and that attract negative press, andour findings may have broader implications forcontexts in which wrongdoing is rarer. The inter-action effect between focal firm and industrywrongdoing in our study suggests that (1) low lev-els of industry wrongdoing provide a context formaximum penalties in the media when a firm vio-lates societal expectations, but high levels of indus-try wrongdoing may provide a safety net for offend-ers, and (2) firms that do not engage in wrongdoingnevertheless experience declines in the positivetenor of media coverage if their industry as a wholehas high levels of wrongdoing. Thus, firms maywant to work together to maintain lower levels ofindustry wrongdoing through higher safety and

quality standards to limit negative media attentionand exposure to regulators (cf. Pfarrer et al., 2008).

The second practical implication of our findingsis that when a firm or its competitors engage inwrongdoing, managers can use information subsi-dies to influence how the firm is covered in themedia. If a firm itself recalls a product, managersshould be aware that announcements of technicalactions are more likely to restore the positive tenorof media coverage than announcements of ceremo-nial actions. However, when the level of industrywrongdoing is high and, especially, when the focalfirm has not engaged in wrongdoing, announce-ments of ceremonial actions are a more effectivestrategy than announcements of technical actions.Taken together, our research findings suggest that“managing the message” is an important part of afirm’s strategy, and firms might benefit from a com-bination of technical and ceremonial actions, de-pending on whether the firm itself or its competi-tors engaged in wrongdoing.

REFERENCES

Abarbanell, J. S., & Lehavy, R. 2003. Biased forecasts orbiased earnings? The role of earnings management inexplaining apparent optimism and inefficiency inanalysts’ earnings forecasts. Journal of Accountingand Economics, 24: 675–723.

Ahmadjian, C. L., & Robinson, P. 2001. Safety in num-bers: Downsizing and the deinstitutionalization ofpermanent employment in Japan. AdministrativeScience Quarterly, 46: 622–654.

Aiken, L. S., & West, S. G. 1991. Multiple regression:Testing and interpreting interactions. ThousandOaks, CA: Sage.

Alvesson, M. 1990. Organization—From substance to im-age. Organization Studies, 11: 373–394.

Arellano, M., & Bond, S. 1991. Some tests of specificationfor panel data: Monte Carlo evidence and an appli-cation to employment equations. Review of Eco-nomic Studies, 58: 1991.

Ashforth, B. E., & Gibbs, B. W. 1990. The double-edge oforganizational legitimation. Organization Science,1: 177–194.

Bandura, A. 2001. Social cognitive theory of mass com-munication. Media Psychology, 3: 265–299.

Barnett, M. L., & Hoffman, A. J. 2008. Beyond corporatereputation: Managing reputational interdependence.Corporate Reputation Review, 11: 1–9.

Barnett, M. L., & King, A. A. 2008. Good fences makegood neighbors: A longitudinal analysis of an indus-try self-regulatory institution. Academy of Manage-ment Journal, 51: 1150–1170.


Bascle, G. 2008. Controlling for endogeneity with instru-mental variables in strategic management research.Strategic Organization, 6: 285–327.

Basdeo, D., Smith, K. G., Grimm, C., Rindova, V. P., &Derfus, P. 2006. The impact of market actions onfirm reputation. Strategic Management Journal, 27:1205–1219.

Baumeister, R. F., Bratslavsky, E., Finkenauer, C., &Vohs, K. D. 2001. Bad is stronger than good. Reviewof General Psychology, 5: 323–370.

Beamish, P. W., & Bapuji, H. 2008. Toys recalls andChina: Emotions vs. evidence. Management andOrganization Review, 4: 197–209.

Benoit, W. L. 1995. Accounts, excuses, and apologies: Atheory of image restoration strategies. Albany:State University of New York Press.

Brief, A. P., & Motowidlo, S. J. 1986. Prosocial organiza-tional behaviors. Academy of Management Review,11: 710–725.

Bromiley, P., & Marcus, A. 1989. The deterrent to dubi-ous corporate-behavior—Profitability, probabilityand safety recalls. Strategic Management Journal,10: 233–250.

Brown, L. 2001. A temporal analysis of earnings sur-prises: Profits versus losses. Journal of AccountingResearch, 39: 221–241.

Burgoon, M., Denning, V. P., & Roberts, L. 2002. Lan-guage expectancy theory. In J. P. Dillard & M. Pfau(Eds.), The persuasion book: Developments in the-ory and practice: 117–136. Thousand Oaks, CA:Sage.

Business Wire. 2001. Hasbro’s new G.I. Joe collectioncommemorates 60th anniversary of attack onPearl Harbor (May 22, Pawtucket, RI). AccessedJuly 14, 2011.

Business Wire. 2001. Hey kids—Come out and play atthe Hasbro/FAO Schwarz Play-a-Thon; children toraise money for the charity of their choice by doingwhat they do best—playing! (October 19, NewYork). Accessed July 14, 2011.

Business Wire. 2003. Greg Proops from the improv tele-vision series “Whose Line is It Anyway?” hosts“How Looney Can You Be?” competition at Toys RUs Times Square (October 24, New York). AccessedJuly 14, 2011.

Business Wire. 2006. Action products brings I DIG(TM)Dinosaurs to elementary school classrooms; na-tionl program introduces premier excavation lineto estimated 1.6m students & families (October 5,Orlando). Accessed July 14, 2011.

Business Wire. 2007. RC2 reports 2007 second quarteroperating results; revises 2007 outlook (August 1,Oak Brook, IL). Accessed July 14, 2011.

Cacioppo, J. T., & Gardner, W. L. 1999. Emotion. In J. T.Spence (Ed.), Annual review of psychology, vol. 50:191–214. Palo Alto, CA: Annual Reviews.

Carroll, C. E. 2009. The relationship between firms’ me-dia favorability and public esteem. Public RelationsJournal, 3(4): 1–32.

Carroll, C. E., & McCombs, M. 2003. Agenda-setting ef-fects of business news on the public’s images andopinions about major corporations. Corporate Rep-utation Review, 6: 36–46.

Chaiken, S., Liberman, A., & Eagley, A. H. 1989. Heuristicand systematic information processing within andbeyond the persuasion context. In J. S. Uleman &J. A. Bargh (Eds.), Unintended thought: 212–252.New York: Guilford.

Chatterjee, S., & Price, B. 1991. Regression analysis byexample (2nd ed.). New York: Wiley.

Chen, C. C., & Meindl, J. R. 1991. The construction ofleadership images in the popular press: The case ofDonald Burr and People Express. AdministrativeScience Quarterly, 36: 521–551.

Coombs, W. T. 1995. Choosing the right words: The de-velopment of guidelines for the selection of the “ap-propriate” crisis-response strategies. ManagementCommunication Quarterly, 8: 447–476.

Coombs, W. T. 2007. Protecting organization reputationsduring a crisis: The development and application ofsituational crisis communication theory. CorporateReputation Review, 10: 163–176.

Darr, E. D., Argote, L., & Epple, D. 1995. The acquisition,transfer, and depreciation of knowledge in serviceorganizations: Productivity in franchises. Manage-ment Science, 41: 1750–1762.

Deephouse, D. L. 2000. Media reputation as a strategicresource: An integration of mass communication andresource-based theories. Journal of Management,26: 1091–1112.

Deephouse, D. L., & Carter, S. M. 2005. An examinationof differences between organizational legitimacy andorganizational reputation. Journal of ManagementStudies, 42: 329–360.

Deephouse, D. L., & Heugens, P. P. M. A. R. 2009. Linkingsocial issues to organizational impact: The role ofinfomediaries and the infomediary process. Journalof Business Ethics, 86: 541–553.

Desai, V. M. 2011. Mass media and massive failures:Determining organizational efforts to defend fieldlegitimacy following crises. Academy of Manage-ment Journal, 54: 263–278.

Dukerich, J. M., & Carter, S. M. 2000. Distorted imagesand reputation repair. In M. Schultz, M. J. Hatch, &M. H. Larsen (Eds.), Expressive organization: Link-ing identity, reputation and the corporate brand:97–112. Oxford, U.K.: Oxford University Press.


Duriau, V. J., Reger, R. K., & Pfarrer, M. D. 2007. Acontent analysis of the content analysis literature inorganization studies: Research themes, data sources,and methodological refinements. OrganizationalResearch Methods, 10: 5–34.

Einwiller, S. A., Carroll, C. E., & Korn, K. 2010. Underwhat conditions do the news media influence cor-porate reputation? The roles of media dependencyand need for orientation. Corporate Reputation Re-view, 12: 299–315.

Elsbach, K. D. 1994. Managing organizational legitimacyin the California cattle industry: The constructionand effectiveness of verbal accounts. AdministrativeScience Quarterly, 39: 57–88.

Elsbach, K. D. 2003. Organizational perception manage-ment. In B. M. Staw & R. I. Sutton (Eds.), Research inorganizational behavior, vol. 25: 297–332. Green-wich, CT: JAI.

Elsbach, K. D., & Kramer, R. M. 1996. Members’ re-sponses to organizational identity threats: Encoun-tering and countering the business week rankings.Administrative Science Quarterly, 41: 442–476.

Elsbach, K. D., & Sutton, R. I. 1992. Acquiring organiza-tional legitimacy through illegitimate actions—Amarriage of institutional and impression manage-ment theories. Academy of Management Journal,35: 699–738.

Elsbach, K. D., Sutton, R. I., & Principe, K. E. 1998.Averting expected challenges through anticipatoryimpression management: A study of hospital billing.Organization Science, 9: 68–86.

Entman, R. M. 2007. Framing bias: Media in the distri-bution of power. Journal of Communication, 57:163–173.

Epstein, S. 1994. Integration of the cognitive and thepsychodynamic unconscious. American Psycholo-gist, 49: 709–724.

Festinger, L. 1957. A theory of cognitive dissonance.Stanford, CA: Stanford University Press.

Fiske, S. T., & Taylor, S. E. 2008. Social cognition: Frombrains to culture. New York: McGraw-Hill.

Floyd, K., Ramirez, A., Jr., & Burgoon, J. K. 1999. Expec-tancy violations theory. In L. K. Guerrero, J. A. De-Vito, & M. L. Hecht (Eds.), The nonverbal commu-nication reader: Classic and contemporaryreadings (2nd ed.): 437–444. Prospect Heights, IL:Wakeland.

Fombrun, C. J. 1996. Reputation: Realizing value fromthe corporate image. Boston: Harvard BusinessSchool Press.

Fombrun, C. J., & Shanley, M. 1990. What’s in a name?Reputation building and corporate strategy. Acad-emy of Management Journal, 33: 233–256.

Forgas, J. P. 1995. Mood and judgment: The affect infu-sion model (AIM). Psychological Bulletin, 117: 39–66.

Freedman, S., Kearney, M., & Lederman, M. 2012. Prod-uct recalls, imperfect information, and spillover ef-fects: Lessons from the consumer response to the2007 toy recalls. Review of Economics and Statis-tics, 94: 499–516.

Godfrey, P. C., Merrill, C. B., & Hansen, J. M. 2009.Institutional stakeholders and the insurance value ofCSR. The relationship between corporate social re-sponsibility and shareholder value: An empiricaltest of the risk management hypothesis. StrategicManagement Journal, 30: 425–445.

Greifeneder, R., Bless, H., & Pham, M. T. 2011. When dopeople rely on affective and cognitive feelings injudgment? A review. Personality and Social Psy-chology Review, 15: 107–141.

Greve, H. R., Palmer, D., & Pozner, J. E. 2010. Organiza-tions gone wild: The causes, processes, and conse-quences of organizational misconduct. In J. P. Walsh& A. P. Brief (Eds.), Academy of Management an-nals, vol. 4: 53–107. Essex, U.K.: Routledge.

Halaby, C. N. 2004. Panel models in sociological re-search: Theory into practice. In K. Cook & J. Hagan(Eds.), Annual review of sociology, vol. 30: 507–544. Palo Alto, CA: Annual Reviews.

Hamilton, B. H., & Nickerson, J. A. 2003. Correcting forendogeneity in strategic management research. Stra-tegic Organization, 1: 51–78.

Hannan, T., & Freeman, J. 1984. Structural inertia andorganizational change. American Sociological Re-view, 49: 149–164.

Hastie, R., & Dawes, R. M. 2001. Rational choice in anuncertain world. The psychology of judgment anddecision making. Thousand Oaks, CA: Sage.

Hoetker, G. 2007. The use of logit and probit models instrategic management research: Critical issues. Stra-tegic Management Journal, 28: 331–343.

Houser, D., & Wooders, J. 2006. Reputation in auctions:Theory, and evidence from eBay. Journal of Eco-nomics and Management Strategy, 15: 353–369.

Janis, I. L., & Fadner, R. H. 1943. A coefficient of imbal-ance for content analysis. Psychometrika, 8: 105–119.

Jensen, M. 2006. Should we stay or should we go? Ac-countability, status anxiety, and client defections.Administrative Science Quarterly, 51: 97–128.

Jin, Y., & Liu, B. F. 2010. The blog-mediated crisis com-munication model: Recommendations for respond-ing to influential external blogs. Journal of PublicRelations Research, 22: 429–455.

Johnson, M. E. 2001. Learning from toys: Lessons in


managing supply chain risk from the toy industry.California Management Review, 43(3): 106–124.

Jones, E. E., & McGillis, D. 1976. Correspondent infer-ences and the attribution cube: A comparative reap-praisal. In J. H. Ickes & R. F. Kidd (Eds.), New direc-tions in attribution research, vol. 1: 389–420.Hillsdale, NJ: Erlbaum.

Jonsson, S., Greve, H. R., & Fujiwara-Greve, T. 2009.Undeserved loss: The spread of legitimacy loss toinnocent organizations in response to reported cor-porate deviance. Administrative Science Quar-terly, 54: 195–228.

Kang, E. 2008. Director interlocks and spillover effects ofreputational penalties from financial reportingfraud. Academy of Management Journal, 51: 537–555.

Kennedy, M. T. 2008. Getting counted: Markets, media,and reality. American Sociological Review, 73:270–295.

Kirsch, D., Goldfarb, B., & Gera, A. 2009. Form or sub-stance: The role of business plans in venture capitaldecision making. Strategic Management Journal,30: 487–515.

Kostova, T., & Zaheer, S. 1999. Organizational legitimacyunder conditions of complexity: The case of the mul-tinational enterprise. Academy of Management Re-view, 24: 64–81.

Krippendorff, K. 2004. Content analysis: An introduc-tion to its methodology (2nd ed.). Thousand Oaks,CA: Sage.

Krishnan, R., Martin, X., & Noorderhaven, N. G. 2006.When does trust matter to alliance performance?Academy of Management Journal, 49: 894–917.

Lamertz, K., & Baum, J. A. C. 1998. The legitimacy oforganizational downsizing in Canada: An analysis ofexplanatory media accounts. Canadian Journal ofAdministrative Sciences, 15: 93–107.

Lange, D., Lee, P. M., & Dai, Y. 2011. Organizationalreputation: A review. Journal of Management, 37:153–184.

Lingle, J. H., Geva, N., Ostrom, T. M., Leippe, M. R., &Baumgardner, M. H. 1979. Thematic effects of per-son judgments on impression organization. Journalof Personality and Social Psychology, 37: 674–687.

Love, E. G., & Kraatz, M. 2009. Character, conformity, orthe bottom line? How and why downsizing affectedcorporate reputation. Academy of ManagementJournal, 52: 314–335.

Luo, Y. 2008. A strategic analysis of product recalls: Therole of moral degradation and organizational control.Management and Organization Review, 4: 183–196.

Lyon, T. P., & Maxwell, J. W. 2011. Greenwash: Corporate

environmental disclosure under threat of audit.Journal of Economics and Management Strategy,20: 3–41.

Marcus, A. A., & Goodman, R. S. 1991. Victims andshareholders: The dilemmas of presenting corporatepolicy during a crisis. Academy of ManagementJournal, 14: 281–305.

Martins, L. L. 2005. A model of the effects of reputationalrankings on organizational change. OrganizationScience, 16: 701–720.

McCarthy, J. D., McPhail, C., & Smith, J. 1996. Images ofprotest: Dimensions of selection bias in media cov-erage of Washington demonstrations, 1982 and 1991.American Sociological Review, 61: 478–499.

McCombs, M. E. 1981. The agenda setting approach. In D.Nimmo & K. R. Sanders (Eds.), Handbook of politi-cal communication: 121–140. Newbury Park, CA:Sage.

McCombs, M., Llamas, J. P., Lopez-Escobar, E., & Rey, F.1997. Candidate images in special elections: Second-level agenda setting effects. Journalism and MassCommunication Quarterly, 74: 703–717.

McCombs, M., & Shaw, D. 1972. The agenda-setting func-tion of mass media. Public Opinion Quarterly, 36:176–187.

Mercer, M. 2005. The fleeting effects of disclosure forth-comingness on management’s reporting credibility.Accounting Review, 80: 723–744.

Mesquita, L. F., & Brush, T. H. 2008. Untangling safe-guard and production coordination effects in long-term buyer-supplier relationships. Academy ofManagement Journal, 51: 785–807.

Mishina, Y., Block, E. S., & Mannor, M. J. 2012. The pathdependence of organizational reputation: How socialjudgment influences assessments of capability andcharacter. Strategic Management Journal, 33: 459–477.

Mishina, Y., Dykes, B. J., Block, E. S., & Pollock, T. G.2010. Why “good” firms do bad things: The effects ofhigh aspirations, high expectations, and prominenceon the incidence of corporate illegality. Academy ofManagement Journal, 53: 701–722.

Morrison, E. W., & Robinson, S. L. 1997. When employ-ees feel betrayed: A model of how psychologicalcontract violation develops. Academy of Manage-ment Review, 22: 226–256.

Near, J. P., Rehg, M. T., Van Scotter, J. R., & Miceli, M. P.2004. Does type of wrongdoing affect the whistle-blowing process? Business Ethics Quarterly, 14:219–242.

Parducci, A. 1968. The relativism of absolute judgments.Scientific American, 219: 84–90.


Pennebaker, J. W., Booth, R. J., & Francis, M. E. 2007.Linguistic inquiry and word count (LIWC): A com-puterized text analysis program. Austin, TX: LIWC.net.

Pew Research Center. 2008. www.pewresearch.org.

Pezzo, S. P., Pezzo, M. V., & Stone, E. R. 2006. The socialimplications of planning: How public predictionsbias future plans. Journal of Experimental SocialPsychology, 42: 221–227.

Pfarrer, M. D., DeCelles, K. A., Smith, K. G., & Taylor,M. S. 2008. After the fall: Reintegrating the corruptorganization. Academy of Management Review, 33:730–749.

Pfarrer, M. D., Pollock, T. G., & Rindova, V. P. 2010. Atale of two assets: The effects of firm reputation andearnings surprises on investors’ reactions. Academyof Management Journal, 53: 1131–1152.

Pfarrer, M. D., Smith, K. G., Bartol, K. M., Khanin, D. M.,& Zhang, X. 2008. Coming forward: The effects ofsocial and regulatory forces on the voluntary restate-ment of earnings subsequent to wrongdoing. Orga-nization Science, 19: 386–403.

Planalp, S., & Fitness, J. 1999. Thinking/feeling aboutsocial and personal relationships. Journal of Socialand Personal Relationships, 16: 731–750.

Pollock, T. G., & Rindova, V. P. 2003. Media legitima-tion effects in the market for initial public offer-ings. Academy of Management Journal, 46: 631–642.

Pollock, T. G., Rindova, V. P., & Maggitti, P. G. 2008. Marketwatch: Information and availability cascades amongthe media and investors in the US IPO market. Acad-emy of Management Journal, 51: 335–358.

Porac, J. F., & Thomas, H. 1990. Taxonomic mental mod-els in competitor definition. Academy of Manage-ment Review, 15: 224–240.

PR Newswire. 2002. Independent monitoring councilcompletes audits of Mattel manufacturing facili-ties in Indonesia, Malaysia and Thailand (Novem-ber 15, New York). Accessed July 14, 2011.

PR Newswire. 2006. Leapfrog recalls to repair chil-dren’s activity centers due to arm entrapment haz-ard (September 7, Washington DC). Accessed July14, 2011.

Range, L. M., Menyhert, A., Walsh, M. L., Hardin, K. N.,Ellis, J. B., & Craddick, R. 1991. Letters of recommen-dation: Perspectives, recommendations, and ethics.Professional Psychology: Research and Practice,22: 389–392.

Rao, H. 1994. The social construction of reputation: Cer-tification contests, legitimation, and the survival oforganizations in the American automobile industry:1895–1912. Strategic Management Journal, 15: 29–44.

Reger, R. K., & Palmer, T. B. 1996. Managerial categori-zation of competitors: Using old maps to navigatenew environments. Organization Science, 7: 22–39.

Rhee, M., & Haunschild, P. R. 2006. The liability of goodreputation: A study of product recalls in the U.S.automobile industry. Organization Science, 17:101–117.

Richey, M. H., Koenigs, R. J., Richey, H. W., & Fortin, R.1975. Negative salience in impressions of character:Effects of unequal proportions of positive and nega-tive information. Journal of Social Psychology, 97:233–241.

Rindova, V. P., Pollock, T. G., & Hayward, M. L. A. 2006.Celebrity firms: The social construction of marketpopularity. Academy of Management Review, 31:50–71.

Rogers, E. M., Dearing, J. W., & Bregman, D. 1993. Theanatomy of agenda-setting research. Journal of Com-munication, 43: 68–84.

Salancik, G. R., & Meindl, J. R. 1984. Corporate attribu-tions as strategic illusions of management control.Administrative Science Quarterly, 29: 238–254.

Scheufele, D. A. 1999. Framing as a theory of mediaeffects. Journal of Communication, 49: 103–122.

Scott, W. R. 1995. Institutions and organizations. Thou-sand Oaks, CA: Sage.

Scott, W. R. 2003. Organizations: Rational, natural, andopen systems. Upper Saddle River, NJ: Prentice Hall.

Semadeni, M., Cannella, A. A., Fraser, D. R., & Lee, D. S.2008. Fight or flight: Managing stigma in executivecareers. Strategic Management Journal, 29: 557–567.

Short, J. C., Broberg, J. C., Cogliser, C. C., & Brigham, K. H.2010. Construct validation using computer-aidedtext analysis (CATA): An illustration using entrepre-neurial orientation. Organizational Research Meth-ods, 13: 320–347.

Siskind, G. 1966. Mine eyes have seen a host of angels.American Psychologist, 21: 804–806.

Skinner, D., & Sloan, R. 2002. Earnings surprises, growthexpectations, and stock returns or don’t let an earn-ings torpedo sink your portfolio. Review of Account-ing Studies, 7: 289–312.

Skowronski, J. J., & Carlston, D. E. 1989. Negativity andextremity biases in impression formation: A reviewof explanations. Psychological Bulletin, 105: 131–142.

Skowronski, J., & Carlston, D. 1987. Social judgmentand social memory: The role of cue diagnosticityin negativity, positivity, and extremity biases.Journal of Personality and Social Psychology, 52:689–699.


Smith, K. G., Grimm, C., Gannon, M., & Chen, M. J. 1991.Organizational information processing, competitiveresponses and performance in the U.S. domestic air-line industry. Academy of Management Journal,34: 60–85.

Suchman, M. C. 1995. Managing legitimacy: Strategicand institutional approaches. Academy of Manage-ment Review, 20: 571–610.

Sullivan, B. N., Haunschild, P., & Page, K. 2007. Organi-zations non gratae? The impact of unethical corpo-rate acts on interorganizational networks. Organiza-tion Science, 18: 55–70.

Sutton, R. I., & Galunic, D. C. 1996. Consequences ofpublic scrutiny for leaders and their organizations.In B. M. Staw & L. L. Cummings (Eds.), Research inorganizational behavior, vol. 18: 201–250. Green-wich, CT: JAI Press.

Taylor, S. E. 1991. Asymmetrical effects of positive andnegative events: The mobilization-minimization hy-pothesis. Psychological Bulletin, 110: 67–85.

Tetlock, P. C., Saar-Tsechansky, M., & Macskassy, S.2008. More than words: Quantifying language tomeasure firms’ fundamentals. Journal of Finance,63: 1437–1467.

Thompson, J. D. 1967. Organizations in action: Socialscience bases of administration theory. New York:McGraw-Hill.

Tong, T. W., Reuer, J. J., & Peng, M. W. 2008. Interna-tional joint ventures and the value of growth op-tions. Academy of Management Journal, 51: 1014–1029.

Wade, J. B., Porac, J. F., Pollock, T. G., & Graffin, S. D.2006. The burden of celebrity: The impact of CEOcertification contests on CEO pay and performance.Academy of Management Journal, 49: 643–660.

Wartick, S. L. 1992. The relationship between intensemedia exposure and change in corporate reputation.Business and Society, 31(April): 33–49.

Westphal, J. D., & Bednar, M. K. 2008. The pacification ofinstitutional investors. Administrative ScienceQuarterly, 53: 29–72.

Westphal, J. D., & Deephouse, D. L. 2011. Avoiding badpress: Interpersonal influence in relations betweenCEOs and journalists and the consequences for pressreporting about firms and their leadership. Organi-zation Science, 22: 1061–1086.

Westphal, J. D., & Zajac, E. J. 1998. The symbolic man-agement of stockholders: Corporate governance re-forms and shareholder reactions. AdministrativeScience Quarterly, 43: 127–153.

Yu, T., Sengul, M., & Lester, R. H. 2008. Misery lovescompany: The spread of negative impacts resultingfrom an organizational crisis. Academy of Manage-ment Review, 33: 452–472.

Zuckerman, E. W. 1999. The categorical imperative: Se-curities analysts and the illegitimacy discount.American Journal of Sociology, 104: 1398–1438.

Zyglidopoulos, S. C. 2001. The impact of accidents onfirms’ reputation for social performance. Businessand Society, 40: 416–441.

Anastasiya Zavyalova ([email protected])is an assistant professor in the Jones Graduate School ofBusiness at Rice University. She received her Ph.D. fromthe University of Maryland. Her research focuses on theprocesses through which firms damage and restore socialapproval assets, such as reputation and celebrity. She isspecifically interested in the role of media in influencingsocial perceptions about firms in light of negative disrup-tive events.

Michael D. Pfarrer ([email protected]) is an assistantprofessor in the Terry College of Business at the Univer-sity of Georgia. He received his Ph.D. from the Universityof Maryland. His research focuses on external percep-tions of firm behavior and how firms manage these inter-pretations to create value. His specific interests includepositive and negative social evaluations (e.g., firm celeb-rity, legitimacy, reputation, and stigma), impression andcrisis management, media accounts, corporate gover-nance, and the role of business in society.

Rhonda K. Reger ([email protected]) is an associ-ate professor of strategic management and entrepreneur-ship at the University of Maryland, College Park. (Prof.Reger will join the University of Tennessee faculty in Jan-uary 2013.) She received her Ph.D. in strategic managementfrom the University of Illinois at Urbana-Champaign. Hercurrent research interests include cognitive approaches tocompetitive and reputation dynamics, organizational iden-tity, and nonmarket strategies.

Debra L. Shapiro ([email protected]) is theClarice Smith Professor of Management at the Universityof Maryland, formerly the Willard J. Graham Distin-guished Professor of Management at UNC-Chapel Hill.She earned her Ph.D. in organization behavior from theJ.L. Kellogg Graduate School of Management at North-western University. Her research centers on how to man-age conflict, including change resistance and perceivedinjustice, in organizations.


MANAGING THE MESSAGE: THE EFFECTS OF … · THE EFFECTS OF FIRM ACTIONS AND INDUSTRY SPILLOVERS ON...

Documents

Transcript of MANAGING THE MESSAGE: THE EFFECTS OF … · THE EFFECTS OF FIRM ACTIONS AND INDUSTRY SPILLOVERS ON...