Consumer response to corporate irresponsible behaviorMoral emotions and virtues

When a company does something genuinely wrong, such as dumping toxins, exploiting workers, or lying to customers, what actually moves people to act? Not to feel bad or grumble quietly, but to march, boycott, or tell everyone they know. Hold that question for a moment. Grappi, Romani, and Bagozzi found that it's not outrage in general. It's three very specific emotions, and they operate differently depending on what kind of wrongdoing was committed. The gap they were filling is a real one. Most corporate social responsibility research has focused on the positive side of the ledger: firms behave well and consumers reward them. When researchers did look at negative consumer responses, they tended to measure things like willingness to pay or purchase intentions — individualistic and transactional outcomes. Grappi and colleagues argued that when companies behave badly, consumers don't just quietly change their shopping habits. They talk, they organize, and they protest. These collective and interpersonal responses had been largely ignored by the literature. So the authors built a framework to explain two specific behaviors: negative word of mouth, which involves saying bad things about a firm, recommending against it, and actively discrediting it, and protest, which is a broader category that includes blogging against the company, picketing, joining boycotts, taking legal action, and participating in collective movements. These are meaningfully different things, and understanding what drives each one is the core puzzle of this paper. The emotional engine the authors identify is a cluster of three moral emotions: contempt, anger, and disgust. In moral psychology, this triad is known as the CAD triad, which stands for contempt, anger, and disgust, and researchers have long noted that these three emotions tend to travel together. They share a common appraisal structure: all three arise when someone perceives a violation of moral norms, all three orient a person against the offending party, and all three carry an urge to distance from or confront the wrongdoer. Prior work by Rozin, Shaver, and others treated the three as theoretically distinct but also noted that they cluster empirically in real situations. Grappi and colleagues tested whether that clustering holds in the consumer context by running a second-order confirmatory factor analysis, which is a statistical technique that estimates whether a set of specific measures can be explained by a single higher-order factor, across four samples. The results were clear: model fit was good, with a comparative fit index of 0.97 and a root mean square error of approximation of 0.07, and the nine item loadings at the first-order level ranged from 0.91 to 0.97. The three loadings connecting the higher-order factor to contempt, anger, and disgust individually were 0.97, 0.98, and 0.90. That's an extraordinarily tight structure. Based on this, the authors created a single composite variable they call CAD, which becomes the central mediator in everything that follows. Importantly, fear loaded separately and was not part of the CAD cluster, which matters later when the authors test whether fear plays a similar role. It does not. Now, here's where the framework gets more nuanced. CAD is the emotional engine, but it doesn't operate alone. Grappi, Romani, and Bagozzi introduce what they call other-regarding virtues, which are character dispositions like justice, beneficence, equality, and communal cooperation. What makes these virtues "other-regarding" is that they orient a person's moral attention outward — toward the community, fairness, and collective welfare — rather than inward toward personal grievance. The authors argue, and find, that these virtues don't just add to the emotional response; they regulate how emotions translate into action. Two consumers can feel the same level of disgust, but the one who also holds strong commitments to justice and communal cooperation is far more likely to convert that disgust into protest. To test all of this, the authors recruited 280 adult Italian shoppers and exposed each one to a single realistic scenario describing a corporate wrongdoing. One scenario featured a fictional confectioner called Dark Chocolate, which was caught using and abusing child labor on cocoa plantations. This scenario presented an ethical transgression and a violation of fundamental moral norms. The other scenario featured a fictional Big Retailer entering a town, demolishing a beloved community center, and threatening local shopkeepers. This scenario represented a social transgression, as it harmed community norms and local institutions. A pretest confirmed these landed differently: on a seven-point seriousness scale, the ethical scenario averaged six point sixty-four, while the social scenario averaged four point ninety-five. Participants also correctly categorized the two scenarios as more ethical and more social in nature, respectively. The manipulation worked. The statistical framework the authors used is called moderated mediation. The idea is that one variable — CAD — carries the effect of the manipulation through to consumer behavior, but the strength of that pathway depends on a third factor, other-regarding virtues. Virtues moderate the mediation. Moreover, they moderate it in different places depending on whether the transgression was ethical or social. The authors refer to these as Case A and Case B. For ethical transgressions, like that of Dark Chocolate and child labor, the manipulation had a significant impact on CAD. The coefficient was negative four point twenty-five, with a t-statistic of negative nineteen point forty-nine. The manipulation provoked strong contempt, anger, and disgust almost automatically. However, whether that emotional response then converted into negative word of mouth or protest depended on how strongly consumers held other-regarding virtues. The interaction of CAD and virtues predicted negative word of mouth with a coefficient of zero point twenty-five and protest with a coefficient of zero point forty-one. The bootstrapped conditional indirect effects tested at the mean and at one standard deviation above and below the mean of virtues were all significantly different from zero, and the indirect effect through CAD was substantially larger when virtues were high. For social transgressions, like the Big Retailer and the demolished community center, the pattern flipped. Here, the manipulation did not reliably produce CAD on its own. Instead, CAD was only evoked among consumers who already held strong other-regarding virtues. The interaction of the manipulation and virtues predicted CAD with a coefficient of negative zero point fifty-seven. Once CAD was felt, it strongly predicted negative word of mouth with a coefficient of zero point sixty-nine and protest with a coefficient of zero point twenty-six. The direct effects of the social manipulation on outcomes, bypassing CAD, were not significant. Virtues entered earlier in the chain for social transgressions, determining whether consumers felt moral emotions at all, while for ethical transgressions, they entered later, determining whether emotions led to action. Fear, the secondary mediator the authors tested, did not show this pattern. Its interaction terms were generally nonsignificant in the moderated mediation tests. In one model, it had a small direct association with protest in the social condition, but it did not produce the same conditional indirect effects as CAD. This finding reinforces that the CAD triad is doing specific, identifiable work — it is not a proxy for generalized negative affect. What this means in practice is that talking and acting have different drivers. Negative word of mouth flows fairly reliably from CAD across both transgression types. Once consumers feel contempt, anger, and disgust, they talk. Protest is more conditional; it depends on who those consumers are, specifically whether they hold strong other-regarding virtues. A firm worried about reputation damage through spreading negative stories faces a different problem than a firm worried about organized public protest. The emotional substrate may be the same, but the population of consumers who will move from feeling to organizing is a subset shaped by moral character. For managers, the framework carries a pointed warning. Ethical transgressions, which directly violate human dignity, are judged nearly two points higher on a seven-point severity scale than social transgressions. They produce CAD almost automatically, without virtues needing to unlock it first. The BP Gulf spill, which the authors mention cost nearly forty billion dollars and briefly erased half the company's market value, is the kind of event this model is built to explain. Monitor consumer emotional reactions early, not just attitudes and purchase intentions. Address the moral dimensions of a crisis, not just the practical ones. And understand that among consumers who hold strong commitments to justice and communal cooperation, felt moral emotions will reliably become organized action. Theoretically, the paper's contribution is in combining moral emotions with character virtues to predict collective consumer behavior, yielding a richer account than either element alone. Moral emotions tell you the direction of the response, while virtues tell you who will act and how far. The authors flag real limitations, stating that scenarios and language-based emotion measures are artificial, and they call for more naturalistic settings, physiological measures of emotion, and the study of a broader range of corporate contexts. Those are fair caveats. However, the architecture they've built — emotions plus virtues, moderated mediation, and ethical versus social transgression — provides researchers and practitioners with a genuinely more precise vocabulary for what happens when companies get things badly wrong. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.