CSR commitments, perceptions of hypocrisy, and recovery

This paper examines perceived hypocrisy when a failure is aligned with prior social performance. It is hypothesized that commitment to a CSR domain creates greater performance expectations thus exacerbating the effects when an aligned failure occurs. Study 1 demonstrates that failure alignment and severity increase perceived hypocrisy which negatively impacts customer evaluations of trust, repurchase intent, and brand attitude. Study 2 evaluates two response strategies of apology and compensation vs. no response. An apology significantly reduced perceptions of hypocrisy only when the failure was unaligned with prior CSR. Compensation significantly reduced hypocrisy in both the unaligned and aligned conditions.

Shifting perceptions regarding the role of business in society has lead firms to consider the broader outcomes of activities beyond simple profit maximization. Pressure from stakeholder groups has translated into firms adopting policies aimed at generating favorable perceptions of corporate citizenship, corporate social responsibility (CSR) and business ethics. Take, for example, a survey by Cone Communications in which 94% of respondents said businesses have at least some higher-order obligations beyond the shareholder wealth maximization (Cone Communications., 2017). These findings support the assertion that social outcomes should be considered alongside traditional strategic goals.
However, while the link between CSR and measurable business outcomes is often unclear (McWilliams & Siegel, 2000), several studies have shown that corporate citizenship may contribute to a positive firm reputation (Lai, Chiu, Yang, & Pai, 2010;Tetrault Sirsly & Lvina, 2019). In other cases, research has shown that CSR adds robustness to a firm's reputation, a "buffering effect" (Godfrey, 2005;Kim, 2014;Peloza, 2006) or the ability to protect firms against the negative outcomes of an adverse event such as a product recall or environmental misstep. For example, CSR lead to more favorable attributions by consumers following a product crisis (Klein & Dawar, 2004), protected firm value following regulatory action (Godfrey, Merrill, & Hansen, 2009), or positively influenced consumer actions following a negative event (Kim & Woo, 2019).
In contrast to a buffering effect, other studies have noted how a reputation for CSR in the presence of social irresponsibility can lead to perceptions of hypocrisy (Janney & Gove, 2011;Lenz, Wetzel, & Hammerschmidt, 2017;Wagner, Lutz, & Weitz, 2009).
Hypocrisy is a state in which CSR claims are inconsistent with a firm's internal processes or a discrete external event (Ioannou, Kassinis, & Papagiannakis, 2018). Paradoxically, a consumer's negative view of the event may actually be amplified. As firms engage in CSR they shape stakeholders' perceptions by creating signals of underlying values which may positively impact a firm's reputation (Tetrault Sirsly & Lvina, 2019). However, in the case of wrong-doing, the event may contradict these signals and engender feelings which call into question a firm's claims. For example, Janney and Gove (2011) found that negative effects were exacerbated when a firm with a positive reputation for good governance was found to be involved in a governance-related scandal.
While studies into the shielding characteristics of a CSR reputation are promising, few consider the multidimensional nature of CSR when ascertaining hypocrisy. A firm can target its social efforts at a variety of CSR subdomains such as its employees, the local community, or environmental sustainability (Öberseder, Schlegelmilch, & Murphy, 2013). Few studies consider how the domain of the crisis situation and the domain of prior CSR actions may interact to alter consumer perception of the event. For instance, would stakeholders perceive a firm as hypocritical if the wrong-doing occurs in an area unrelated to a firm's CSR efforts? In order to fill this gap, this study proposes that when a failure occurs in a domain aligned with the firm's prior focus of citizenship efforts, the resulting effect will be greater feelings of hypocrisy among consumers compared to an unaligned domain.
Secondly, this paper contributes to the reputation and recovery literature by examining domain alignment effects on three generic recovery strategies, In doing so, this paper helps answer the call for further investigation into the process of loss and recovery of a firm's approval (Westphal & Deephouse, 2011).

Corporate social responsibility
CSR is the collection of responsibilities that a firm has towards society (Carroll, 1991). It represents the efforts by an organization to improve the lives of various stakeholders or create positive changes in the external environment (Dahlsrud, 2008), or a means by which a firm seeks social legitimacy (Pollach, 2015).
Rather than existing as a single construct, the umbrella term CSR can be conceptualized as a collection of various domains, sub domains, and behaviors which highlight responsibilities aimed at various stakeholder groups. For purposes of sensemaking and analysis, several scholars have adopted differing domain classifications. For example, in his foundational paper, Archie Carroll (1991) described the umbrella of CSR in 4 parts, namely economic, legal, ethical, and philanthropic responsibilities. Matten and Moon (2008) partitioned CSR into "explicit" and "implicit" components. Explicit actions are classified as voluntary philanthropic actions directed towards internal or external stakeholders for the purposes of social or shared value creation. Implicit CSR are actions undertaken seemingly through the normal course of business which creates social value such as high wages and benefits for employees. Further, (Cornelius, Todres, Janjuha-Jivraj, Woods, & Wallace, 2008) examined CSR in the context of internal vs. external CSR denoting actions targeted towards stakeholders based upon their relationship to the firm. Reporting frameworks such as the Global Reporting Initiative and assess firms across a variety of domains such as performance regarding the natural environment, labor, corporate governance and fiscal responsibility.
Likewise, this paper adopts a multidimensional view of reputation for CSR. Rather than existing as a generalized assessment of a firm, a reputation can consist of multiple underlying domains (Lange, Lee, & Dai, 2011). A firm can be known for a "…particular attribute or characteristic: An organization has a reputation for something, such as having high quality products…" (Fischer & Reuber, 2007 p. 57). Assessments of individual components of reputation arise from past actions, and influence what behaviors a stakeholder expects to see from a firm in the future (Deutsch & Ross, 2003). In a similar vein, firms can engender assessments of various CSR domains. I.E. a firm which concentrates on strong environmental performance may have separate assessments of its actions regarding its employees, governance, etc. (Fombrun, Ponzi, & Newburry, 2015). This idea forms one of the central questions of paper which asks what happens if a firm which has a reputation for strong performance in one CSR domain experiences a negative event in another.
For purposes of experimentation this paper focuses on two distinct categories of CSR. Since the research question does not pertain to a specific category per se, but rather to the general effects of category alignment, care was taken to select two domains which would be viewed as properly differentiated from each other. Previous authors have distinguished between CSR domains utilizing stakeholder focus as the category boundary. For example, (Hameed, Riaz, Arain, & Farooq, 2016) utilized internal vs. externally focused actions for their analysis of employee-company identity. Other authors restricted their analyses to popular CSR disclosure dimensions such as community involvement and the natural environment (Hou & Reber, 2011;Jin & Lee, 2019). Following this precedent, conventions of Cornelius et al. (2008) the paper investigates employee relations (or employee-oriented) CSR and actions related to the natural environment. This utilizes one internally and one externally focused CSR domain targeting two distinct stakeholder groups. Further, these domains were selected with the assumption that continuing media focus on issues such as sustainability and labor relations would render these categories comprehensible by the average respondent.

Corporate hypocrisy
Put simply, corporate hypocrisy occurs when a firm claims to be something that it is not (Wagner et al., 2009), or its actions run counter to the image it is attempting to portray to its stakeholders. As a firm commits resources to a cause it is attempting to create associations between itself and socially desirable attributes in the eyes of stakeholders (i.e. "we care about the environment") (Brown & Dacin, 1997), and is building expectations of future performance (de Quevedo-Puente, de la Fuente-Sabaté, & Delgado-García, 2007) .
We contend that this is analogous to both a claim about current state of being and an implicit promise, or an understood commitment to future performance. When information is revealed regarding behavior that contradicts previously held positions, the resulting dissonance may lead consumers to question the motives or authenticity of previous attributions (Lange & Washburn, 2012), and may enhance negative reactions to wrongdoing (Wagner et al., 2009). This is evidenced by Janney and Gove (2011) who found that previous CSR reputation for corporate governance served to exacerbate the negative effects of governance-related wrongdoing. This is similar to logic found in the literature on "greenwashing" or the lack of substantive environmental action to match with firm rhetoric. A firm may claim to be green without offering any material commitments to sustainability (Walker & Wan, 2012). In this situation, revelations about a firm's true nature may exacerbate the effects.
Conversely, when a failure occurs in a non-aligned area, consumers must evaluate the behavior in the absence of distinct claims regarding intended commitment. While previous research has established that good works in one CSR domain may carry over to beliefs in other areas (Klein & Dawar, 2004), we hypothesize that the lack of concrete assertions in one area by the firm will lead to fewer future expectations regarding performance in that specific domain which will, in turn, produce fewer unfavorable attributions. Thus we arrive at the following hypothesis: H1a: Perceived hypocrisy will be greater when the failure aligns with prior CSR reputation.

Severity
In addition to alignment, it is useful to consider the impact of failure severity on consumer reactions. Hypocrisy is a representation of the gap between communication and reality (Fassin & Buelens, 2011). A relatively minor failure may represent a trivial contradiction to firm communication and may be seen by consumers as unavoidable, or largely unimportant. Conversely, a severe failure may represent a deliberate lapse in organizational conduct and may signal to consumers that there are significant problems with underlying processes. These attributions may drive the perceived dissonance between ideal conduct and actual behavior thereby leading to greater perceptions of hypocrisy (Fassin & Buelens, 2011). Therefore: H1b: Failure severity will have a direct effect on perceived hypocrisy such that a more severe failure will lead to greater perceptions of hypocrisy.
According to attribution theory, consumers look for causal explanations when presented with information (Kelley & Michela, 1980). In the case of hypocrisy, when actions conflict with prior reports of CSR, consumers may elaborate upon the event and infer negative attributes as factors driving the behavior. These perceptions, in turn, may lower overall brand evaluations and decrease repurchase intent. This is supported by Wagner et al. (2009) who found that consumers experienced stronger negative affect when perceptions of hypocrisy were high. Further, since trust is dependent upon anticipated consistency of behavior, actions which contradict prior CSR reports may lead individuals to rate a firm as less trustworthy. Based on this assessment, we predict.
H2: Perceived Hypocrisy will have a negative relationship with overall Brand Attitude, repurchase intent, and trust.

Study 1
Study 1 employed a 2 Alignment (aligned/unaligned) × 2 Severe (yes/no) mixed-methods experimental design. The impact of perceived hypocrisy on attitudes towards brands is a phenomenon which is expected to be found across the general population. Thus, we did not exclude anyone between the ages of 18-64 years old from participating in the study. That said, data was collected in the United States and thus will more than likely contain more U.S. citizens. In order to efficiently capture this population, data was collected using an online subject pool, Amazon Mechanical Turk, which provides high-quality data quickly and relatively inexpensively and Qualtrics, an online survey design and data collection tool. An initial sample of 223 respondents was achieved. However, 15 respondents were removed due to incomplete data or for failure to pass a comprehension check thereby bringing the total sample size to 207 (mean age = 31, 65% male, 75% Caucasian).
In order to establish experimental groups based on alignment, respondents were first exposed to one of two hypothetical scenario detailing a fictitious firm's social practices (either environmental or employee oriented CSR, see Appendix A) and asked to rate perceptions of the firm's CSR practices. Next, respondents were randomly presented with a scenario in which the same fictitious firm's actions resulted in a service failure either in the same domain of CSR practices (aligned) or in another domain of CSR practices (unaligned, see in Appendix B). Subjects were then asked to rate perceptions of the firm's CSR practices, as well as, level of perceived hypocrisy and a number of other outcome variables (e.g. Attitude Toward the Brand, Purchase Intentions, and Trust). All constructs, except for Brand Attitude, were measured using established threeitem 7-point Likert scales. Brand Attitude was measured using an established four-item 7-point semantic differential scale.
A pre-test was conducted to ensure the efficacy of the severity manipulations. A pool of 104 respondents were exposed to either the mild or severe failure in the environmental or employee-oriented CSR domain. Respondents were then asked to rate the severity of the failure using a 5-point likert scale. Independent sample t-tests were conducted to compare conditions. Results indicated a significant different between the mild environmental (M = 2.96, SD = 1.42) and severe environmental (M = 4.07, SD = 1.05) conditions; T(49) = − 3.185, p = .003. Likewise, results indicated a significant difference in the mild employee-oriented (M = 3.2, SD = 1.08) and severe employee-oriented (M = 3.79, SD = .88) conditions; T(47) = − 2.094, p = .042. Further tests revealed no significant differences between the environmental and employee-oriented scenarios in both the mild (T(48) = −.670, p = .506) and severe (T(48 = 1.032, P = .307) conditions, thus indicating no perceived difference between failure type.

Method of analysis
To test the hypotheses, we measured each variable of interest using vetted extant reflective multi-item scales. Each scale utilized seven-point rating scales with reflective multi-item measures. Constructs were developed using items adapted from previous literature (Bougie, Pieters, & Zeelenberg, 2003;Shim & Yang, 2016 ;Wagner et al., 2009). A mean score index of six items was used to measure perceived hypocrisy (Shim & Yang, 2016;Wagner et al., 2009). On a 7point scale, ranging from 1 (strongly disagree) to 7 (strongly agree), participants were asked to indicate how strongly they agreed with statements such as "Star Inc. acts hypocritically," "Star Inc. is pretending to be something it is not," "Star Inc. pretends to be something that it is not," and "Star Inc. does exactly what it says." The index was M = 5.30, SD = 0.088 and reliability was α = .953.
To assess the overall brand attitude, a mean score index of four items was used (dislike-like, favorableunfavorable, bad-good, and positive-negative) on 7point semantic differential scales. The index (M = 3.07, SD = 0.105) produced acceptable reliability (α = .983). Trust was assessed using a mean score index of three item 7-point Likert scales asking for agreement to items such as "Star Inc. is trustworthy" and "Star Inc. would meet my expectations." The index was M = 2.99, SD = 0.094 and reliability was α = .913. Lastly, to assess repurchase intentions, we utilized a mean score index of three items using a 7point Likert scales asking for agreement to items such as "Following the incident, I would consider buying a product from this company" and "Following the incident, there is a strong likelihood that I will buy a product from this company." The index was M = 3.17, SD = 0.106 and reliability was α = .965.
Structured equation modeling (SEM) using AMOS 21 was performed to statistically analyze the data. Test for data normality indicated no violations of assumptions. A confirmatory factor analysis (CFA) measurement model was conducted to assess unidimensionality, as well as, convergent and discriminate validity for each construct (Jöreskog & Sörbom, 1986). The model indicated acceptable overall fit for all constructs individually and for the full model (Chi-Square = 301.86, p < .000), RMSEA = .063; CFI = .98; NFI = .95). Furthermore, convergent and discriminant validity tests were within acceptable levels and, therefore, indicate no issues in the data.
Since a questionnaire was used for data collection, there was an increased possibility of common methods bias (CMB). A second measurement model analysis including a common latent factor was performed and indicates no CMB present (Liang, Saraf, Hu, & Xue, 2007). In addition, a recent paper determined that gender could have an impact on perceptions of CSR, moderating effect on the relationship between CSR perceptions and corporate brand equity, as well as CSR perceptions and perceptions of hypocrisy (Hur, Kim, & Jang, 2016). Thus, we use gender as a control variable in all analyses.

Results
A structural equation model was constructed to test hypothesized relationships. Overall the model indicated good fit (RMSEA = .078, CFI = .992, NFI = .985). As shown in Fig. 1, both alignment (0 = unaligned, 1 = aligned) and failure severity (0 = low, 1 = high) were positively related to perceived hypocrisy (0.27 and 0.26, respectively) indicating support for H1a and H2b. Furthermore, perceived hypocrisy was negatively related to brand attitude, repurchase intentions and trust (−.83, −.75 and − .91 respectively). Figure 1 shows the results of the research model.

Study 2
While study 1 demonstrates that perceived hypocrisy can have a deleterious effect on consumer responses and attitude, study 2 extends study 1 by investigating the effect of two post-failure response strategies (apology and compensation). Corporate communications after a negative event can be an effective tool for reducing financial and reputational harm. Specifically, the issuance of an apology or compensation to victims is a common tactic used by organizations (Hearit, 2006) among other initiatives such as internal policy changes (Pfarrer, Decelles, Smith, & Taylor, 2008), or social investments meant to restore the firm's image as an ethical actor (Noack, Miller, & Smith, 2019).
Corporate crises and other reputation damaging events present the possibility of serious harm to a firm which can manifest in the form of reduced customer satisfaction (Xie & Peng, 2009) negative reviews (Gelbrich, 2010) or even calls for boycotts (Makarem & Jae, 2016). In essence, a hypocrisyinducing event can not only damage the trust afforded by various stakeholder groups, but may also lead to negative expectations about future performance. When apologizing, firms use rhetoric and public statements in an attempt to take responsibility for a crisis and repair damaged stakeholder relationships (Coombs & Holladay, 2008). An apology represents an attempt to communicate remorse and guilt while asserting that such an action will not happen in the future (Pace, Fediuk, & Botero, 2010).
Indeed, research has shown that apologies can blunt the negative effects of a corporate crisis. For example, Manika, Papagiannidis, and Bourlakis (2017) found that consumers were more likely to forgive a business, report higher levels of perceived trustworthiness, and had higher perceptions of overall brand favorability. Further, consumers were less likely to spread negative word-of-mouth following an apology regardless of whether they were directly or indirectly harmed by the firm (Casidy & Shin, 2015).
Despite these findings, there have also been conflicting findings regarding apology type. Coombs and Holladay (2008) found that three recovery strategies (sympathy, apology and compensation) each had a similar effect on consumer anger following an organizational crisis. A partial explanation for this effect.
We hypothesize that in a situation where perceived hypocrisy is high (such as the aligned condition), rhetoric may be less effective due to the already perceived violation of an organization's implicit promises (I.E. "Talk is cheap"). When the organizational crisis is within the same domain as the firm's purported social responsibility competencies, an apology may be insufficient to rebuild trust amongst stakeholders or encourage forgiveness. However, when no apparent violation of previous commitments occurs (such as the unaligned condition), rhetoric may be sufficient to counter hypocrisy due to lower skepticism regarding the authenticity of past claims. In essence, when a firm commits a violation, stakeholders may have less Therefore: H3a: When failure type is unaligned with prior CSR, perceived hypocrisy will be significantly reduced when a firm issues an apology vs. no response.
Conversely, a firm can respond with more substantial resource commitments such as through a pledge to pay for damages or to compensate stakeholders affected by the crisis. This works to restore inequity caused by the organization's failure (Smith, Bolton, & Wagner, 1999). Using previous logic, when failure is aligned more intense resource commitments may be required to overcome the feelings of hypocrisy generated by the incident (I.E. "putting money where one's mouth is"), as substantive commitments may be more effective in repairing organizational image (Pfarrer et al., 2008). Therefore, in the case of an aligned failure we predict that a substantive commitment of resource via compensation will lead to the greatest reduction in perceived hypocrisy compared to other response options.
H3b: When failure type is aligned with prior CSR, perceived hypocrisy will be significantly reduced when a firm responds with monetary compensation compared to an apology or no response.

Method of analysis
To test the hypotheses in study 2, we again measured each variable of interest using vetted extant reflective multi-item scales. Each scale utilized seven-point rating scales with reflective multi-item measures. Constructs were developed using items adapted from previous literature (Bougie et al., 2003;Shim & Yang, 2016;Wagner et al., 2009). We utilized the same mean score index of six items to measure perceived hypocrisy (Shim & Yang, 2016;Wagner et al., 2009). For the Study 2, the index was M = 4.37, SD = 1.69 and reliability was α = .961.
A 2 Alignment (aligned/unaligned) × 2 Severe (yes/no) × 3 Response (no response/ apology/monetary compensation) mix method experimental design was employed for Study 2. Similar to study 1, we expect the phenomenon of study to be found across the general population. Thus again, we did not exclude anyone between the ages of 18-64 years old from participating in the study. Data was collected in the United States and thus will more than likely contain more U.S. citizens. Data was collected using an online subject pool, Amazon Mechanical Turk, in Study 2. A total of 279 respondents completed the survey. However, of those, 35 respondents were removed due to incomplete data or for failure to pass a comprehension check thereby bringing the usable sample size to 244 (mean age = 34, 57% male, 75% Caucasian). The age was distributed across a wider range than in Study 1, with 91% of respondents being between 19 and 54 years old. While allows for more confidence the generalizability of results.
Similar to study 1, respondents were first exposed to one of two hypothetical scenario detailing a fictitious firm's social practices (either environmental or employee oriented CSR). Immediately following this prompt subjects assessed the firm's level of social responsibility. Next, subjects read a scenario involving a firm-level failure (see Appendix C for recovery scenarios). Those in the aligned condition were exposed to a failure in the same domain as the previously established social practices (I.E. environmental CSRenvironmental failure) while those in the unaligned condition were given a failure in the unrelated domain (I.E. environmental CSRemployee related failure). In addition to type, the failure scenarios were varied to reflect either a minor or severe failure. Subjects then rated the severity of the failure. Departing from study 1, subjects were then given information about one of three organization responses consisting of either: 1) no response, 2) an apology, or 3. information regarding monetary compensation. After the prompts, perceived hypocrisy was measured and demographic information was collected.
A confirmatory factor analysis (CFA) measurement model was conducted to assess unidimensionality, as well as, convergent and discriminate validity for each construct using AMOS 21 (Jöreskog, et al., 1988). The model indicated acceptable overall fit for all constructs individually and for the full model (Chi-Square = 550.43 (P = .000), RMSEA = .056; CFI = ..98; NFI = .96). Furthermore, convergent and discriminant validity tests were within acceptable levels and, therefore, indicate no issues in the data.
Since a questionnaire was used for data collection, there was an increased possibility of common methods bias (CMB). A second measurement model analysis including a common latent factor was performed and indicates no CMB present (Liang et al., 2007). As in the first study, we were concerned that gender might have a confounding effect on results, thus we control for it.

Results
An ANCOVA was conducted on participants' rating of corporate hypocrisy. Results revealed a significant main effect for the covariate, Pre_CSR_Perc F(1, 248) = 10.633, p < .01. There was a significant main effect for Alignment, F(1,248) = 13.479, p < .001. In general, participants had greater perceptions of hypocrisy when the failure was aligned with prior CSR (M = 4.67, SD = 1.63), than when it was unaligned (M = 4.03, SD = 1.60). Perception of failure severity exerted a significant effect F(1,248) = 87.660, p < .001. Hypocrisy was rated higher by those who perceived the failure to be more egregious (M = 5.04, SD = 1.56) compared to those with a less severe rating (M = 3.54, SD = 1.35). A significant main effect for Response (none, apology, compensation) F(2,248) = 24.615, p < .001. A significant three-way interaction of response, severity and alignment F(2,248) 4.755, p < .01 was present. To better understand the interaction of alignment, failure severity and response type, the effect of both failure severity and response type on hypocrisy was interpreted within levels of alignment.

Unaligned failure
In the unaligned failure condition, both main effects of perceived severity, F(1,124) = 57.130, p < .001 and response type, F(2,124) = 16.400, p < .001 were significant, but these effects were qualified by a significant two way interaction between perceived severity and response, F(2,124) = 11.257, p < .001. Simple contrast effects in the low perceived severity condition revealed no significant difference based on response. However, when perceived severity was high, simple contrasts showed a significant difference of hypocrisy when comparing both the no response and apology condition (M = 5.847 vs. M = 4.888, t = 2.59, p < .05), and the no response and compensation condition (M = 5.88 vs. M = 3.24, t = 6.73, p < .001). The effect of response type in the unaligned condition is displayed in Fig. 2. Collectively, these results provide support for H3a.

Aligned failure
In the aligned failure condition significant main effects were observed from both response type, F(2, 123) = 8.982, p < .001, and perceived severity, F(1, 123) = 37.185, p < .001. Interestingly, unlike in the unaligned condition, prior perceptions of CSR had no significant effect on hypocrisy. This suggests that beliefs about a firm's social performance my shield a firm from skepticism only when failures do not match that social domain. In contrast to the unaligned condition, the effect of response type did not vary based on perceived severity. Simple contrasts were conducted to compare response types across both failure severity conditions. In support of H3b, compensation produced significantly lower feelings of hypocrisy compared to no response (M = 3.958 vs. M = 5.183,

Fig. 2 Response Type by Perceived Severity for Unaligned Failures
Smith and Rhiney International Journal of Corporate Social Responsibility (2020) 5:1 t = 4.206, p < .001). There was no significant difference between apology and no response.

Discussion
Organizational failings present a serious problem for firm's as they might incur monetary loss, decreased sales, or potentially damage a business' reputation. Given the severity of these effects, scholars have devoted much research to understanding how loss following a negative event can be mitigated or repaired.
To this end, much progress has been made in understanding how corporate citizenship and social performance can act as a buffering mechanism during a crisis, effectively shielding an organization from reputation damage. However, treating social performance as a singular construct when analyzing its effects after a negative event ignores the multidimensionality of CSR. To address this, and to contribute to the burgeoning literature on corporate hypocrisy, this study investigated how CSR sub-domains interact with failure type to influence perceptions of hypocrisy. Both Study 1 and Study 2 investigate the effect of alignment and failure severity on perceived hypocrisy. Consistent with predictions, both the degree of failure alignment and failure severity had significant impacts on perceived hypocrisy which in turn led to a variety of deleterious effects on firm rating. Study 2 extends study 1 by investigating the effects of three response types on perceived hypocrisy following either an aligned or unaligned failure. Consistent with H3a an apology resulted in a significant decrease in perceived hypocrisy when compared to no response. Likewise, as predicted in H3b, compensation resulted in a significant decrease in perceived hypocrisy while an apology had no effect. However, it should be noted that overall, the compensation strategy lead to the most significant reduction in perceived hypocrisy across both groups. This suggests that while a simple apology might be sufficient in limited contexts, overall customers may respond more favorable to substantive commitments of resources regardless of failure type.
The significance of these findings are twofold: first, it partially refutes the conceptualization of CSR as a buffer against firm wrongdoing. The buffering effect was observed but most notably in conditions where the organizational crisis did not match with the area of prior CSR reputation. Further, while a simple apology was enough to counteract some of the negative effects in the unaligned condition, in the aligned condition apology was no different from a non-response. This suggests that managers should consider the nature of the failure context in relation to firm reputation when crafting an organization response strategy.

Limitations and future Research
Both studies have limitations which provide a basis for future research. First, the samples for both studies were taken from a single cultural region which limits its generalizability. Culture and other regional characteristics might influence an individual's judgements about hypocrisy. In terms of hypocrisy, according to Effron, Markus, Jackman, Muramoto, and Muluk (2018) more interdependent, collectivist societies place greater emphasis the subordination of personal interests and are therefore more accustomed to behavior changes based upon context. In this situation, behavior which is inconsistent with past claims might generate less cognitive dissonance and may subsequently be viewed as less hypocritical. In contrast, a more individualistic society places greater emphasis on consistency of behavior (Suh, 2002) and actions which run contrary to past claims may be viewed as more self-serving.
In a similar vein, there is evidence to suggest that reactions to apology and apology type may differ across cultural contexts. Apologies utilized in the United States were found to have a greater emphasis on explanation whereas apologies in Japan placed more importance on showing remorse and promises to prevent future transgressions (Sugimoto, 1997). Some cultural groups have more concern regarding the status of the individual delivering the apology when accounting for communication channel (Sengupta, Ray, Trendel, & Vaerenbergh, 2018). Further, in a China-US comparison respondents differed in importance placed on professional dress, body-language, and expressions of remorse (Song, Eslami, & Galindo, 2018). Taken together, these factors provide future papers with ample opportunity to expand upon the relationship between hypocrisy, apology and recovery by including variations in national context.
While the paper controls for the effect of gender in the analysis, research in cognitive psychology suggests that gender differences may actually influence how an individual reacts to corporate apology both as the receiver and originator of the apology (Wei & Ran, 2019). Future studies could investigate demographic differences and how they interact to determine perceptions of hypocrisy and the effectiveness of apology.
Due to a lack of a theoretical justification for selecting one CSR domain over another, this paper utilizes two categories which were meant to be immediately identifiable and comprehensible for respondents. Further research is needed to determine if perceived hypocrisy is influenced by aligned and unaligned pairings in other CSR domains. Star Inc. takes pride in its commitment to the environment and sustainability and demonstrates this by only using environmentally friendly materials in its clothing. The company ensures that all waste generated by its manufacturing plant is disposed of in a safe and sustainable manner. Further, the company works to promote environmental sustainability outside of its business by donating 10% of each purchase to various conservation groups.
As proof of their efforts, Star Inc. has received many awards and recognition for environmental performance.

Reputation for Employee-Oriented CSR
Star Inc. manufactures and sells a variety of affordable, yet fashionable clothing for all age groups.
Star Inc. takes pride in its commitment to fair labor and human rights and demonstrates this by enforcing safety standards in its plants which are far more stringent than the law requires. The company routinely inspects its manufacturing plants both domestically and abroad to ensure workers are treated fairly and paid a livable wage. Further, the company works to promote human rights outside of its business by donating 10% of each purchase to various groups which fight to improve working conditions in less-developed countries.
As proof of their efforts, Star Inc. has received many awards and recognition for their commitment to people. Star Inc. Criticized for Labor Conditions USA -Star Inc. has drawn criticism from local interest groups after a report was released today which found unsafe working conditions at 1 of its 6 international plants. According to the report, inadequate safety precautions at the manufacturing facility have raised concerns regarding the potential for future injuries.

Appendix B
High Severity Star Inc. Cited for Violation USA -The local Environmental Services Office has issued a fine to Star Inc. after a report released yesterday found the company's manufacturing plant to be in violation of several environmental statutes. The report concludes that Star Inc. was dumping untreated waste from its manufacturing plant into local streams and rivers. Many environmental groups are angry with Star and cite the fact that its dumping is causing serious harm to local fish and wildlife.
Star Inc. criticized for labor conditions USA -Star Inc. has drawn criticism from local interest groups after a report was released today which found poor working conditions at many of its international plants. According to the report, at some sites it was common for people to work up to 16 h a day at rates that were less than minimum wage. Many groups are angry with Star and cite the fact that its practices are causing serious harm to the local community.
Smith and Rhiney International Journal of Corporate Social Responsibility (2020) 5:1