Mixed or mismatch? Novel approaches to the interplay of prior associations and crisis response strategies

In today's marketplace marked by heightened public scrutiny and rising consumer expectations, brand crises have become not only more visible but also more consequential. Consumers increasingly expect companies to act competently and responsibly, and they often respond strongly when corporate actions or crisis responses fall short of these expectations. Consequently, crisis communication has become a critical mechanism through which firms attempt to contain reputational damage while also reaffirming, repairing, or redirecting stakeholder evaluations. Companies therefore invest substantial effort in cultivating favorable prior associations that shape consumers' perceptions, most notably through corporate ability (CA) and corporate social responsibility (CSR). Whereas some firms become strongly associated with a single reputational dimension, others cultivate broader, hybrid associations that combine both CA and CSR. This distinction becomes particularly important in crisis contexts, where consumers rely on prior associations as cognitive reference points for interpreting new and potentially threatening information.

Despite these efforts, favorable expectations can quickly collapse when a crisis occurs. For example, Paris Baguette, a renowned high-quality bakery chain in Korea, faced severe backlash and boycotts following concerns over poor working conditions and its handling of an employee's death (Lee, 2022), while United Airlines experienced substantial reputational damage after the forceful removal of a passenger and a poorly managed corporate response (Weller, 2017). These cases illustrate how mismanaged crisis communication can intensify negative consumer reactions and accelerate reputational damage.

Because crises often threaten the very associations firms have worked to build, understanding how prior associations shape responses to crisis communication is critical. Prior research has shown that consumers rely on pre-crisis brand associations as cognitive anchors when evaluating a company in times of crisis (Kim, 2011, 2014; Tao and Wilson, 2016). Some studies have therefore suggested that reaffirming a company's established CA or CSR strengths may mitigate negative reactions (Kim, 2014; Kim et al., 2009; Tao and Song, 2020). However, findings remain mixed, suggesting that response effectiveness may depend on more than simple alignment between prior associations and crisis type.

Prior research has focused largely on the content of prior associations, typically examining CA and CSR as separate assets, while paying less attention to their dimensional structure. In practice, however, many firms cultivate not only single-dimensional associations but also hybrid associations that combine both CA and CSR (e.g. Chen and Tao, 2020; Kim et al., 2019; Lee et al., 2018). This distinction may matter because consumers holding single-dimensional versus hybrid associations may interpret crisis responses differently. At the same time, prior research has not sufficiently explained why a response strategy that is strategically non-aligned with the damaged domain may sometimes be effective rather than counterproductive.

To address these gaps, this research examines how prior corporate associations and crisis response strategies jointly shape consumer reactions across two crisis contexts. More specifically, it investigates whether response novelty explains when a crisis response is effective, particularly when it highlights an alternative, favorable dimension of the brand. Study 1 focuses on a competence-related crisis by comparing CA versus hybrid associations, whereas Study 2 focuses on a morality-related crisis by comparing CSR versus hybrid associations. Focusing on response novelty, the study argues that post-crisis communication can do more than repair violated expectations: it can redirect consumer evaluations by activating an undamaged, favorable dimension of the brand. In doing so, the study extends Expectancy Violations Theory, reconceptualizes prior associations in terms of their dimensional structure (single-dimensional vs. hybrid), and qualifies the dominant congruence-based view of crisis communication by identifying conditions under which strategic mismatch can be beneficial.

For any company, fostering favorable consumer associations is crucial because such associations shape how stakeholders perceive and evaluate the firm (Brown and Dacin, 1997; Fombrun, 1996; Kim, 2014; Kim and Rader, 2010). Brand associations generally refer to the psychological perceptions and judgments that consumers form on the basis of a company's activities and communications. Prior research has typically distinguished two major dimensions of corporate associations: corporate ability (CA) and corporate social responsibility (CSR) (Biehal and Sheinin, 2007; Brown and Dacin, 1997). The CA dimension reflects a company's expertise and competence in producing high-quality products and services, whereas the CSR dimension reflects its moral and social commitments, such as ethical conduct, environmental sustainability, and concern for societal welfare (Kim, 2011, 2014; Sen and Bhattacharya, 2001). These associations shape consumer expectations and influence how consumers respond to subsequent corporate information (Kim, 2014).

Importantly, prior associations do not merely function as favorable brand assets; they also serve as cognitive anchors that guide information processing, including under conditions of uncertainty such as a crisis (Kim, 2011, 2014; Tao and Wilson, 2016). This role becomes especially consequential when the crisis appears to threaten the very dimension on which the company's reputation is built.

Most previous research has focused on associations in single-dimensional terms, examining either CA- or CSR-based reputations in isolation. However, contemporary firms increasingly cultivate dual-dimensional associations, or hybrid associations, that combine both CA and CSR dimensions (Chen and Tao, 2020; Kim et al., 2019; Lee et al., 2018). For example, firms like Microsoft and IKEA are recognized not only for product quality and innovation but also for environmental or social responsibility, thereby creating a broader and more multifaceted reputational structure (Jackson, 2026). This trend suggests that prior associations may matter not only because of their substantive content (i.e. whether they prioritize CA versus CSR), but also because of their dimensional structure (i.e. whether firms are perceived through a single dominant lens or a hybrid CA–CSR lens). Despite this significant industry shift, empirical evidence regarding the specific role and mechanisms of hybrid reputations remains scarce, leaving a critical gap in our understanding of how these complex structures function in practice.

To address this research gap, the present research classifies prior corporate associations into two categories: single-dimensional associations and dual-dimensional (hybrid) associations. Single-dimensional associations reflect a relatively narrow representation of the firm, centered primarily on either CA or CSR. By contrast, hybrid associations reflect a broader representation in which consumers simultaneously recognize both competence-related and responsibility-related strengths. This distinction is theoretically meaningful because consumers may not all approach crisis information with the same cognitive structure. Consumers holding a single-dimensional association may rely on a relatively narrow evaluative frame, whereas those with hybrid associations may draw on a broader set of favorable meanings when interpreting crisis-related information. As a result, hybrid associations may provide greater resilience when one dimension of the brand becomes threatened, while single-dimensional associations may make consumers more vulnerable when the crisis directly targets the firm's central reputational strength. Thus, distinguishing between single-dimensional and dual-dimensional associations provides the conceptual foundation for understanding how consumers respond to post-crisis communication.

This study draws on Expectancy Violations Theory (EVT) to understand why the dimensional structure of prior associations matters in crisis settings. EVT posits that individuals form expectations on the basis of prior information and respond when those expectations are violated (Afifi and Metts, 1998). In the corporate context, prior brand associations shape what consumers expect from a firm. A company associated strongly with CA is expected to behave competently and deliver reliable products or services, whereas a company associated strongly with CSR is expected to behave responsibly and uphold ethical or social commitments. When a crisis occurs, these prior expectations become the basis against which the firm is judged.

A crisis therefore creates a discrepancy between favorable prior expectations and the reality of a negative event. When that discrepancy is especially pronounced, consumers may revise their evaluations downward and respond more negatively toward the firm. This mechanism may be particularly severe when the crisis directly aligns with the company's dominant prior association (Hegner et al., 2016; Kim, 2014; Lewin and Warren, 2025; Smith and Rhiney, 2020; Wagner et al., 2009; Wei and Diddi, 2022). For example, a product-harm crisis may be especially damaging for a firm known primarily for competence (Hegner et al., 2016; Kim, 2014), whereas a morality-related crisis may be especially damaging for a firm known primarily for social responsibility (Lewin and Warren, 2025; Smith and Rhiney, 2020; Wagner et al., 2009; Wei and Diddi, 2022). In such cases, the crisis threatens not only the firm's overall reputation, but also the very basis on which consumers previously evaluated the firm positively. This is consistent with research suggesting that congruence between the crisis domain and the firm's core association can produce a “boomerang effect,” intensifying negative consumer reactions rather than containing them (Lee and Chung, 2018; Sohn and Lariscy, 2015).

This logic has direct implications for association-based crisis response strategies. Within the CA–CSR framework, crisis responses are often categorized into CA responses, which emphasize past competence and capability, and CSR responses, which emphasize moral obligations, social responsibility, and concern for stakeholder welfare (Coombs and Holladay, 2015; Tao, 2023; Tao and Song, 2020). A straightforward implication of congruence logic would be that firms should respond to crises by reaffirming the same association on which their reputation is built. However, prior empirical evidence is mixed. Some studies suggest that aligned responses can mitigate reputational harm, whereas others show little advantage—or even potential drawbacks—when firms emphasize the very dimension that has already been called into question (Kim et al., 2009; Kim, 2014; Park et al., 2024). These inconsistent findings suggest that the effectiveness of CA-versus CSR-based responses cannot be understood solely in terms of fit.

Rather, response effectiveness likely depends on how consumers interpret post-crisis messages in light of their prior associational structure. If EVT explains why crises trigger strong negative reactions, the next question is what kind of post-crisis message can soften those evaluations. The present research argues that one important answer lies in response novelty. More specifically, it asks whether consumers respond differently to crisis communication depending on whether their prior associations are single-dimensional or hybrid, particularly when the crisis directly threatens a focal associational dimension. To investigate these dynamics, the present research examines two parallel crisis contexts: Study 1 focuses on a product-harm (competence-related) crisis by comparing CA versus hybrid associations, whereas Study 2 examines a morality-harm (CSR-related) crisis by comparing CSR versus hybrid associations. This parallel design allows the research to compare the vulnerabilities of single-dimensional associations with the broader resilience of hybrid associations across two theoretically relevant crisis domains.

Response novelty refers to the degree of newness or uniqueness in a crisis response's content relative to consumers' prior knowledge or the thematic domain of the crisis (Tao and Song, 2020). In crisis communication, novelty matters because consumers do not process all messages equally. A response may be perceived as more novel when it presents information that is relatively unfamiliar in light of what consumers already know about the company or in light of the crisis domain itself. Such novel information tends to attract attention, increase salience, and interrupt routine or biased defensive processing (Ahmadjian and Robinson, 2001; Förster et al., 2010; Pfarrer et al., 2008; Wyer, 2019). Rather than simply reaffirming the association that has already been damaged, a novel response may introduce an alternative, favorable perspective that encourages consumers to reconsider the firm from a different angle (Tao and Song, 2020).

This possibility is especially relevant following expectancy violations. When a crisis undermines a salient prior association, consumers may become locked into a negative interpretive frame. Under such conditions, a response that merely repeats the damaged association may do little to shift evaluations. By contrast, a response that highlights an undamaged but favorable dimension may function as a buffering cue, redirecting attention and facilitating a more favorable post-crisis reappraisal. In this way, response novelty helps explain why a strategically non-aligned response may sometimes be effective rather than counterproductive.

The extent to which a response is perceived as novel should therefore depend on the interplay between the response strategy and the consumer's prior associational structure. In a competence-related crisis, a CSR response should be especially novel to consumers who hold single-dimensional CA associations, because CSR information lies outside their dominant evaluative frame. The same response should be less novel to consumers with hybrid associations, who already recognize the company along both CA and CSR dimensions. Conversely, in a morality-related crisis, a CA response should appear especially novel to consumers who hold single-dimensional CSR associations, whereas the same response should appear less novel to consumers with hybrid associations. Thus, the following hypotheses are proposed:

Beyond shaping novelty perceptions, response novelty should also influence broader post-crisis consumer reactions. When a crisis response introduces unexpected but favorable information, it can disrupt consumers' established expectancy framework and redirect attention away from the damaged association. Prior research on information processing suggests that novel stimuli capture attention, increase salience, and interrupt biased or defensive reasoning (Sheinin et al., 2011; Wyer, 2019). In this sense, response novelty should function as a mediating mechanism linking prior associational structure and crisis response strategy to post-crisis consumer outcomes. By shifting attention toward an undamaged, favorable dimension of the brand, novelty can soften the negative consequences of expectancy violations, buffer the EVT “boomerang effect,” and facilitate more favorable post-crisis judgments (Tao and Song, 2020). Novelty, therefore, is not merely an incidental feature of the message; it is the process through which an alternative response strategy becomes persuasive.

To capture these effects comprehensively, the present research examines three dependent variables: attitudes toward the company, purchase intention, and forgiveness intention. Attitudes toward the company capture consumers' overall evaluative judgments following a crisis, making them a fundamental indicator of reputational damage and recovery (Tao and Song, 2020). Purchase intention reflects the behavioral implications of crisis communication for marketplace outcomes and thus indicates whether more favorable evaluations translate into consumption-related responses (Hegner et al., 2016). Forgiveness intention captures consumers' willingness to restore a damaged relationship with the firm, making it especially relevant in post-crisis recovery contexts (Yoruk et al., 2025). Together, these variables capture evaluative, behavioral, and relational dimensions of crisis response effectiveness. Accordingly, the following hypotheses are proposed:

The purpose of Study 1 is to test the joint effects of brand associational dimensions and crisis response types in a competence-related context. A 2 (Brand's dimensional associations: CA vs. hybrid) × 2 (Brand crisis response type: CA response vs. CSR response) between-subjects experimental design was employed. The brand's dimensional associations and crisis response types were manipulated, while participants' perceived novelty, attitudes toward the company, purchase intentions, and forgiveness intentions were measured. Pre-crisis company attitudes were measured as a covariate to control for potential confounding effects (Tao and Song, 2020).

A fictitious electronics company, Bettertronics, was created to minimize potential confounding effects associated with preexisting brand attitudes (see Appendix A). To manipulate brand association type, participants were exposed to either a CA-focused company description or a hybrid description incorporating both CA and CSR elements (see Appendices B and D). To induce a competence-related crisis, participants then read a news article describing a laptop explosion caused by the company's negligent use of faulty batteries (see Appendix E), based on prior research (Jeon and Baeck, 2016; Tao, 2018). Finally, participants were presented with either a CA-focused or CSR-focused crisis response message (see Appendices G and H).

A total of 114 responses were collected from U.S. participants via Amazon Mechanical Turk (MTurk). To ensure data quality, two instructional attention-check items were embedded in the survey, one in the middle and one near the end, each asking participants to select “Agree.” After applying predetermined screening criteria, including attention-check failures, 111 usable responses were retained for the final analysis (female: 46.9%; M_age = 41.41, SD = 13.99, range = 22–78). Participants were randomly assigned across four experimental conditions: CA association with a CA-focused apology (n = 26), CA association with a CSR-focused apology (n = 23), hybrid association with a CA-focused apology (n = 32), and hybrid association with a CSR-focused apology (n = 30). A post hoc power analysis conducted using G*Power (Faul et al., 2007) indicated that the ANCOVA interaction effect achieved a statistical power of 0.85, suggesting an adequate level of statistical power. Participants were first given an overview of the study and asked for their consent to participate. They were then presented with a brief description of the fictitious brand, Bettertronics and randomly assigned to one of two association conditions: CA or Hybrid. After reading about the company, they were asked to rate their thoughts about Bettertronics. Next, they were exposed to a news article describing a laptop explosion caused by the company's careless use of defective batteries and again asked to evaluate the company. Participants were then randomly assigned to one of two corporate crisis response conditions: CA-focused vs. CSR-focused crisis response messages. After reading the apology message, they completed measures assessing their responses to the message and the experimental manipulations. Lastly, they were asked demographic questions before being debriefed and thanked. The survey took about 18–20 min, and participants were compensated with $1.00 for their participation.

Perceived novelty was measured using 4 items on a 7-point semantic scale (Tao and Song, 2020) (e.g. “The content of Bettertronics's response provides no/substantial amount of new information to me,” “I have already known all/known little of the information described in Bettertronics' response,”) (M = 4.28, SD = 1.50, α = 0.94). Attitude toward the company was measured with 3 items on a 7-point bipolar scale (i.e. unfavorable/favorable, bad/good, and negative/positive) (M = 4.60, SD = 1.54, α = 0.97). Purchase intention was assessed using 3 items on a 7-point bipolar scale (i.e. unlikely/likely, improbable/probable, and impossible/possible) (M = 3.94, SD = 1.76, α = 0.97). Forgiveness intention was assessed using a single item on a 7-point Likert scale (adapted from Xie and Peng, 2009) (e.g. “I would forgive the company based on their response.”) (M = 4.38, SD = 1.40)

A series of t-tests were performed to check manipulations. First, to check the manipulations of the brand's dimensional associations, participants were asked to rate the extent to which they thought the company was competent and socially responsible. To measure perceived corporate competence, participants rated their degree of agreement for five items on a 7-point Likert scale, with 1 being “strongly disagree” and 7 being “strongly agree” (e.g. “I think Bettertronics is a competent laptop producer.” “I associate Bettertronics with innovative products,” and “I associate Bettertronics with good quality products.”) (α = 0.88). To measure perceived CSR, participants rated their degree of agreement for five items on a 7-point Likert scale, with 1 being “strongly disagree” and 7 being “strongly agree” (e.g. “I think Bettertronics is a socially responsible company.” “I associate Bettertronics with taking good care of their employees,” and “I associate Bettertronics with environmental responsibility.”) (α = 0.94). (Brown and Dacin, 1997; Kim, 2011, 2014; Kim and Rader, 2010; Sen and Bhattacharya, 2001; Tao, 2018). An independent t-test indicated that participants who were exposed to both the positive CA and hybrid associations condition showed higher positive CA associations (M_CA = 6.27 vs. M_hybrid = 6.29; t = −0.584, p > 0.05). No significant difference in positive CA associations was found between the CA and hybrid conditions. On the other hand, participants who were exposed to the positive CA associations condition indicated that they had significantly lower positive CSR associations than those with hybrid associations (M_CA = 4.65 vs. M_hybrid = 6.25; t = −8.279, p < 0.001).

Additionally, to check whether the crisis response types were appropriately manipulated, participants were asked to rate their thoughts on whether the crisis response message focused on competence or CSR for 2 items on a 7-point bipolar scale (i.e. “1 = Competence/the company's ability, 7 = corporate social responsibility/the company's morality”; α = 0.80). An independent t-test revealed that participants exposed to the CA crisis response evaluated that the response was focused more on the company's competence and ability, while those exposed to CSR crisis response evaluated that the response was focused more on the company's CSR and morality (M_CA-response = 2.94 vs. M_CSR-response = 5.59; t = −10.401, p < 0.001). The manipulations of the independent variables were successful.

To examine the proposed hypothesis (H1a), an analysis of covariance (ANCOVA) was conducted with two independent variables (i.e. brand's dimensional associations and crisis response type) and a covariate (i.e. pre-crisis company attitudes) on the dependent variables (i.e. perceived novelty, attitudes toward the company, purchase intention, forgiveness intention).

In terms of perceived novelty of a crisis response (H1a), the results demonstrated significant main effects of prior dimensional associations (M_CA = 4.47, SD_CA = 1.71; M_hybrid = 4.13, SD_hybrid = 1.31; F(1,106) = 4.88, p = 0.029, η_p² = 0.04) and the crisis response types (M_CA = 3.99, SD_CA = 1.59; M_CSR = 4.60, SD_CSR = 1.34; F(1,106) = 6.13, p = 0.015, η_p² = 0.06) on crisis response novelty. The results also indicated pre-crisis attitudes (F(1,106) = 5.66, p < 0.05, η_p² = 0.05) significantly influenced crisis response novelty. Further, the findings showed a significant interaction effect of prior dimensional associations, and the crisis response types on crisis response novelty (F(1,106) = 4.31, p < 0.05, η_p² = 0.04). Table 1 illustrates ANCOVA results.

A planned contrast was performed to evaluate those relations in more detail. The results indicated a significant difference, when CSR response was provided after a competence-related crisis, participants with CA associations showed a higher crisis response novelty (M = 5.13, SD = 1.38) than those with hybrid associations (M = 4.20, SD = 1.18; F(1,106) = 8.73, p < 0.01, $η$_p² = 0.076). On the other hand, when CA response was provided, no significant difference was found between participants in the CA associations (M = 3.89, SD = 1.79) and those in the hybrid associations (M = 4.07, SD = 1.44; F (1,106) = 0.05, p > 0.05, $η$_p² = 0.000). Figure 1 presents the relationship in detail. Therefore, the results supported H1a.

To test H2, a moderated mediation analysis was performed using the PROCESS macro for SPSS (Hayes, 2012, Model 8, 5,000 bootstrap samples). Prior dimensional associations were entered as the independent variable, crisis response type as the moderator, perceived response novelty as the mediator, and attitude toward the company, purchase intention, and forgiveness intention as the dependent variables. Pre-crisis attitude was included as a covariate. The results revealed a significant interaction effect of prior dimensional associations and crisis response type on perceived novelty (β = −0.57, SE = 0.27, p < 0.05). In addition, the interaction did not directly predict attitude toward the company (p = 0.899), purchase intention (p = 0.994), and forgiveness intention (p = 0.109). In contrast, the results revealed that the significant effects of perceived novelty on attitude toward a company (β = 0.31, SE = 0.10, p = 0.002, 95% CI [0.1162, 0.4973]), purchase intention (β = 0.34, SE = 0.11, p = 0.003, 95% CI [0.1159, 0.5642]), forgiveness intention (β = 0.24, SE = 0.09, p = 0.005, 95% CI [0.0736, 0.4125]). Table 2 presents detailed coefficient pathways in the moderated mediation model.

This implies that, in a competence crisis, crisis response novelty mediated the interaction effects of prior dimensional associations and corporate response strategy on attitude toward the company, purchase intention, and forgiveness intention, after controlling for pre-crisis attitude. Table 3 and Figure 2 illustrate the moderated mediation results. Therefore, the results supported H2.

The findings of Study 1 show that prior associational structure and crisis response strategy jointly shape consumers' evaluations in a competence-related crisis. When a CSR-focused response was provided, consumers with single-dimensional CA associations perceived greater response novelty than those with hybrid associations. In contrast, when a CA-focused response was provided, perceived novelty did not differ between the two association conditions. This finding is theoretically important because it suggests that the effectiveness of a crisis response depends not only on whether it aligns with the damaged domain, but also on how consumers cognitively organize the brand before the crisis. In this respect, Study 1 extends an EVT-based account of crisis processing by showing that a strategically non-aligned response can prompt reappraisal when it introduces novel and favorable information. Perceived novelty increased attitudes toward the company, purchase intentions, and forgiveness intentions. Although the interaction between prior associations and response strategy did not directly affect these outcomes, it operated indirectly via novelty, indicating that non-aligned responses can improve evaluations by introducing unexpected positive information.

The findings further demonstrate that the weaker novelty effect observed among consumers with hybrid associations reflects the greater resilience of multidimensional reputational structures. When CA is damaged, consumers with hybrid associations are less influenced by response strategies that either reaffirm CA or emphasize CSR because they can draw on a broader set of favorable brand meanings. This multidimensional representation provides a more stable interpretive framework, reducing reliance on any single response cue. Thus, Study 1 shows that associational dimensionality shapes post-crisis responses through the mechanism of response novelty.

Study 2 employed a 2 (brand's dimensional associations: CSR vs. hybrid) × 2 (crisis response type: CA response vs. CSR response) between-subjects design to test H1b and H3 in a morality-related crisis context. As in Study 1, the brand's dimensional associations and crisis response types were manipulated, perceived novelty and post-crisis consumer responses were measured, and pre-crisis company attitude was included as a covariate.

To maintain internal consistency across studies, Study 2 used the same fictitious electronics company, Bettertronics. To manipulate brand association type, participants were exposed to either a CSR-focused company description or the same hybrid description used in Study 1 (see Appendices C and D). To induce a morality-related crisis, participants read a news article describing severe environmental damage caused by the company's careless waste handling (see Appendix F), based on prior research (Jeon and Baeck, 2016; Tao, 2018). Participants then received either a CA-focused or CSR-focused crisis response message (see Appendices G and H).

A total of 103 responses were collected from U.S. participants via Amazon Mechanical Turk (MTurk). Using the same screening procedure as in Study 1, including two instructional attention checks, 101 usable responses were retained for the final analysis (female: 42.6%; M_age = 39.63, SD = 12.60, range = 23–74). Participants were randomly assigned across four experimental conditions: CSR association with a CA-focused apology (n = 25), CSR association with a CSR-focused apology (n = 21), hybrid association with a CA-focused apology (n = 28), and hybrid association with a CSR-focused apology (n = 27). The corresponding post hoc power analysis for Study 2 indicated that the ANCOVA interaction effect achieved a statistical power of 0.87. The procedure followed the same sequence as Study 1, except that participants viewed either the CSR-focused or hybrid company description and read a morality-related crisis article. After reading the assigned company description, participants reported their pre-crisis attitudes toward Bettertronics. They then read the crisis article, evaluated the company again, viewed the assigned crisis response, and completed the outcome measures, manipulation checks, demographic questions, and debriefing. Participants received $1.00 for completing the study.

Study 2 used the same measures as Study 1. Perceived novelty was assessed with 4 items on a 7-point semantic scale (Tao and Song, 2020) (M = 4.01, SD = 1.39, α = 0.85). Attitude toward the company (M = 4.57, SD = 1.34, α = 0.94) and purchase intention (M = 4.65, SD = 1.53, α = 0.95) were measured using 3 items on 7-point bipolar scales. Forgiveness intention was assessed using a single 7-point Likert item adapted from Xie and Peng (2009) (M = 4.63, SD = 1.64).

Identical measures from Study 1 were adopted to check manipulations. An independent t-test indicated that participants who were exposed to both the positive CSR and hybrid associations condition showed higher positive CSR associations (M_CSR = 6.18 vs. M_hybrid = 6.17; t = 0.108, p > 0.05). No significant difference in positive CSR associations was found between the CSR and hybrid conditions. On the other hand, participants who were exposed to the positive CSR associations condition indicated that they had significantly lower positive CA associations than those with hybrid associations (M_CSR = 5.84 vs. M_hybrid = 6.36; t = −3.614, p < 0.001).

Additionally, in terms of the crisis response types, an independent t-test revealed that participants who were exposed to CA crisis response evaluated that the response was focused more on company's competence and ability, while those who were exposed to CSR crisis response evaluated that the response was focused more on company's CSR and morality (M_CA-response = 3.47 vs. M_CSR-response = 5.40; t = −6.843, p < 0.001). The manipulations of the independent variables were successful.

To examine the proposed hypotheses (H1b), an ANCOVA was conducted with two independent variables (i.e. brand's dimensional associations and crisis response type) and a covariate (i.e. pre-crisis company attitudes) on the dependent variables (i.e. perceived novelty, attitudes toward the company, purchase intention, forgiveness intention).

Regarding the perceived novelty of crisis response (H1b), the results demonstrated no significant main effects of crisis response types (F(1,96) = 1.66, p = 0.200, η_p² = 0.017) on crisis response novelty, while the prior dimensional associations significantly influenced crisis response novelty (M_CSR = 4.33, SD_CSR = 1.35; M_hybrid = 3.76, SD_hybrid = 1.39; F(1, 96) = 4.48, p < 0.05, η_p² = 0.045). The results also indicated no significant influence of pre-crisis attitudes (F(1,96) = 0.300, p > 0.05, η_p² = 0.003) on crisis response novelty. Further, the findings showed a significant interaction effect of prior dimensional associations, and the crisis response types on crisis response novelty (F(1,96) = 4.44, p < 0.05, η_p² = 0.044) (see Table 4).

Planned contrasts were conducted to examine this interaction. When a CA-focused apology was presented following a morality-related crisis, participants with CSR associations perceived greater response novelty (M = 4.41, SD = 1.32) than did those with hybrid associations (M = 3.29, SD = 1.24; F (1, 96) = 9.42, p < 0.01, η_p² = 0.089). In contrast, when a CSR-focused apology was presented, perceived response novelty did not differ significantly between participants with CSR associations (M = 4.23, SD = 1.41) and those with hybrid associations (M = 4.24, SD = 1.39; F (1, 96) = 0.002, p > 0.05, η_p² = 0.000). Therefore, the results supported H1b (see Figure 3).

To test H3, the same moderated mediation procedure was applied using PROCESS Model 8 with 5,000 bootstrap samples. Prior dimensional associations were entered as the independent variable, crisis response type as the moderator, perceived response novelty as the mediator, and pre-crisis attitude as a covariate. Attitude toward the company, purchase intention, and forgiveness intention served as the dependent variables. The results revealed a significant interaction effect of prior dimensional associations and crisis response type on perceived novelty (β = 1.13, SE = 0.54, p < 0.05). The interaction did not directly predict attitude toward the company (p = 0.550), purchase intention (p = 0.385), or forgiveness intention (p = 0.176). In contrast, the results revealed that the significant effects of perceived novelty on attitude toward a company (β = 0.37, SE = 0.10, p < 0.001, 95% CI [0.1869, 0.5474]), purchase intention (β = 0.30, SE = 0.11, p = 0.006, 95% CI [0.0863, 0.5129]), forgiveness intention (β = 0.39, SE = 0.11, p < 0.001, 95% CI [0.1764, 0.6120] (see Table 5).

This implies that, in a morality crisis, crisis response novelty mediated the interaction effects of positive prior dimensional associations and corporate response strategy on attitude toward the company, purchase intention, and forgiveness intention, after controlling for pre-crisis attitude. Therefore, H3 was supported (see Table 6 and Figure 4).

Study 2 shows that the proposed pattern extends to a morality-related crisis. When a CA-focused response followed the CSR-related crisis, consumers with single-dimensional CSR associations perceived greater response novelty than those with hybrid associations, whereas no corresponding difference emerged for a CSR-focused response. Consistent with Study 1, these findings suggest that strategic mismatch generates novelty when it activates a favorable but previously underemphasized brand dimension. Perceived novelty, in turn, was positively associated with attitudes toward the company, purchase intention, and forgiveness intention. Consumers with hybrid associations, by contrast, again exhibited weaker novelty effects, reflecting the broader evaluative framework provided by representations that already incorporate both CA and CSR meanings. Because the alternative dimension is already familiar within such representations, a non-aligned response is less likely to introduce genuinely new information. Thus, although hybrid associations may offer resilience by providing multiple favorable meanings on which consumers can draw, they also limit the novelty generated by strategic mismatch. Together, the findings identify response novelty as a mechanism linking prior associational structure and crisis response strategy to consumer reactions, while revealing an important boundary condition: highlighting an alternative favorable dimension may be effective when that dimension is underrepresented in consumers' prior associations, but may offer little novelty when it is already embedded in a hybrid representation.

As businesses navigate increasingly complex crises under heightened public scrutiny, understanding how consumers interpret post-crisis communication becomes critical. In this context, the present research examined how prior corporate associations and crisis response strategies jointly shape consumer reactions across two crisis contexts. Across both studies, the novelty of the response depended on consumers' prior brand associations: highlighting an alternative favorable brand dimension increased perceived novelty among consumers with single-dimensional associations, but reduced novelty among those with hybrid associations.

In addition, response novelty mediated the effects of prior associations and crisis response strategy on post-crisis consumer responses, underscoring its central role in shaping post-crisis evaluations. Taken together, the findings advance crisis communication theory in three related ways.

First, the present research extends Expectancy Violations Theory (EVT) by showing that post-crisis communication does more than simply repair violated expectations. Prior EVT-based crisis research has largely emphasized the negative consequences of expectancy violations, particularly when a crisis directly undermines a favorable prior association or contradicts the reputational basis on which consumers evaluate the firm (Kim, 2014; Sohn and Lariscy, 2015; Tao, 2018). Related research further suggests that product-harm and morality-related crises can damage different dimensions of consumer trust and corporate evaluation, depending on whether the crisis threatens competence- or responsibility-based expectations (Dutta and Pullig, 2011; Hegner et al., 2016; Lee et al., 2021). Consistent with this line of work, the present findings show that prior associations provide an expectancy framework through which consumers interpret crisis information. Yet they also add an important refinement: once expectations have been violated, a crisis response can still play a buffering role when it introduces novel, favorable information through an alternative and undamaged brand dimension. In this sense, response novelty functions as a mechanism of post-crisis reappraisal. Rather than merely restoring what was damaged, a novel response can redirect attention and enable consumers to evaluate the company through a different interpretive route (Tao and Song, 2020). Thus, this study extends EVT by showing how response communication can mitigate the effects of expectancy violations by generating novelty.

Second, this study reconceptualizes prior corporate associations not only in terms of their substantive content, namely CA versus CSR, but also in terms of their dimensional structure. Prior empirical research has established that CA and CSR associations shape consumer evaluations and can function as important reputational resources in crisis contexts (Brown and Dacin, 1997; Kim, 2014; Sen and Bhattacharya, 2001; Tao, 2018). More recent work has examined how firms use CA- and CSR-based communication strategies before and after crises, showing that consumers' pre-crisis associations and response strategies jointly shape post-crisis outcomes (Tao and Song, 2020). However, much of this work has focused on which type of association or response strategy is more effective under particular crisis conditions. By contrast, the present findings show that what also matters is whether consumers hold a single-dimensional or hybrid representation of the company. This distinction offers a new insight for crisis communication theory and management because dimensional structure conditions both vulnerability and resilience in crisis processing. Single-dimensional associations appear to heighten sensitivity to novel, non-aligned responses, whereas hybrid associations provide a broader and more stable basis for post-crisis evaluation. In other words, the way consumers cognitively organize a brand before a crisis is not merely background information; it is a theoretically meaningful condition that shapes how crisis responses are interpreted and whether they are likely to be effective.

Third, the findings qualify the dominant fit- or congruence-based logic in crisis communication research. Existing literature has often suggested that responses are more effective when they align with the basis of consumers' prior attitudes, address the focal crisis domain, or match the nature of the reputational threat (Dutta and Pullig, 2011; Pullig et al., 2006; Tao and Song, 2020). Yet prior empirical findings have not always converged. Some studies suggest that association-consistent responses can mitigate reputational harm, whereas others indicate that favorable prior associations or aligned responses may produce limited benefits, or even backfire, when the crisis directly threatens the core basis of the firm's reputation (Kim, 2014; Sohn and Lariscy, 2015; Tao and Song, 2020). The present study helps explain these mixed findings by identifying response novelty as a mechanism through which strategic mismatch can sometimes become persuasive. The present study does not reject fit-based logic altogether. Rather, it identifies an important boundary condition: under certain circumstances, especially when consumers hold single-dimensional associations, a strategic mismatch between prior associations and response strategy can be beneficial because it generates response novelty and facilitates reappraisal. This suggests a limited but meaningful conceptual shift from asking whether a response simply “fits” to asking when strategic incongruence may be more effective. In doing so, the study also implies a broader and more diversified direction for crisis management guidelines, namely that effective crisis responses may sometimes require not reaffirming the damaged core, but strategically activating an alternative favorable dimension of the brand.

Taken together, the two studies also clarify the distinct role of hybrid associations. Prior research on hybrid or multidimensional reputations suggests that stakeholders may respond differently to crises when organizations are associated with multiple favorable dimensions rather than a single dominant reputational basis (Lee et al., 2018). Building on this view, the present research shows that hybrid associations are not merely stronger reputational assets, but distinct cognitive configurations that shape how consumers interpret crisis responses. Because hybrid associations incorporate both CA and CSR meanings, they provide consumers with multiple favorable dimensions through which to understand the company. At the same time, however, this broader representational structure reduces the likelihood that a response emphasizing either dimension will be perceived as genuinely new. This helps explain why strategically non-aligned responses generated greater novelty among consumers with single-dimensional associations, whereas corresponding novelty effects were weaker among those with hybrid associations. Thus, hybrid associations may provide a broader evaluative framework, but they do not ensure that strategic incongruence will facilitate reappraisal, particularly when the alternative dimension is already salient in consumers' prior representation of the company. This nuance adds to existing knowledge by showing that hybrid associations shape not only the range of favorable meanings available to consumers, but also the extent to which a strategically non-aligned response can generate novelty.

From a practical standpoint, the findings suggest that managers should treat the dimensional structure of consumers' pre-crisis associations as a strategic input rather than as background reputation, and response novelty as audience-relative rather than as an inherent message property. Before selecting a response strategy, firms should assess whether consumers primarily associate the brand with CA, CSR, or both, and whether the proposed response introduces a credible and previously underrepresented basis for evaluation. For brands with single-dimensional associations, activating an underdeveloped, undamaged dimension may generate novelty and broaden the basis on which the company is evaluated. For brands with hybrid associations, however, simply shifting emphasis from CA to CSR, or vice versa, is unlikely to appear meaningfully new because both dimensions are already embedded in consumers' brand representations. This creates a strategic paradox: hybrid associations provide a broader repertoire of favorable meanings, yet reduce the novelty value of switching between them. In such cases, novelty may need to come from the substance or execution of the response—such as concrete corrective actions, new evidence, stakeholder involvement, or an unexpected communication format—rather than from the focal association alone. Strategic incongruence should therefore be used selectively, when the alternative dimension is both credible and underrepresented, and should complement rather than replace an appropriate response to the focal crisis. Regular association audits and audience segmentation can help firms identify where novelty is likely to emerge. Effective crisis communication is thus not a matter of choosing fit or mismatch in the abstract, but of matching the source of novelty to what consumers already know and expect from the brand.

While this study offers valuable insights, several limitations suggest directions for future research. First, the use of a fictitious company and experimentally constructed crisis scenarios may limit the generalizability of the findings to real-world crises, where brand histories, prior attitudes, and consumer–brand relationships are often more complex and deeply rooted. Future studies could strengthen external validity by examining actual crisis events or well-known brands. Second, the present research focused on two crisis contexts and two response domains, CA and CSR. Future research should examine whether the same novelty-based process extends to other crisis response message characteristics, such as emotional versus rational appeals, sincerity, corrective action, or attribution of fault in apology messages. Third, although the findings suggest that hybrid associations serve as a buffering resource, the present design does not fully identify the boundary conditions of this effect. Future studies should examine when hybrid associations merely buffer against damage and when they can be more actively leveraged in post-crisis recovery. Relatedly, given that this study focused on perceived novelty as a response-strategy mechanism, future studies could investigate additional mediators that explain when matched or mismatched crisis-response strategies enhance or undermine the protective effects of hybrid associations. Finally, additional crisis-related variables, such as crisis responsibility, crisis severity, and consumers' emotional responses, may further refine the proposed framework. Future work should also investigate whether the effects observed here persist, weaken, or change over time as consumers reassess the firm in the longer term.

This study involved experimental research with human participants. All procedures involving human subjects were conducted in accordance with the ethical standards of the institutional research committee. The study protocol was reviewed and approved by the Institutional Review Board of The University of Texas at Austin.

The supplementary material for this article can be found online.