Corporate Hypocrisy

Wagner et al. (Reference Wagner, Lutz and Weitz2009) proposed that when a firm’s CSR engagement is contrary to its stated standards of social responsibility, this can give rise to perceptions of corporate hypocrisy and negative attitudes toward the firm. In a similar vein, Scheidler et al. (Reference Scheidler, Edinger-Schons, Spanjol and Wieseke2019) reported a tendency for corporate hypocrisy to be perceived when firms’ CSR engagement implies an inconsistency between treatment of external and internal stakeholders. Indeed, reinforcing the connection to negative attitudes, their findings associate employees’ perceptions of corporate hypocrisy with emotional exhaustion and higher staff turnover.

However, building on the concept of organized hypocrisy (Brunsson, Reference Brunsson2002), Christensen, Morsing, and Thyssen (Reference Christensen L, Morsing and Thyssen2013) and Cho, Laine, Roberts, and Rodrigue (Reference Cho, Laine, Roberts and Rodrigue2015) challenged the notion that corporate hypocrisy is always detrimental. For example, Christensen et al. (Reference Christensen L, Morsing and Thyssen2013) argued that CSR communication can be performative, announcing firms’ ideals and aspirations. Thus, it does not have to be a neutral tool to describe or report a firm’s CSR practice. Given that CSR standards and goals are continuously expanding and evolving, aspirational CSR communication can be “an important resource for social change, even when organizations do not fully live up to their aspirations” (Christensen et al., Reference Christensen L, Morsing and Thyssen2013: 372). Cho et al. (Reference Cho, Laine, Roberts and Rodrigue2015) pointed out that organizations often face irreconcilable societal and institutional demands. Thus, corporate hypocrisy is needed to establish and maintain legitimacy. As a result, firms’ sustainability reports tend to be symbolic rather than substantive (Cho et al., Reference Cho, Laine, Roberts and Rodrigue2015). These studies suggest that corporate hypocrisy can benefit firms.

As discussed above, such mixed results arise partly because the extant literature fails to differentiate between inconsistency that arises from a firm lacking a genuine CSR motive and using its CSR engagement to claim undeserved moral benefit(s) and inconsistency performed as a result of lack of ability and/or resources. Building on the SCM (Cuddy et al., Reference Cuddy, Fiske and Glick2008; Fiske et al., Reference Fiske, Cuddy, Glick and Xu2002), this article argues that individuals judge firms in the same way they judge other social targets, namely, on perceived warmth and competence (Aaker et al., Reference Aaker, Vohs and Mogilner2010; Bolton & Mattila, Reference Bolton and Mattila2015; Cuddy et al., Reference Cuddy, Fiske and Glick2008; Kervyn et al., Reference Kervyn, Fiske and Malone2012). We further argue that warmth underlies individuals’ judgments of whether a firm’s CSR initiatives are driven by genuine motives. The next section discusses this in detail.

Prior CSR, Firm Warmth, and CSR Motives

The SCM identifies warmth and competence as the two universal dimensions underlying perceptions of different social targets, such as different groups of people (Cuddy et al., Reference Cuddy, Fiske and Glick2008). The warmth dimension comprises traits such as trustworthiness, kindness, and sincerity. It “represents an accommodating orientation that profits others more than the self” (Cuddy et al., Reference Cuddy, Fiske and Glick2008; 63). The competence dimension captures the efficacy of relevant social targets in executing on their intentions, with those targets that are able to carry out their intentions being considered as capable, intelligent, and skillful (Cuddy et al., Reference Cuddy, Fiske and Glick2008). Firms are often anthropomorphized into human beings possessing thoughts, desires, intentions (Knobe & Prinz, Reference Knobe and Prinz2008), character (Godfrey, Reference Godfrey2005), goals, and identities (Pfeffer, Reference Pfeffer, Cummings and Staw1981). Consistent with this, the SCM has been used to explain perceptions of nonhuman entities such as firms (Aaker et al., Reference Aaker, Vohs and Mogilner2010; Bolton & Mattila, Reference Bolton and Mattila2015; Shea & Hawn, Reference Shea and Hawn2019). For example, Aaker et al. (Reference Aaker, Vohs and Mogilner2010) found that not-for-profit firms were considered as warmer but less competent. In contrast, profit-driven firms were considered more competent but less warm.

Relating to our research, Carroll (Reference Carroll1991) argued that CSR is a morally colored concept, reflecting a firm’s concern for its stakeholders’ moral rights that goes beyond legal requirements. In other words, CSR initiatives can reflect a firm’s moral value in accommodating stakeholders’ welfare that is beyond the economic interests of the firm and what is required by law (Carroll, Reference Carroll1991). Therefore, such activities can be considered caring and trustworthy (Aaker et al., Reference Aaker, Vohs and Mogilner2010; Bolton & Mattila, Reference Bolton and Mattila2015). Of course, firms differ in their commitment to CSR initiatives. Some firms show a great and consistent commitment of resources to their CSR initiatives—a strong prior record of CSR. In contrast, other firms show a commitment to CSR that is relatively insubstantial and/or inconsistent over time, depending on a variety of factors such as their profits—a weak prior record of CSR.

Given that CSR activities tend to direct the firm’s resources to the improvement of social welfare by addressing social and/or environmental issues, they are well-suited to promote inferences of firm warmth. As Cuddy, Glick, and Beninger (Reference Cuddy, Glick and Beninger2011) pointed out, “warmth is inferred from actions that appear to serve self-interest versus others’ interests” (Cuddy et al., Reference Cuddy, Glick and Beninger2011: 76), such that a strong record of contributions to the welfare of others tends to promote warmth. Thus, applied to evaluations of firms, SCM predicts greater warmth is inferred for firms with a strong prior record of CSR. Indeed, Shea and Hawn (Reference Shea and Hawn2019) demonstrated that individuals perceive firms engaging in CSR as warm, which in turn influences their purchase intentions and/or reputational judgments toward the firms. This is supported by Gao and Mattila (Reference Gao and Mattila2014). Their research suggested that consumers are more likely to express satisfaction towards green (vs. non-green) hotels, partly because green hotels are perceived to be warmer.

We now turn to individuals’ perceptions of the motives that underlie a firm’s CSR activities and the tendencies for these to be influenced by the strength of a firm’s prior record of CSR and firm warmth. With regard to the connection between warmth and inferred motives, Cuddy et al. (Reference Cuddy, Glick and Beninger2011: 74) argued that “warmth judgments affect how much we trust versus doubt others’ motives.” This reflects a broader tendency for warmth to have a primary role in evaluations, such that “warmth judgments are made more quickly than competence judgments and have a greater impact on overall attitudes toward others” (Cuddy et al., Reference Cuddy, Glick and Beninger2011: 76). This point has received empirical support in the context of the evaluations of firms, as Gao and Mattila (Reference Gao and Mattila2014) and Shea and Hawn (Reference Shea and Hawn2019) demonstrated that perceptions of firm warmth mediate individuals’ judgments of, and behavioral intentions toward, the evaluated firms. Such primacy for warmth in the SCM, such that it forms a basis from which other evaluations of the target are made, predicts that firm warmth influences perceptions of corporate motives. In particular, a perception of greater warmth is expected to increase the likelihood that an individual infers that a firm is motivated to undertake CSR activities by a genuine regard for social welfare.

Thus, alongside the prediction that a strong record of CSR promotes perceptions of firm warmth, this effect of warmth on perceived motive implies that a firm’s prior record of CSR affects individuals’ inferences of firm’s motives. Specifically, a stronger prior record of CSR promotes perceptions that the firm is motivated by a genuine regard for social welfare. Moreover, this effect is mediated by firm warmth. In making this argument, we echo a prominent theme in the extant literature: that individuals’ corporate evaluations depend on the strength of a firm’s prior record of CSR (Luo & Bhattacharya, Reference Luo and Bhattacharya2009). Indeed, this notion has received some empirical support. For example, Du, Bhattacharya, and Sen (Reference Du, Bhattacharya and Sen2010) argued that individuals’ inferences about the motives underlying a firm’s CSR initiatives depend on its perceived commitment to those initiatives. Specifically, a greater and more consistent commitment promotes the inference that CSR activities are motivated by a genuine regard for social welfare. In addition, Skarmeas and Leonidou (Reference Skarmeas and Leonidou2013) found that an inconsistent commitment to CSR leads individuals to suspect that a firm is motivated by a desire to manipulate perceptions through window-dressing.

Prior CSR, CSR Motives, and Corporate Hypocrisy

We now turn to perceptions of hypocrisy that can arise in the wake of a CSI event, and apply the notion of moral licensing for this purpose. The moral licensing theory considers the tendency for observers to license transgressions (i.e., “reduce their condemnation of morally dubious behavior or the actor who committed it in light of the actor’s prior good deeds” [Effron & Monin, Reference Effron and Monin2010: 1618]). Within this view, licensing can arise for two reasons: moral credits and moral credentials. For our purposes, the first of these implies that a strong prior record of CSR would provide a firm moral credits in the eyes of individuals, whereas a CSI event would bring moral debits (Ormiston & Wong, Reference Ormiston and Wong2013). Thus, a firm would expectably avoid negative assessments in the wake of a CSI event as long as any moral debits are more than balanced by the moral credits it has previously accumulated through its prior record of CSR. The second reason for licensing—moral credentials—implies that a firm’s prior CSR might change the way individuals construe a CSI event. That is, a firm’s strong prior CSR record can license a CSI event by causing individuals to differently construe the firm’s role in that event, such that the firm is judged to be less culpable. Effron and Monin (Reference Effron and Monin2010) summarized these two reasons for licensing as follows: “If moral credits are like a deposit that allows one to ‘purchase’ the right to transgress, moral credentials provide a lens through which subsequent behavior is construed more favorably” (Effron & Monin, Reference Effron and Monin2010: 1619).

In this way, the moral licensing theory helps us to understand how past good deeds can mitigate perceptions of hypocrisy by providing context for the interpretation of subsequent bad deeds (Effron & Monin, Reference Effron and Monin2010). Transgression may cause individuals to infer that past good deeds were intended to claim underserved moral benefits—leading to perceptions of hypocrisy. Alternatively, past good deeds may cause individuals to interpret transgression as unintentional (e.g., an accident)—thereby mitigating perceptions of hypocrisy (Effron & Monin, Reference Effron and Monin2010). The former tendency reflects a lack of moral license, as a relatively great transgression is perceived to outweigh (the moral credits or credentials accrued from) prior good behavior. In this case, the motives for past good deeds are reinterpreted in light of bad deeds, and the firm is judged to have behaved hypocritically. The latter tendency is redolent of licensing a la moral credentials, as an impressive prior record of good behavior influences individuals to interpret morally dubious behavior more favorably. This lessens the degree to which the transgression undermines positive evaluations of the firm’s motives.

Furthermore, the tendency for transgression to be viewed as unintentional in light of past good deeds echoes a notable contribution to the CSR literature from Godfrey (Reference Godfrey2005), who argued that a firm’s prior record of CSR can influence individuals’ judgments of its culpability in a socially harmful event. This is because a strong prior record of CSR can act as evidence of good corporate character, mitigating individuals’ negative assessments regarding the firm’s intention of wrongdoing (Godfrey, Hatch, & Hansen, Reference Godfrey, Hatch and Hansen2010). In the wake of a potential transgression, a positive evaluation of a firm’s prior CSR activities “encourages stakeholders to give the firm the benefit of the doubt regarding intentionality, knowledge, negligence, or recklessness” (Godfrey, Reference Godfrey2005: 788).

In a similar vein, Lange and Washburn (Reference Lange and Washburn2012) proposed that a firm’s prior record of CSR can serve “as a cognitive anchor, biasing the observer as he or she makes adjustments from the anchor to reach an interpretation of the meaning of a new firm action” (Lange & Washburn, Reference Lange and Washburn2012: 312). Indeed, the extant CSR literature has repeatedly demonstrated that individuals tend to reward firms perceived to have a genuine CSR motive, but respond negatively toward firms perceived to have a profit-driven motive for its CSR activities (e.g., Brunk & de Boer, Reference Brunk and de Boer2020; Gao & Mattila, Reference Gao and Mattila2014; Kim & Choi, Reference Kim and Choi2018). For example, Gao and Mattila (Reference Gao and Mattila2014) demonstrated that consumers were more satisfied when they perceived a hotel’s CSR to be driven by a genuine motive. Recently, Kim and Choi (Reference Kim and Choi2018) documented that individuals responded more positively toward a firm’s post-crisis CSR communication if they considered the communication was driven by a genuine CSR motive. Consistent with these findings, and reflecting arguments drawn from the moral licensing theory above, we propose the following hypotheses:

Taken together, the arguments presented above imply a double mediation—by both firm warmth and a genuine CSR motive—of the impact of prior CSR on corporate hypocrisy. If the impact of prior CSR on CSR motive is mediated by firm warmth (H2), and the impact of prior CSR on corporate hypocrisy is mediated by CSR motive (H4), this implies that the impact of prior CSR on corporate hypocrisy is mediated by a genuine CSR motive, which in turn is mediated by firm warmth. To reflect this double mediation, we present the following hypothesis:

The Moderating Role of CSR–CSI Domain

As mentioned above, moral licensing is understood to arise from prior good deeds for two reasons: moral credits and moral credentials. The former requires moral credits (accrued from prior good deeds) to outweigh moral debits (from any subsequent transgression); the latter relies upon a tendency for moral credentials earned from prior good deeds to cause observers to positively construe subsequent morally dubious behavior. However, for Effron and Monin (Reference Effron and Monin2010), some transgressions are unlikely to be licensed in these ways. This is because the nature of such transgressions undermines the positive evaluations of prior behavior, thereby undermining the basis for moral credits and credentials. Moreover, these transgressions are relatively likely to bring perceptions of hypocrisy, because they tend to be viewed as more directly contradictory to prior good deeds.

Effron and Monin (Reference Effron and Monin2010) made this argument with reference to the distinction between different-domain transgressions (i.e., prior good deeds and transgression are associated with different, relatively unrelated social issues) and same-domain transgressions (i.e., prior good deeds and transgression are associated with the same or highly related social issues). The latter are expected to be more strongly associated with perceived hypocrisy. Crucially, they argue that transgressions “in the same domain as prior moral behavior can . . . create the appearance of making, and then contradicting, a claim about one’s values” (Effron & Monin, Reference Effron and Monin2010: 1620). In contrast, “transgressions in a different domain should likely not appear hypocritical, because they are not directly inconsistent with prior good deeds” (1620). In the context of individuals’ evaluations of a person, rather than a firm, Effron and Monin (Reference Effron and Monin2010) provided empirical support for this by demonstrating a tendency for moral licensing to be present for different-domain transgressions but absent for same-domain transgressions.

We apply this argument to the evaluation of firms by first recalling the previously discussed tendency for individuals to make anthropomorphic judgments of firms (Godfrey, Reference Godfrey2005). It is well-documented that individuals are willing to ascribe character attributes (e.g., selfishness, trustworthiness) and motives to firms (e.g., Davies, Chun, da Silva, & Roper, Reference Davies, Chun, Silva and Roper2001). Consistent with this, the notion of moral licensing has been used to explain a link between CSR and CSI, wherein CEOs may be willing to run greater risk of CSI if they believe their firm has accumulated sufficient moral credits from its prior CSR record (Ormiston & Wong, Reference Ormiston and Wong2013). Second, we note that CSR is a multidimensional concept that spans the plurality of social and environmental issues. These issues are in turn arrayed across a host of disparate domains, e.g., from habitat degradation to community cohesion, from climate change to employee health, and so forth (Mattingly & Berman, Reference Mattingly and Berman2006). Therefore, a firm’s CSR activities and CSI events may pertain to the same domain or different ones. Indeed, empirical evidence points to a common tendency for firms to exhibit a mix of positive and negative indicators not only across domains, but also within each domain of CSR (Oikonomou et al., Reference Oikonomou, Brooks and Pavelin2014).

When applied to this context, Effron and Monin’s (Reference Effron and Monin2010) argument implies that the effect of a firm’s prior record of CSR affects perceptions of hypocrisy is moderated by the domain(s) of CSR and CSI. Specifically, a firm’s prior record of CSR mitigates perceptions of hypocrisy more strongly if CSR and CSI are in different domains (vs. the same domain). When a CSI event does not relate to the same domain as a firm’s prior record of CSR, that event is less likely to be perceived as a direct contradiction of a firm’s prior CSR. For example, when faced with an apparent corporate misdeed that has brought damage to wildlife, individuals may consider that such a CSI event does not contradict a firm’s prior CSR contributions to employee welfare. That is, the revelation of poor environmental stewardship does not undermine the notion that the firm acts with due regard for employee welfare.

This argument is somewhat supported by previous empirical findings. Janney and Gove (Reference Janney and Gove2011) found that when CSR and CSI were in the same domain (vs. different domains) firms tended to be punished more harshly by the market. In a similar vein, Carlos and Lewis (Reference Carlos and Lewis2018: 130) found that, with respect to membership of the Dow Jones Sustainability Index, “the threat of hypocrisy is amplified for firms with stronger reputations in the same domain as the certification.” However, it is worth noting that both studies (Carlos & Lewis, Reference Carlos and Lewis2018; Janney & Gove, Reference Janney and Gove2011) mainly adopt a firm-level perspective in documenting how the domains of CSR and CSI can influence market reactions and firm strategy without direct evidence regarding individuals’ judgments of corporate hypocrisy. To reflect the arguments presented above and complement these studies by providing more direct evidence of hypocrisy judgments, we predict:

Sample

Using age and gender as criteria for stratified sampling, 181 participants were recruited for Study 1 via one of the biggest online panels—Amazon Mechanical Turk (MTurk).Footnote ¹ We recruited participants from MTurk because online participants show similar characteristics to our target population—the general public. In addition, the recruitment of participants online is cheaper than paper-and-pencil methods (Berinsk, Huber, & Lanz, Reference Berinsk, Huber and Lanz2012; Buhrmester, Kwang, & Gosling, Reference Buhrmester, Kwang and Gosling2011). As MTurk’s participants are mainly in the United States (Berinsk et al., Reference Berinsk, Huber and Lanz2012; Buhrmester et al., Reference Buhrmester, Kwang and Gosling2011), Study 1 recruited American residents as participants. However, among the 181 participants, 81 failed our attention-check questions, leaving a reduced, yet sufficient, and effective sample size of 100. It is worth noting that this failure rate for our attention check was consistent with Oppenheimer, Meyvis, and Davidenko (Reference Oppenheimer, Meyvis and Davidenko2009), who suggested 30 to 40 percent of online participants may fail to read instructions carefully. Following Zhou and Fishbach (Reference Zhou and Fishbach2016), we compared the demographic characteristics (age and gender) of participants that failed the attention-check questions with those that did not. The results showed no statistically significant difference. We further compared the sample attrition rate across conditions. The results again showed no statistically significant difference. Thus, our results were not likely to be biased by sample attrition.

We used the United States Census Bureau (2017) as a guide to check that our sample’s age and gender—our stratified sampling criteria—matched the current American population. According to United States Census Bureau (2017) data, 50.8 percent of the US population is female and 15.2 percent of the population are aged 65 or over. A one-sample binomial test indicated that our sample’s gender distribution did not differ significantly from the general population (p = 0.24), with 51 percent of participants in our sample being female. In addition, our sample’s age distribution did not differ significantly from the general population (p = 0.86), with 15 percent of participants in our sample aged 65 or over. Thus, our sample matched the current US population in age and gender.

Design and Procedure

The experiment was a one-factor (prior CSR: strong vs. weak) between-subject design. Participants’ perceptions of firm warmth and CSR motive were collected as mediator and dependent variable. In order to avoid demand artifacts (Sawyer, Reference Sawyer1975), participants were told that the purpose of Study 1 was to investigate how people attribute human traits, emotions, or intentions to nonhuman entities such as firms. After the collection of their demographic information, each participant was invited to read one of the prior-CSR scenarios documented below. Following Goodman and Paolacci (Reference Goodman and Paolacci2017) and Wessling, Huber, and Netzer (Reference Wessling, Huber and Netzer2017), participants were asked to answer the focal firm’s name and industry as an attention check. This was to ensure that participants’ responses were not based on guessing (i.e., without reading the scenarios). Only participants that correctly answered the attention-check questions could proceed. Next, participants answered manipulation check questions, and then their perceptions of firm warmth and CSR motive were collected, together with control variables.

Prior-CSR Manipulation

Consistent with Bolton and Mattila (Reference Bolton and Mattila2015), a fake retailer was chosen as the focal firm to avoid the potential biases associated with participants’ pre-existing knowledge of and attitudes toward existing retailers. In order to increase the external validity of the study, we developed CSR scenarios by drawing upon items within the MSCI ESG—formerly Kinder, Lydenburg & Domini (KLD) —ratings database and an analysis of news media content, via the LexisNexis database. As a result, rural youth unemployment was chosen as the focal issue for CSR, because community CSR is a key index in the MSCI/KLD ratings and the issue has been consistently featured in news coverage of corporate behavior. Furthermore, many key retailers emphasize their commitment to such community-related issues in their annual reports of CSR engagement (e.g., Marks & Spencer, 2016).

Participants were randomly assigned to read either a strong or weak prior-CSR scenario. For strong prior CSR, participants were told that the focal firm was a leader in the retail industry in reducing rural youth unemployment. In addition, participants were told that between 2006 and 2015 the focal firm donated 80 percent more than the industry average and donated £2 million every year regardless of its profitability. Furthermore, in 2015, the focal firm encouraged its employees to spend 240,000 paid work hours (equivalent to a monetary value of £1.1 million) to help various charities reduce rural youth unemployment. As a result, the focal firm had been consistently ranked in the top three of the “50 best retailers for youth.”

In contrast, for weak prior CSR, participants were told that the focal firm played an active role in the retail industry in reducing rural youth unemployment, but that between 2006 and 2015 it donated only 30 percent of the industry average and every year its donation depended on its profitability (failing to contribute in some years because of low profitability). As a result, the focal firm had never been ranked among the “50 best retailers for youth.” All other information was identical across the two conditions.Footnote ²

Measures

Study 1 Discussion

Study 1 provides evidence that strong (vs. weak) prior CSR increases perceptions of a genuine CSR motive. Furthermore, it demonstrates that this effect is partly mediated by perceived firm warmth. Building on Study 1, Studies 2, 3, and 4 address how prior CSR, perceived firm warmth and a genuine CSR motive influence perceptions of corporate hypocrisy.

Sample

We recruited 300 participants for Study 2 from the British general public via Qualtrics. As in Study 1, stratified sampling was used to ensure that the sample reflected the demographic characteristics (age and gender) of the British general public, as reported in the most recent census data (Office for National Statistics, 2011). However, 110 participants failed our attention check questions, leaving a reduced, yet sufficient, sample size of 190.Footnote ⁶ As in Study 1, and following Zhou and Fishbach (Reference Zhou and Fishbach2016), we compared the demographic characteristics of participants who failed the attention-check questions with those who did not. The results showed no statistically significant difference. We further compared the sample attrition rate across conditions. The results again showed no statistically significant difference. Thus, our results were not biased by sample attrition.

The Office for National Statistics (2011) suggests that currently 50 percent of the UK population is female. In terms of age, 15 percent of the UK population is aged 16 to 24, 22 percent is aged 25 to 34, 22 percent is aged 35 to 44, 22 percent is aged 45 to 54, and 19 percent is aged 55 to 64 (Office for National Statistics, 2011). A one-sample binomial test indicated that our sample’s gender distribution did not differ significantly from the general population (p = 0.28), with 54 percent of participants in our sample being female. Similarly, a chi-square goodness-of-fit test indicated that our sample’s age distribution did not differ significantly from the general UK population (χ ² (4) = 0.47, p = 0.79). Overall, 10 percent of our sample were aged 16 to 24, 22 percent were aged 25 to 34, 21 percent were aged 35 to 44, 23 percent were aged 45 to 54, and 24 percent were aged 55 to 64. Thus, our sample matched the British general public in age and gender.

Design and Procedure

Participants were randomly assigned to one of the prior-CSR conditions: strong vs. weak vs. control. Participants’ perceptions of firm hypocrisy, firm warmth, and a genuine CSR motive were collected as key mediators and dependent variables. In order to avoid demand artifacts (Sawyer, Reference Sawyer1975), participants were told that the purpose of the research was to examine how multitasking and single tasking differently impacted on people’s memory and thoughts. Participants were also told that the research contained different and unrelated sections. After collecting their demographic information, participants in the strong and weak prior-CSR conditions were asked to read one of the prior-CSR scenarios reported in Study 1. Participants in the control group were exposed to the same firm profile without any prior-CSR information. Then participants were asked to answer attention check questions and manipulation check questions as in Study 1.

Participants were then invited to undertake twenty simple mathematical calculations as a distraction task. This was intended to reduce the likelihood of subjects guessing the true purpose of the research project and, by implication, to avoid demand artifacts (Sawyer, Reference Sawyer1975). Following this, all participants were asked to read a CSI-event scenario on bird population decline, described below. While reading, they were required to click the dollar signs ($) embedded in the article and remember the total number of such signs found (as they would need to report it later). This supported the cover story that this project was intended to examine how multitasking and single tasking differently impacted people’s memory and thoughts. After reading the scenario, participants were asked to answer manipulation check questions relating to the CSI event, and provide their views of corporate hypocrisy, a genuine CSR motive, firm warmth, and control variables. Finally, participants were invited to report the number of dollar signs ($) found in the CSI scenario, after which they were thanked and debriefed.

Manipulation and Measures

Study 2 Discussion

Study 2 extends Study 1 by providing further support for our hypotheses. The results of Study 2 indicate that firms with a strong prior record of CSR are more likely to be perceived as genuinely motivated in their CSR activities. This, in turn, lessens perceptions of corporate hypocrisy. Furthermore, perceived firm warmth mediates the effect of prior CSR on the perception of a genuine CSR motive, which in turn mediates the effect of prior CSR on perceived corporate hypocrisy. However, to address Hypothesis 6, we must next ask: Does CSR–CSI domain moderate the impact of prior CSR on the perception of corporate hypocrisy?

Sample

Using the same sampling method as Study 2, 515 participants were recruited for Study 3 from the UK public via Qualtrics. However, 190 failed the attention-check questions, leaving an effective sample size of 325.Footnote ⁸ The same as in the previous two studies, and following Zhou and Fishbach (Reference Zhou and Fishbach2016), we compared the demographic characteristics of participants who failed the attention-check questions against those who did not. The results showed no statistically significant difference. We further compared the sample attrition rate across conditions. The results again showed no statistically significant difference. Thus, our results were less likely to be biased by this sample attrition.

As Study 3 recruited participants from the British general public, we again used data from the Office for National Statistics (2011) to guide stratification on age and gender. According to a one-sample binomial test, our sample’s gender distribution did not differ significantly from the general population (p = 0.65), with 51 percent of participants in our sample being female. According to a chi-square goodness-of-fit test, our sample did not significantly differ from the general population with respect to age distribution (χ ²[4] = 1.21, p = 0.55). Overall, 10 percent of our sample were aged 16 to 24, 21 percent were aged 25 to 34, 23 percent were aged 35 to 44, 22 percent were aged 45 to 54, and 24 percent were aged 55 to 64. This suggested our sample matched the UK general public in age and gender.

Design, Procedure, and Measurement

Study 3 was a three-by-two (Prior CSR: strong vs. weak vs. control) * (CSR–CSI domains: same vs. different) between-subject design. Just as in Study 2, participants’ perceptions of corporate hypocrisy, firm warmth, and a genuine CSR motive were collected as key mediators and dependent variables. In terms of procedure, Study 3 differed from Study 2 in two respects: first, in Study 3, the CSI event always related to rural youth unemployment; second, in Study 3, the focal firm’s prior CSR either prevented bird population decline or reduced rural youth employment. Thereby, for some participants, the firm’s prior CSR and the CSI event were in the same domain (i.e., rural youth unemployment), and for other participants, they were in different domains (i.e., bird population decline and rural youth unemployment, respectively). The rest of Study 3’s experimental procedure and measurement were identical to Study 2.

Manipulation

Study 3 Discussion

Study 3 provides further evidence in support of our framework. It demonstrates that firms with a strong prior record of CSR can lessen perceptions of corporate hypocrisy because stakeholders are more likely to infer a genuine motive in their CSR activities. More importantly, Study 3 suggests that the CSR–CSI domain moderates the impact of prior CSR on perceived corporate hypocrisy such that the moderating effect is stronger when CSR and CSI are in different domains, rather than the same one.

Sample

Using the same sampling method as Study 1, we recruited 800 participants for Study 4 from MTurk. However, 224 of these failed our attention-check questions, leaving an effective but sufficient sample size of 576.Footnote ⁹ The same as in the previous two studies, and following Zhou and Fishbach (Reference Zhou and Fishbach2016), we compared the demographic characteristics of participants who failed the attention-check questions against those who did not. The results showed no statistically significant difference. We further compared the sample attrition rate across conditions. The results again showed no statistically significant difference. Thus, our results were less likely to be biased by this sample attrition.

As per Study 1, we used the United States Census Bureau (2017) as a guide to our sample’s age and gender—our stratified sampling criteria—and sought to match the current American population. Currently, according to United States Census Bureau (2017) data, 50.8 percent of the US population is female and 15.2 percent of the population are aged 65 or over. A one-sample binomial test indicated that our sample’s gender distribution did not differ significantly from the general population (p = 0.17), with 51 percent of participants in our sample being female. In addition, our sample’s age distribution did not differ significantly from the general population (p = 0.73), with 15.4 percent of participants in our sample aged 65 or over. Thus, our sample matched the current American population in age and gender.

Design, Procedure and Measurement

Study 4 had the same design as Study 3: a 3 (prior CSR: strong vs. weak vs. control) X 2 (CSR–CSI domains: same vs. different) between-subject design. Participants’ perceptions of firm warmth, a genuine CSR motive and firm hypocrisy were collected as key mediators and dependent variables. In terms of procedure and measurement, Study 4 made the following changes to Study 3: first, unlike Study 3 where all participants were exposed to the same CSI event, in Study 4 participants were randomly exposed to either an environmental CSI event or a community CSI event, as described below. Second, in Study 4, all participants were exposed to prior CSR activity in the same domain—environment. Thereby, for some participants, the firm’s prior CSR activity and the CSI event were in the same domain (environment) and, for other participants, the firm’s prior CSR activity and the CSI event were in different domains (environment and community, respectively). Third, unlike Study 3, where we measured perceptions of firm warmth and a genuine CSR motive after a CSI event, in Study 4 both variables were measured before a CSI event. Finally, in order to increase scale sensitivity, all constructs were measured on eleven-point scales (rather than the seven-point scales used in the preceding studies). The rest of Study 4’s experimental procedure and measurement were identical to the earlier studies.

Manipulation

Structural Model

Overall, the estimations fit the data well. The explained variance (adjusted R ²) for perceived firm warmth was 0.74, followed by perceptions of a genuine CSR motive at 0.64 and corporate hypocrisy at 0.43; all larger than the 0.1 suggested by Falk and Miller (Reference Falk and Miller1992). The results of the blindfolding procedure in SmartPLS indicated the values of Stone–Geisser’s Q ² test were all greater than 0 (values ranged from 0.36 to 0.62), providing strong evidence of the model’s predictive power (Tenenhaus, Vinzi, Chatelin, & Lauro, Reference Tenenhaus, Vinzi, Chatelin and Lauro2005). Our model’s standardized root mean square residual (SRMR) was 0.03, below the acceptable threshold of 0.08. The normed-fit index (NFI) had a value of 0.97, above the required threshold of 0.95, indicating a good fit (Hair et al., Reference Hair, Hult, Ringle and Sarstedt2017). In addition, none of the constructs’ variance inflation factors (VIFs) were larger than 5 (values ranged from 3.10 to 3.29), indicating that multicollinearity was not an issue at the structural model level.

The Moderating Role of CSR–CSI Domain (Hypothesis 6)

We used multigroup analysis in SmartPLS to test this moderation. When CSR and CSI were in different domains, prior CSR had a significant negative impact on perceived corporate hypocrisy (β = -0.39, p < 0.001). When CSR and CSI were in the same domain, prior CSR also had a significant negative impact on perceived corporate hypocrisy (β = -0.30, p < 0.01). More importantly, the difference between these two impacts was marginally significant (p < 0.06), providing support to H6 (see Figure 3).

Figure 3: The Interactive Effect between CSR–CSI Domain and Prior CSR on Perceived Corporate Hypocrisy (Study 4)

Study 4 Discussion

Using a different prior-CSR manipulation and a different method of data analysis (PLS-SEM), Study 4’s results are consistent with those of our preceding studies: high perceptions of firm warmth generated by a firm’s strong prior record of CSR makes stakeholders more likely to infer that a genuine motive underlies its CSR activities, which in turn lessens perceptions of corporate hypocrisy; the effect of prior CSR on perceived corporate hypocrisy is moderated by CSR–CSI domain.

Theoretical Contributions

In terms of the corporate hypocrisy literature, there is an unresolved debate about whether prior CSR can reduce perceived corporate hypocrisy in the wake of a CSI event. Both Godfrey (Reference Godfrey2005) and moral licensing theory (Miller & Effron, Reference Miller and Effron2010) have argued that prior CSR can accrue goodwill among individual stakeholders, encouraging them to give a firm the benefit of the doubt in relation to its wrongdoings. This is consistent with the notion that a prior record of CSR lessens perceptions of corporate hypocrisy. However, Shim and Yang (Reference Shim and Yang2016) found that prior CSR had no impact on perceived hypocrisy, while Wagner et al. (Reference Wagner, Lutz and Weitz2009) even reported that prior CSR could increase hypocrisy perceptions. We argue that such mixed results are partly due to a failure in the extant literature to differentiate inconsistency performed because a firm lacks a genuine CSR motive and uses its CSR engagement to claim undeserved moral benefit(s) from inconsistency that arises from a lack of ability and/or resources (Effron et al., Reference Effron, O’Connor, Leroy and Lucas2018; Monin & Merritt, Reference Monin, Merritt, Mikulincer and Shaver2011). Here, we demonstrate that a genuine CSR motive can mitigate the perception that a firm’s inconsistency is performed to claim undeserved moral benefits (Effron et al., Reference Effron, O’Connor, Leroy and Lucas2018). This highlights the importance of understanding individuals’ inferences about the motives that underlie a firm’s CSR activities.

Guided by the SCM (Cuddy et al., Reference Cuddy, Fiske and Glick2008; Fiske et al., Reference Fiske, Cuddy, Glick and Xu2002), our research further identifies the source of individuals’ inferences of a genuine CSR motive—the perception of firm warmth generated by its prior record of CSR. By doing so, we provide a basis for empirical study that tests not only the relationship between prior CSR and perceived corporate hypocrisy, but also the mechanisms underlying this relationship. Consistent with the imperative for evidence-based theory development, our four experimental studies provide detailed empirical support for our key theoretical arguments.

Second, we provide evidence in support of an extension to corporate evaluations of Effron and Monin’s (Reference Effron and Monin2010) arguments that hypocrisy is more evident when good and bad deeds relate to the same (vs. different) domains. Our results consistently suggest that when CSR and CSI are in different domains (vs. the same one), individuals are less likely to judge the focal firm as hypocritical. This complements the extant literature by providing direct evidence of the underlying mechanisms by which markets tend to enact greater punishment when a firm’s CSR and CSI lie in the same domain (Carlos & Lewis, Reference Carlos and Lewis2018; Janney & Gove, Reference Janney and Gove2011).

In terms of the CSR literature, our results suggest that stronger prior CSR can reduce perceptions of corporate hypocrisy. This is redolent of Godfrey’s (Reference Godfrey2005) argument that prior CSR can function as an insurance mechanism, preserving a firm’s value in the wake of negative events. Our article not only supports Godfrey (Reference Godfrey2005) but also provides unique insights into why superior prior CSR can insure firms against negative judgments in the wake of a CSI event. In addition, in sampling respondents from the general public, we address the criticism of the extant CSR-related literature for its narrow focus upon a subset of primary stakeholders (e.g., investors and consumers), leaving the nature of the relationship between the interests of business and those of broader society largely unexplored (Barnett, 2019).

Limitations and Future Directions

Although our research contributes to both CSR literature and corporate hypocrisy literature, any conclusions should be treated with caution because of the following limitations. First, based on the SCM (Cuddy et al., Reference Cuddy, Fiske and Glick2008; Fiske et al., Reference Fiske, Cuddy, Glick and Xu2002), we propose that perceived firm warmth affects stakeholders’ inferences of a firm’s motive for its CSR activities. However, we collected perceptions of warmth and perception of a genuine CSR motive at the same time. In order to rule out reverse causation—a genuine CSR motive leads to firm warmth, we retested H2 with perceived firm warmth as the dependent variable and a genuine CSR motive as the mediator. The indirect conditional effect became nonsignificant in all studies, suggesting a genuine CSR motive was not a mediator of prior CSR on perceived firm warmth. However, future research might manipulate perceived firm warmth to extend the evidence on the direction of causation.

Second, this research recruited participants from only two developed countries—the United States and the United Kingdom. This may limit the generalizability of our findings, because Matten and Moon (Reference Matten and Moon2008) suggested that different countries have differing views of CSR practices. Indeed, in our Studies 2 and 3 that sampled from the British general public, perceived firm warmth fully mediated the impact of prior CSR on a genuine CSR motive. However, in Studies 1 and 4 where we sampled from the American general public, perceived firm warmth only partly mediated the impact of prior CSR on a genuine CSR motive. In addition, in Studies 2 and 4, perceptions of firm warmth and a genuine CSR motive fully mediated the effect of prior CSR on perceptions of corporate hypocrisy. However, in Study 3, this mediation was partial. Thus, variation in results across our studies may reflect country differences, and future research may investigate the degree and implications of any such international variations in findings.

Third, this research tests the impact only of proactive CSR on corporate hypocrisy perceptions, given that participants were asked to read about a firm’s prior record of CSR activities and were then informed of a CSI event. Wagner et al. (Reference Wagner, Lutz and Weitz2009) differentiated proactive CSR and reactive CSR. While proactive CSR refers to those CSR actions and statements preceding CSI events, reactive CSR refers to those CSR actions and statements made in response to CSI events (Wagner et al., Reference Wagner, Lutz and Weitz2009). Future studies can replicate our research with reactive CSR activity to see whether it gives rise to similar results.

Fourth, this research focuses on only two domains of CSR: environment and community. CSR is a multidimensional concept comprising several domains, such as community, environment, employees, product safety, human rights, corporate governance, and so forth. Future studies can replicate our research in other CSR or CSI domains to test the robustness of our framework.

Finally, in this research all CSI events were perceived to be of moderate strength. However, both Godfrey (Reference Godfrey2005) and Lange and Washburn (Reference Lange and Washburn2012) pointed out that the strength of CSI events is a key factor influencing stakeholders’ judgments of the focal firm. Thus, future research can manipulate CSI event strength to see how our results hold for other strengths. In addition, future research can test other potential moderators of our framework. For example, whether our framework is still valid for stigmatized industries (e.g., tobacco) awaits future research, as individuals may not consider firms in these industries to engage CSR with a genuine motive. Product substitutability may also moderate the impact of prior CSR on perceived corporate hypocrisy. If individuals have to use the products (because they cannot be substituted), then they may less likely to perceive corporate hypocrisy to justify their continued use of the products. Furthermore, future research can underpin how individual stakeholders attribute the cause of a CSI event based on the focal firm’s prior CSR record. This can complement our research to provide more detailed insights about individuals’ perceptions of corporate hypocrisy.

Practical Implications

In addition to the theoretical contributions described above, this research has implications for managerial practice. For example, our findings suggest that strong (but not weak) prior-CSR activities can contribute to risk management by providing some degree of insulation against crises. Thus, our findings potentially inform ongoing debate about whether prior CSR is an asset or a liability in negative events (e.g., Godfrey et al., Reference Godfrey, Merrill and Hansen2009; Oikonomou et al., Reference Oikonomou, Brooks and Pavelin2014). On the one hand, some scholars have suggested that prior CSR can protect against losses of share value (Godfrey et al., Reference Godfrey, Merrill and Hansen2009); on the other hand, other studies have demonstrated that prior CSR can worsen a firm’s financial performance (Oikonomou et al., Reference Oikonomou, Brooks and Pavelin2014). The results of our research provide support to the view that strong (but not weak) prior CSR is an asset that offers protective value in the wake of negative events. In particular, this research demonstrates that superior prior CSR encourages stakeholders to infer a genuine motive underlying a firm’s CSR activities. This, in turn, reduces their perceptions of corporate hypocrisy when the focal firm is involved in a CSI event.