POLITICAL INTENSITY, CUSTOMER LOYALTY, AND BRAND ADVOCACY: A NOVEL PERSPECTIVE ON POLITICAL POLARIZATION WITH STRATEGIC IMPLICATIONS By Udit Sharma A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Business Administration – Marketing – Doctor of Philosophy 2024 ABSTRACT Political polarization continues to grow in the U.S., prompting considerable research in marketing on the effects of individuals’ political identities on their behaviors as customers. Underpinning this stream of research is an assumption that individuals’ political orientations drive their customer behaviors. While this approach has generated important insights, it covers only one dimension of political polarization – ideologically-driven asymmetries differentiating those on the “left” versus the “right” – and thus provides an incomplete picture of the implications of political polarization to marketers. The authors introduce a previously overlooked dimension of political polarization – political intensity – focusing on how those closer to the far ends of the political spectrum (on the “left” and “right”) are similar to each other and different from those with weaker and moderate identities. They propose distinct effects of political intensity (beyond those of political orientation) on customer loyalty and brand advocacy, outline the mechanisms underlying these effects, and show that they are stronger in market-based contexts (competitor-related, product launches and promotions; firm-related, failures) that otherwise threaten customer-brand relationships, but weaker in a nonmarket context (customer- brand political misalignment) contradicting customers’ foundational political values. These relationships are tested across nine studies that utilize a variety of data sources and estimation approaches. Copyright by UDIT SHARMA 2024 ACKNOWLEDGEMENTS Foremost, it is imperative that I thank God for providing me with the necessary strength and drive. Furthermore, I extend my sincere gratitude to Dr. Tomas Hult, Dr. Forrest Morgeson, Dr. Suman Basuroy, and Dr. Ayalla Ruvio, who served as invaluable members of my dissertation committee, offering unwavering support and feedback in the development of my research. I am also thankful to the other members of the faculty at Michigan State University for equipping me with the necessary resources and opportunities for academic growth. My heartfelt thanks go to my colleagues in the program for enriching my experience over the course of the last five years. Above all, I am deeply thankful to my family for their unwavering support, encouragement, and the sacrifices they have made throughout this journey. iv TABLE OF CONTENTS INTRODUCTION .......................................................................................................................... 1 THEORETICAL FRAMEWORK AND HYPOTHESES .............................................................. 7 METHODOLOGY ....................................................................................................................... 18 DISCUSSION AND IMPLICATIONS ......................................................................................... 41 FIGURES ...................................................................................................................................... 48 TABLES ........................................................................................................................................ 49 REFERENCES ............................................................................................................................. 53 APPENDIX A: A REVIEW OF LITERATURE ........................................................................... 65 APPENDIX B: DESCRIPTION OF THE CONTROL VARIABLES.......................................... 68 APPENDIX C: PRELIMINARY EVIDENCE ............................................................................. 71 APPENDIX D: CORRELATIONS AND DESCRIPTIVE STATISTICS (STUDY 1A) .............. 73 APPENDIX E: ALTERNATE MEASURE OF INTERNAL THREAT ....................................... 74 APPENDIX F: ALTERNATE (CURVILINEAR) ESTIMATION STUDY 1A ........................... 77 APPENDIX G: ADDRESSING ENDOGENEITY (NEIGHBOR POLITICAL INTENSITY) ... 80 APPENDIX H: ADDRESSING ENDOGENEITY (POLITICAL ADVERTISING) ................... 86 APPENDIX I: TABLES AND FIGURES FOR STUDY 1B ........................................................ 91 APPENDIX J: TABLES AND FIGURES FOR STUDY 2A ....................................................... 93 APPENDIX K: TABLES AND FIGURES FOR STUDY 2B ...................................................... 97 APPENDIX L: DETAILS OF COMPLEMENTARY STUDY 2BW ......................................... 101 APPENDIX M: TABLES AND FIGURES FOR STUDY 3 ...................................................... 106 APPENDIX N: TABLES AND FIGURES FOR STUDY 4 ....................................................... 109 APPENDIX O: TABLES AND FIGURES FOR STUDY 5 ........................................................ 111 v APPENDIX P: DETAILS OF COMPLEMENTARY STUDY 5W .............................................115 APPENDIX Q: CURVILINEAR ESTIMATION (REMAINING STUDIES) ........................... 121 APPENDIX R: POLITICAL INTENSITY AND POLITICAL ORIENTATION ...................... 127 APPENDIX S: HIGH COMPLEXITY CONTEXTS ................................................................. 135 vi INTRODUCTION Recently, a heightened interest among both marketing researchers and practitioners in the “politics of consumption” and its strategic implications for brands has emerged (e.g., McConnell et al. 2018; Panagopoulos et al. 2020). While the convergence and intermingling of politics and customer attitudes and behaviors is not a new phenomenon (Bennett 1998), political dynamics have changed. These changed dynamics have resulted in the migration of individuals from the moderate center of the political spectrum towards the extremes, resulting in dramatically amplified political polarization (Boxell, Gentzkow, and Shapiro 2022; McConnell et al. 2018). For example, in a study of U.S. citizens conducted by the Pew Research Center, the gap between the opinions of Republicans and Democrats on ten key political values is the largest it has been since measurement began (Pew 2017). Similarly, Republican and Democratic members of the U.S. Congress indicate being more divided than at any time in the last 50 years (Pew 2022). Indeed, since 1980, political polarization in the U.S. has increased more dramatically than in virtually any other developed democracy (Boxell, Gentzkow, and Shapiro 2022). Research suggests that growing political polarization has spill-over effects that influence not only the political sphere but also economic interactions (e.g., Bhagwat et al. 2020; Peterson and Godby 2020). For example, customers’ perceptions of brands change considerably when they become aware of the brand’s political orientation (Hydock, Paharia, and Blair 2020; Panagopoulos et al. 2020). Moreover, a significant majority of both Democrats and Republicans now indicate that knowing a brand’s political orientation is important before deciding to buy (or reject) its products/services (Piacenza 2018). Under these conditions, the uptick in interest among marketers in how political identity informs customers’ attitudes and behaviors is understandable. Efforts to elucidate the range of effects of customers’ political identity have 1 grown across the scholarly sub-fields of marketing, including consumer behavior and psychology (e.g., Jost 2017a; Jung et al. 2017a), retailing (e.g., Jung and Mittal 2020), and strategy (e.g., Bhagwat et al. 2020; Fernandes et al. 2022; Hydock, Paharia, and Blair 2020). A review of the expanding body of literature on political identity in marketing published in leading journals since 2010 (Table A1, Appendix A), as well as related review studies (Jost 2017a; Jung and Mittal 2020), reveal a noteworthy characteristic of this literature that informs our work. That is, the studies in this literature have focused almost exclusively on uncovering “ideological asymmetries” that differentiate “liberals and conservatives” in their attitudes and behaviors as consumers. Such a perspective corresponds to the paradigm dominant within political psychology over the last two decades that also focuses on the psychological differences between those on the left versus the right of the political spectrum (Jost et al. 2003; Jost 2017b). As such, underpinning this stream of research in marketing is an assumption that differences in individuals’ political orientations have natural implications for their attitudes and behaviors in the consumption domain (Jost 2017a). While this approach has led to important insights for marketers, especially from the perspective of customer segmentation (Table A1, Appendix A), it covers only one dimension of political polarization – i.e., ideologically-driven asymmetries differentiating those on the left from those on the right – and hence provides an incomplete picture of the implications of political polarization for customers and marketers (see Figure 1, a representative structure of political identity research in marketing). Recent work in political psychology suggests that beyond political orientation (i.e., left vs. right), another critical dimension of political polarization – the intensification of political 2 identities, or political intensity1 – has significant implications in shaping individuals’ decisions concerning issues and events across various aspects of life (Brandt, Evans, and Crawford 2015). This approach focuses on how individuals closer to the extremes on both the left and right ends of the political spectrum are similar to each other and different from those who are politically moderate or neutral, and how such differences inform their emotions, cognition, and behaviors (van Prooijen and Krouwel 2019). For example, compared to moderates, individuals who are more politically intense (on both ends of the political spectrum) are more likely to overestimate their understanding of issues (Fernbach et al. 2013), more likely to believe in conspiracy theories (van Prooijen, Krouwel, and Pollet 2015), and more likely to exhibit stronger convictions in their beliefs (Zwicker, van Prooijen, and Krouwel 2020). Nevertheless, this alternative perspective focused on the commonalities of the more politically intense remains largely unexplored in the marketing literature, and thus little is known about how political intensity drives individuals’ emotions, cognition, and behaviors as consumers. This currently overlooked perspective guides our research, addressing questions that have the potential to meaningfully inform both academic inquiries into the role of political identity in marketing as well as strategic managerial decision- making towards more effectively navigating an increasingly polarized business landscape, with a special focus on customer segmentation strategies. Our work complements and advances political identity research in marketing by developing and testing a framework outlining the role of political intensity in driving customers’ 1Throughout this paper, we utilize the term “political intensity” to refer to the extent to which citizens are polarized into, and (more) strongly identify with, generic left or right-wing ideological outlooks on society (van Prooijen, Krouwel, and Pollet 2015; van Prooijen and Krouwel 2019). In the political psychology literature, this concept is also referred to as “political extremism.” We choose to adopt the alternate term "political intensity," as it more accurately reflects our research focus – the degree or strength of political identity (versus the term “extremism” that tends to more narrowly convey a focus on only the endpoints of the spectrum). 3 post-purchase decisions concerning the brands from which they buy. Specifically, we examine political intensity as a determinant of customers’ loyalty to and advocacy for their chosen brands. Our selection of these outcomes is deliberately anchored to prior findings that those with more intense political identities are the most fervent loyalists to political parties and often “go the extra mile” in their support in the form of likelihood to vote, monetary donations, voluntary service, and promotional support through active party evangelism (Hudy, Mason, and Aaroe 2015). These outcomes are of particular relevance in the consumption domain, as brands similarly seek segments of customers likely to remain loyal purchasers of and advocates for their offerings due to the financial-performance-related advantages of such customers (e.g., cash flow maintenance and growth, profitability) (Villanueva, Yoo, and Hassens 2008; Watson et al. 2015). Furthermore, we propose and test the mechanisms underlying the effects of political intensity on customer loyalty and advocacy. We draw on the extant research on political intensity in political psychology and argue that the unique psychological traits shared by individuals with more intense political identities (on both sides of the political spectrum), as opposed to those with weaker or neutral political identities (van Prooijen and Krouwel 2019), drive their beliefs about the superiority of their consumption choices and their subsequent level of post-purchase involvement with those choices. This, in turn, drives their inclination to remain loyal to and advocate for their choices i.e., product/service or the brand they choose to purchase from. Next, we add further specificity to our understanding of the effects of political intensity on these critical business outcomes by examining boundary conditions – exogenous factors (i.e., those beyond the control or predictive capacity of the individual consumer) that present significant threats to the continuity of the customer-brand relationship by lowering customers’ perceived switching costs. We propose that the effects of political intensity on customer loyalty 4 and brand advocacy are stronger in the presence of market-based factors, i.e., external threats originating from the actions or inactions of the focal brand’s close competitors (product launches and price promotions) and internal threats originating from the actions or inactions of the focal brand itself (product/service failure). We argue that politically intense customers’ beliefs about the superiority of their consumption choices provide a cushion against these threats to customer- brand relationships. This resilience stems from the motivated reasoning that these customers employ when faced with such challenges. On the other hand, we propose that the effects of political intensity on customer loyalty and brand advocacy are weaker in the presence of a nonmarket threat of customer-brand political misalignment (where the customer and the brand hold opposing political orientations). This attenuation in effect occurs because such a mismatch creates inconsistency between the foundational political values and consumption choices of politically intense customers. We test this theoretical framework empirically across nine interlinked and complementary studies, including a large-scale secondary data study and eight experimental studies. Employing diverse estimation strategies and proxies, operationalizations, and experimental manipulations of key variables, we find consistent support for our proposed relationships. Collectively, our research offers significant contributions to both academic inquiry and managerial practice. We expand the scope of political identity research in marketing by introducing a previously overlooked dimension of political polarization: political intensity. We identify the distinct effects of political intensity – over and above those of political orientation – on customer loyalty and brand advocacy, delineate the mechanisms underlying these effects, and show that such positive effects are stronger in market-based contexts that typically threaten the continuity of customer-brand relationships, but weaker in a nonmarket context that exhibits 5 incompatibility between politically intense consumers’ foundational political values and consumption decisions. By integrating this novel perspective into marketing scholarship, we hope to stimulate further research examining the implications of political intensity on outcomes previously linked to political orientation (Figure 1), as well as identifying unexplored outcomes that are driven uniquely by political intensity rather than political orientation. Additionally, from a customer segmentation perspective, we offer valuable insights for managers seeking to cultivate a more loyal customer base that actively advocates on the brand’s behalf. First, our findings underscore the importance for brands to gain a comprehensive understanding of the effects of political polarization on their customer base, one focused not only on the differences rooted in political orientation but also on the commonalities of those closer to the ends of the political spectrum. We highlight the strategic potential of more politically intense customers (on both the left and the right) as a target segment that is likely to remain more loyal to and advocate on behalf of their chosen brands – a cohort that has grown and appears likely to continue to grow (Boxell, Gentzkow, and Shapiro 2022). Second, this segment of politically intense customers may be particularly valuable as a “cushion” for brands when they encounter market-based external (competitor actions; product launches and promotions) and internal (product/service failures) threats. Such findings have significant relevance for marketers, especially considering prior research suggesting that brands’ investments (time, effort, and resources) aimed at maintaining customer loyalty in the face of such threats are not always effective (Pick and Eisend 2014). Conversely, our findings concerning the attenuating role of customer-brand political misalignment indicate that brands should carefully consider their political positions and target these customers with marketing strategies that are apolitical (c.f., Hydock, Paharia, and Blair 2020). 6 THEORETICAL FRAMEWORK AND HYPOTHESES Main Effects and Mediation: Political Intensity, Belief Superiority of Consumption Choice, Post-Purchase Involvement, and Loyalty and Advocacy In what follows, we draw on recent research in political psychology, along with relevant marketing literature, to propose that politically intense individuals often view their consumption choices as superior. Such beliefs motivate them to become more involved with the product/service or brand from which they choose to purchase. Once deeply involved, they demonstrate a propensity to stay loyal to those choices and advocate for them to others. Political Intensity. A dominant theme in the research at the intersection of marketing and political identity is a focus on asymmetries between political identity groups resulting in dissimilarities in attitudes and behaviors differentiating those on the left and those on the right. This literature has been influenced by work in political and social psychology over the past two decades examining the foundations of political conservatism (Jost et al. 2003) and its emphasis on ideological asymmetries differentiating liberals/Democrats from conservatives/Republicans (Jost 2017a; Jost 2017b). This latter research has connected conservatism with a host of interrelated psychological characteristics. Collectively, these characteristics serve as the foundation of the theory of right-wing authoritarianism (Altemeyer 1988) and the “rigidity of the right” hypothesis, both of which have been deployed in recent marketing studies (e.g., Han et al. 2019; Irmak, Murdock, and Kanuri 2020; Jost 2017a). However, both early (e.g., Hoffer 1951) and more recent (e.g., Toner et al. 2013; van Prooijen and Krouwel 2019; Zmigrod, Rentfrow, and Robbins 2019) research in political psychology has observed significant commonalities uniting those with more intense political identities – regardless of the underlying ideological or partisan orientation – while distinguishing them from those with weaker and neutral political identities. Drawing on this research stream, 7 our central assertion in what follows is that individuals who are more politically intense (versus those with less intense or neutral political identities), and regardless of the directionality of their political beliefs (i.e., liberal/Democrat vs. conservative/Republican), share certain traits which drive their beliefs about the superiority of their consumption choices, the subsequent level of post-purchase involvement with those choices, and ultimately, lead them to become more loyal to and advocate for their chosen brands. Belief Superiority of Consumption Choice. Drawing from extant research on the mindset of more politically intense individuals (van Prooijen and Krouwel 2019), we investigate the impact of political intensity on individuals’ behavior as consumers, specifically, on their beliefs about the superiority of their consumption choices. Belief superiority of consumption choice is an individual’s conviction that his or her choices as a consumer, such as the choice of the product/service or the brand from which to purchase, are better or more correct compared to all other available alternatives (c.f. Ruvio et al. 2020). We propose that the psychological traits exhibited by more politically intense individuals (versus those who are less intense or neutral) lead them to believe that they make superior consumption choices.2 The first trait associated with a more politically intense mindset is psychological distress, which corresponds to a sense of meaninglessness and fear stemming from anxious uncertainty. In the political-economic context, more politically intense individuals have been shown to exhibit higher levels of anxiety about the future and to deploy strong convictions about their choices to ameliorate such feelings (McGregor, Prentice, and Nash 2013; van Prooijen, Krouwel, and Pollet 2015). In the consumption domain, given this distress, such politically intense customers may 2These psychological traits are supported by empirical evidence and contribute to a parsimonious understanding of the mindset of the politically intense (van Prooijen and Krouwel 2019). 8 also consider their choices as consumers to be superior in order to compensate for similar feelings (Hall and Raimi 2018). The second trait, cognitive simplicity, concerns adhesion to a simplistic, black-and-white perception of the social world. Driven by this simplistic worldview, more politically intense individuals tend to interpret the world via binary mental categories comprised of opposites – right vs. wrong, superior vs. inferior, and so forth (Lammers et al. 2017; van Prooijen, Krouwel, and Pollet 2015). Within the consumption domain, and given this tendency towards simplistic thinking, we expect that after making a consumption choice politically intense customers may believe their choice to be superior to the alternatives. This perception arises due to the binary nature of their thinking, which offers a level of simplicity that is absent in more intricate and elaborate decision-making processes (Kruglanski et al. 2006). The third trait, overconfidence in judgment, is the tendency to overestimate the accuracy of one’s views and to place higher confidence in one’s ability to predict future outcomes. In the context of socio-political issues, prior work shows that more politically intense individuals overestimate their abilities and knowledge even when they have little or no expertise (Brandt, Evans, and Crawford 2015; Fernbach et al. 2013; Toner et al. 2013). Accordingly, in the consumption domain, politically intense consumers may believe that they have a deeper understanding of the product/service or the brands available in the marketplace and feel that their choice is better or superior to that of others. Finally, the fourth trait, intolerance of opinions, refers to the tendency of individuals to be resistant to considering alternative perspectives. In the political context, this intolerance is reflected in the tendency of more politically intense individuals to reject opposed ideological beliefs that differ from their own and are considered inferior (Brandt et al. 2014). In the consumption domain, politically intense customers may believe that they make superior choices because they perceive their own beliefs and preferences 9 to be more valid than those of others (in this case, the choices of other consumers). Based on the above discussion, we argue that customers who are more politically intense (on both sides of the political spectrum) regard their chosen option as consumers to be superior to alternatives. Next, we propose that this belief in the superiority of consumption choices motivates politically intense customers to become more involved with their chosen option (product/service or brand). Post-Purchase Involvement. Prior research defines involvement as “a person’s perceived relevance of the object based on inherent needs, values, and interests” (Zaichkowsky 1985). Because it is a matter of individual perception, involvement levels can differ among individuals even for the same objects due to underlying differences in their values and beliefs (c.f. Antil 1984; Havitz and Mannell 2005). In the post-purchase context, customers’ involvement with their chosen option manifests behaviorally, via such behaviors as actively searching for additional information and a willingness to learn more about the chosen option (c.f., Pizzutti, Gonçalves, and Ferreira 2022). Furthermore, prior research shows that choices made by individuals shape their preferences such that the act of making a choice itself enhances the value of, and possibly, their involvement with the choice (Voigt, Murawski, and Bode 2017). Such effects have been shown to occur even in the absence of initial preference for the choice and are enduring in nature (Sharot et al. 2010; 2012). They correspond to an epistemic interpretation of self-perception theory, wherein the presence of conflicting information or experience is unnecessary and reflects a choice-driven positive motivation based on the memory of previous choices of the type: “because I recall choosing it, I must have a favorable opinion about it” (Friston et al. 2016). We expect that this choice-induced value-updating process is magnified among politically intense consumers, precisely because their beliefs regarding the superiority of their choices 10 facilitate and reinforce this value-updating process. Driven by the resulting motivation of the kind “I have made the best choice, I will/must make the most out of it,” such consumers are likely to proactively seek out additional information, usage cases, and positive experiences related to their chosen option, i.e., exhibit higher post-purchase involvement to better appreciate and derive maximum utility from their believed superior purchases (Bryant 2021; Pizzutti, Gonçalves, and Ferreira 2022). Further, such higher post-purchase involvement drives them to remain loyal to and to talk favorably about their choices, as we discuss next. Customer Loyalty and Brand Advocacy. At their core, customers’ post-purchase evaluations of product offerings or brands from which they buy are opportunities for them to reflect on the discrete stages of their customer journey (Lemon and Verhoef 2016). While these evaluations can inform a variety of customers’ subsequent decisions, in both research and practice two critical decisions concern whether to continue buying from the same brand (i.e., customer loyalty) (Morgeson et al. 2020) and whether to recommend the brand to others (i.e., brand advocacy) (You, Vadakkepatt, and Joshi 2015). There are at least two strands of research suggesting that consumers’ post-purchase involvement with their chosen brands should lead to a greater tendency to stay loyal to and advocate for brands to others. The first concerns the sunk cost effect, a concept widely studied in behavioral economics and pervasive in decision-making across domains (Arkes and Blumer 1985; Emich and Pyone 2018). It suggests that individuals continue investing in a project or endeavor based on cumulative prior investments, even if the future benefits of doing so do not outweigh the costs. Applied to the post-purchase context, this means that the more customers become involved with their chosen product/service or brand, which in the case of politically intense customers entails investments (i.e., time, money, resources, etc.) made in trying to make the most out of their 11 superior choice, the more likely they are to remain loyal to it.3 This is because doing otherwise (i.e., switching to an alternate brand) would negate previous investments which likely cannot be fully recovered. The second strand of research concerns the stronger attachment bonds that customers develop with their chosen offering or brand as a result of their higher involvement with it after purchase. A high level of post-purchase involvement entails discovering both utilitarian and hedonic benefits from the chosen option, leading to positive cognitive and emotional bonds to the brand (Thomson, MacInnis, and Park 2005). Previous research shows that attachment leads customers to develop a long-term and stable preference for a brand’s offerings, make efforts to promote the brand, and defend the brand in public (Bhattacharya and Sen 2003; Belaid and Behi 2011). Given the above, we hypothesize that: H1: Political intensity is associated with customers’ higher likelihood of remaining loyal to and advocating on behalf of a chosen brand. H2: The association between political intensity and customers’ loyalty to and advocacy for a chosen brand is mediated through their beliefs about the superiority of their consumption choice and subsequent post-purchase involvement with that choice. Moderation by Threats to Customer-Firm Relationships from Market-Related Factors Inarguably, developing a loyal and expanding customer base is a business imperative. Above, we identify political intensity as a customer-level factor that drives individuals to be more loyal and more likely to advocate on behalf of their chosen brand. There are, however, 3For instance, Apple Inc. designs a highly differentiated and integrated array of products and services that require significant investments by consumers to learn to operate them. Such a strategy, capitalizing on the sunk cost effect, has been identified as a key driver of loyalty that Apple enjoys among its customer base (Kerwin 2023). 12 market-based factors exogenous to the individual customer emerging from the actions/inactions of both the competitor brands in the marketplace and the focal brand that threaten the continuity of customer-brand relationships by lowering switching costs (Baron 1995). Switching costs are the perceived and/or experienced costs incurred by a customer when changing from one supplier to another (Burnham, Frels, and Mahajan 2003). Research has identified and differentiated exogenous market-based factors originating from the actions/inactions of the focal brand’s competitors (i.e., competitor-related, external) from the actions/inactions of the focal brand (i.e., brand-related, internal) that lower switching costs for a brand’s customers and elevate the likelihood that they might end the relationship (Pick and Eisend 2014). Relatedly, research has shown that when confronted with challenging information or experiences, more politically intense individuals undertake motivated reasoning (often manifesting as confirmation/disconfirmation biases) to protect their views, decisions, and choices (Hall and Raimi 2018; Morgeson et al. 2022). In a post-purchase context, upon experiencing market-related threats that may challenge customers’ preexisting choices, we expect that politically intense customers will likewise undertake motivated reasoning to prevent experiencing dissonance from such threats. Specifically, to maintain the integrity of their beliefs in the superiority of their consumption choices, politically intense customers will engage in various forms of confirmation biases, resulting in a smaller decline (relative to customers with weaker beliefs in the superiority of their consumption choice) or even an uptick in their likelihood to remain loyal to and advocate for the focal brand (c.f., Pratt et al. 2022). On the other hand, the more rational evaluations of less politically intense and moderate customers with lower levels of belief superiority suggest that their inclination to maintain the relationship with the focal brand should decline more steeply (relative to customers with stronger beliefs in the 13 superiority of their consumption choice) upon experiencing such threats, as they are more open to rationally internalizing information about such threats (Morgeson et al. 2022). Competitor-Related External Threats. In competitive industries, where customers are regularly exposed to information about alternatives, competitors’ actions that increase the availability (e.g., product launches) or attractiveness (e.g., price promotions) of their offerings can strongly affect customers’ perceived switching costs away from their chosen option (Ahluwalia, Burnkrant, and Unnava 2000; Pick and Eisend 2014). While brands invest in a variety of mechanisms aimed at shielding them from the adverse effects of competitive interventions (e.g., branding and advertising) (Frennea, Han, and Mittal 2019), customer churn rates remain high (Ascarza, Netzer, and Hardel 2018). As such, the identification of customer segments less likely to be swayed by competitors’ actions is a critical endeavor (Watson et al. 2015). We propose that politically intense customers are a segment likely to be less sensitive to competitor-related threats to their relationships with their chosen brands. Customers who are more politically intense, given their deeply ingrained beliefs regarding the superiority of their consumption choice, may disregard information from competitors (e.g., new product launches) that could otherwise lower their switching costs by increasing knowledge about available alternatives. Evidence for such selective information processing is available in the political context in the form of “crippled epistemology,” through which politically intense individuals have been found to form closed information echo chambers, solely attending to and accepting information originating from their chosen ingroup while dismissing information from external sources (Hardin 2002; van Prooijen, Krouwel, and Pollet 2015). Additionally, research suggests that politically intense individuals may abstain from actively seeking new information about alternatives (e.g., new products or promotions by 14 competitors) altogether out of a perceived lack of need, given that they consider their existing choices to be well-informed and superior (Hall and Raimi 2018). Furthermore, consumers who are more politically intense may engage in biased processing of competitors’ information, for example, by selectively interpreting a promotional offer from a competitor as an indication of bad quality or “brand desperation” (Gedenk and Neslin 1999). Consequently, this biased interpretation may even strengthen their intention to remain loyal to their chosen brand (c.f., Pratt et al. 2022). Therefore, we hypothesize that: H3: The association between political intensity and customers’ loyalty to and advocacy for a chosen brand is stronger in the presence of competitor-related external threats such as product launches and price promotions. Brand-Related Internal Threats. Originating from the actions or inactions of the focal brand, a critical factor that can lower customers’ perceived switching costs is a negative incident such as a product/service failure (Morgeson et al. 2020). A significant body of research has demonstrated that such failures have the potential to undo the investments made by brands in delivering satisfying experiences to their customers (Pick and Eisend 2014). Consequently, identifying customer segments with a lower inclination to switch after experiencing a product/service failure is crucial (Khamitov, Gregoire, and Suri 2020). We identify politically intense customers to be one segment less likely to be sensitive to such brand-related threats. Similar to the case of competitor-related threats and the deployment of motivated reasoning, politically intense customers may interpret product/service failures selectively to minimize the challenge to their pre-existing perceived choice superiority, attributing the cause of the failure to sources other than the focal brand and becoming more forgiving (c.f., Dawar and Pillutia 200). Evidence of such framing biases exists in the context of politics, where individuals 15 with more intense political views have been observed to become even more certain about their existing beliefs when confronted with contradictory information. This is because they actively seek evidence that aligns with their pre-existing views even from uncongenial sources (Raimi and Leary 2014). On the other hand, individuals with weaker or neutral political identities, who are less prone to motivated reasoning biases, are less likely to be forgiving when experiencing such failures. H4: The association between political intensity and customers’ loyalty to and advocacy for a chosen brand is stronger in the presence of a focal brand-related internal threat of product/service failure. Moderation by Threats to Customer-Firm Relationships from a Nonmarket Factor Customer-Brand Political Misalignment. The hypotheses outlined above are agnostic to customers’ knowledge of the political orientation of the brands from which they choose to purchase. We have posited that even when they possess no knowledge regarding the political leanings of their chosen brands, customers who are more politically intense will develop stronger loyalty and be more likely to advocate for them. This is of considerable importance, as various firms choose to remain apolitical and are successful in doing so (Center for Political Accountability 2022). However, many brands actively employ nonmarket strategies concerning interactions with participants other than those in the marketplace (Baron 1995), such as engaging in socio-political activism and corporate political activities (Bhagwat et al. 2020). Moreover, an increasing number of customers express the significance of being aware of the political orientations of brands when making purchase and re-purchase decisions (Piacenza 2018). Consequently, a misalignment between customers’ political orientations and those of their chosen brands represents a nonmarket factor that has the potential to threaten the political 16 intensity–loyalty/advocacy relationships. When politically intense customers experience dissonance due to market-related threats, they deploy motivated reasoning to dismiss or selectively process information about such threats to preserve their preexisting beliefs about the superiority of their chosen brands. However, when they experience dissonance stemming from a political identity mismatch between themselves and their chosen brands, it is less likely that such information can be easily dismissed, as it threatens their foundational political values (Iyengar and Westwood 2014). In the case of customer-brand political misalignment, continuing to strongly believe that their chosen brand is superior would be inconsistent with politically intense customers’ fundamental political stance, hence insinuating dissonance. As a result, to reduce such dissonance and to maintain the internal consistency of their foundational political values, politically intense customers may be more compelled to modify their previously held superior beliefs regarding the chosen brands (White, Agro, and Sengupta 2012). Conversely, less politically intense or moderate customers may not experience the same degree of dissonance from such a misalignment, given the narrower ideological divide between them and their chosen brands, making it less likely that they perceive the misalignment as detrimental to their relationships with the brands. H5: The association between political intensity and customers’ loyalty to and advocacy for a chosen brand is weaker in the presence of a nonmarket threat of customer-brand political misalignment. 17 Roadmap of Empirical and Experimental Studies METHODOLOGY To test our hypotheses, we adopt a multi-method, multi-study approach and analyze nine unique but complementary samples of data. In Study 1A, we begin with an analysis of a large dataset compiled from multiple independent secondary sources to test the main effect of political intensity on customers’ loyalty and advocacy (H1), as well as the moderating role of the external threat of competitor actions (via product launches) (H3) and the internal threat of product/service failure (H4). While we employ a quasi-experimental instrumental variable estimation to draw causal inferences in Study 1A, in Study 1B we experimentally manipulate the political intensity of customers and replicate its effects on customers’ loyalty to and advocacy for their chosen brands. In Studies 2A, 2B, and 2BW, we validate the main effects of political intensity (H1) and examine the underlying mechanisms that drive these effects, specifically, the belief superiority of consumption choice and post-purchase involvement with the choice (H2). We accomplish this in two separate controlled contexts: the purchase and professional installation of windows (Study 2A) and web-based photo enhancement services (Study 2B), as well as a recall context that utilizes customers’ experiences with actual cell phone brands (Study 2BW). In Study 3, we replicate the main effect of political intensity (H1) and the moderating role of the external threat of competitor actions by exogenously manipulating the presence (versus absence) of a price promotion from a close competitor (H3) in a controlled context involving a non-durable product (soft drinks). In Study 4, we replicate the main effect (H1) and test the moderating role of an internal threat of service failure by exogenously manipulating customer experience in a controlled context involving a service-embedded non-durable product (restaurant dining 18 experience) (H4). In Studies 5 and 5W, we once again substantiate the main effects of political intensity (H1) and test the moderating role of customer-brand political misalignment (H5) by exogenously manipulating it through brands’ socio-political issue-based activism (Study 5) and political party affiliation (Study 5W). To enhance the validity of our findings, across these studies, we employ diverse measures (and manipulations) of key constructs of political intensity, customer loyalty, brand advocacy, and external and internal threats. Moreover, in all studies, we account for the political affiliation (based on party or ideology) of customers to demonstrate that the effects of political intensity are distinct from and over and above that of the direction (i.e., left vs. right) of customers’ political affiliation. Table 1 provides details regarding the research designs and empirical methodologies utilized in these nine studies. Study 1A: Main Effect (H1) and Moderation (H3 and H4) Design and Sample. In Study 1A, we examine the main effect of political intensity on customers’ loyalty and advocacy (H1), as well as the moderating roles of a competitor-related external threat posed by competitors’ product launches (H3) and a focal brand-related internal threat posed by product/service failures (H4). We compile a large secondary dataset from four independent sources. First, data about customers’ political party affiliation, loyalty to, and advocacy for the brands from which they have recently purchased was obtained from the American Customer Satisfaction Index (ACSI). ACSI is a U.S.-based syndicated market research company that specializes in measuring the experiences of customers with the products and services provided by numerous private sector firms spanning nearly 50 economic industries and federal government entities. As per ACSI’s data collection protocol, after completing a private sector questionnaire for a subset of applicable private sector industries and brands, respondents 19 are randomly screened for inclusion in the federal government study. For the latter, ACSI interviews several thousand customers of federal government agencies annually, including survey items measuring political party affiliation as a relevant segmentation variable. From these samples, we obtained data on all respondents who completed both the federal government study and the private sector study for the years 2016 to 2021, making information available about their political party affiliation and loyalty to and advocacy for their chosen brands, as well as a proxy measure of an internal threat (product/service failure). This dataset includes 22,489 observations collected from a Census-demographic-adjusted panel of customers residing in all 50 U.S. states and Washington, D.C. regarding their recent experiences with 131 U.S. brands. These brands belong to four B2C industries, namely, commercial airlines (N=1,472; 6.56%), retail (department/discount stores N=5,461; 24.33% and specialty stores N=8,916; 39.73%); internet service providers (N=5,072; 22.60%); and insurance (life insurance N=538; 2.40%, personal property insurance N=388; 1.73%, and health insurance N=597; 2.66%).4 Second, we obtain data on competitors’ product launches (external threat) in the form of 4ACSI collects data to measure customer experiences across 11 sectors (10 private sectors and the federal government). Per its data collection protocol, the data collection for federal government is combined with that of four economic industries: commercial airlines, consumer retail, internet services providers, and insurance. There are two key benefits of using this data. First, since only a sub-sample of respondents are chosen randomly to answer the federal government study after they respond to a private sector study, we were able to obtain unbiased information about customers’ self-reported political party affiliation (captured in the federal government study) and experiences with actual private brands (captured in the private sector study). Second, the combined dataset comprises information about customers from a large number of brands across four industries over six years. Prior studies either use a small sample (often not more than a few brands) of self-reported customer political affiliations or use a large sample with distant, aggregate proxies of political affiliations (e.g., voting data to approximate ideology at the county/state level). Our study uniquely uses both a large sample of customers’ experiences with actual brands and a precise measure of political affiliation at the customer level. The complete list of brands is available on the ACSI website (www.theacsi.org). While ACSI’s federal government data has been used extensively in public policy research (e.g., Morgeson et al. 2022), our study is the first to use it in the marketing literature. 20 product market fluidity, which represents the intensity of changes in competitors’ products and services relative to those of the focal brand. Developed by Hoberg, Phillips, and Prabhala (2014), product market fluidity is a dynamic measure of product market competition that draws on the business descriptions included in companies’ annual 10-K reports (item 101). Third, to enhance the robustness of our results, we compute an alternate measure of internal threat (product/service failure) to customer-brand relationships via comprehensive news coverage data on product/service-related failures experienced by each brand obtained from the Ravenpack (Edge) News Analytics database. Lastly, we obtained comprehensive data on political advertisements aired on television networks in the U.S. from 2013 to 2020 from the Wesleyan Media Project (Kantar Media/CMAG). Using this data, as described in the robustness checks section below, we develop two instrumental variables and employ a control function estimation to address the potentially endogenous nature of our predictor of political intensity. Next, we elaborate on the measures adopted to operationalize variables, followed by a discussion of the model specification. Variables and Measures Dependent Variables: Customer Loyalty and Brand Advocacy. To operationalize customer loyalty and brand advocacy, we use measures from the ACSI private sector study that have been widely adopted by both academic researchers and firms of various sizes across national markets (Mintz et al. 2019). Customer loyalty is measured as a customer’s stated likelihood to repurchase from the same brand the next time they purchase the same type of product or service (e.g., Larivière et al. 2016). Brand advocacy is measured as a customer’s stated likelihood to recommend the brand to a friend or a colleague (e.g., Umashankar, Ward, and Dahl 2017). Both variables are measured using a 1-10 scale, with higher values representing 21 a higher likelihood to stay loyal to or advocate for the chosen brand. Predictor Variable: Political Intensity. To operationalize political intensity, we adopt a widely used measure developed within the political science literature. This measure is derived from information captured using a set of items about a respondent’s political party affiliation and the intensity of such affiliation (1= “Strong Democrat,” 2= “Not Very Strong Democrat,” 3= “Leaning Democrat,” 4= “Independent,” 5= “Leaning Republican,” 6= “Not Very Strong Republican,” and 7= “Strong Republican).5 Following procedures used in previous research (Brandt, Evans, and Crawford 2015; Fernbach et al. 2013; Zwicker, van Prooijen, and Krouwel 2020), we transform these political party affiliation and affiliation strength ratings into a measure of political intensity by subtracting the midpoint of the scale (4) and taking the absolute value. This transformed measure provides a linear estimate of political intensity. Subsequently, in a relevant sub-section within our robustness checks, we also adopt an alternate curvilinear (quadratic) approach to testing the effects of political intensity. Moderators: Research in relationship marketing shows that lowered switching costs arising from competitor-related (e.g., new products) or focal brand-related (e.g., negative incidents and experiences) factors pose significant threats to the relationship between customers and brands (Pick and Eisend 2014). Therefore, we use product market fluidity, which arises from changes in competitors’ products and services relative to the focal brand, and service failures, which arise from the focal brand’s actions or inactions, as measures of external (H3) and internal threats (H4), respectively. 5We exclude from our sample those respondents who selected ‘other party’ (3.3% of the initial sample) in their responses. This is because data on their political intensity is not available. We also exclude those respondents who did not disclose their political party affiliation (6.0% of the initial sample). 22 External Threat of Competitor Action (via Product Launches). Our objective is to capture competition from the perspective of the customers of a focal brand, as competition would present a threat to the continuity of the focal brand’s customer-brand relationships only if it provides its customers with reasons to switch. Therefore, we use a unique brand-level measure of product market competition – product market fluidity – developed by Hoberg, Phillips, and Prabhala (2014). 6 This dynamic measure, derived from business descriptions reported by companies in their 10-K reports, captures the intensity of changes in competitors’ products/services relative to the focal brand on an annual basis. The higher values of product market fluidity represent higher levels of threats (opportunities to switch) for the focal brand’s customers from its unique and dynamically defined set of close competitors. While recent studies have used this approach to measure dynamic market shares (Bhattacharya, Morgan, and Rego 2022; Morgeson et al. 2023), to the best of our knowledge, this is the first study in marketing to use this approach to capture the intensity of product market threats to brands.7 Internal Threat (via Product/Service Failure). We use instances of a customer complaint, 6We choose this measure as opposed to traditional approaches to measuring market competition (e.g., the Herfindahl-Hirschman Index (HHI)) as those measures only reflect aggregate competition at the industry level but do not accurately capture how competition influences switching costs for the customers of a focal brand. For example, if a higher level of product market competition is driven by the focal brand’s new product launches but not by the competitors’, then there would not be a customer-switching threat to the focal brand from its existing customers. However, if the close competitors of the focal brand launch more products relative to the focal brand, then that would represent a threat to the existing customers of the focal brand. Another shortcoming of the traditional measures of market competition such as HHI is that they typically rely on government-defined industry classification systems, which have the shortcoming of being intertemporally static. The structure of market activity changes much more rapidly and dynamically than is captured in these classification systems, thereby creating a disparity between the actual market structure and the structure assumed by such schema (Bhattacharya, Morgan, and Rego 2022; Morgeson et al. 2023). 7The data on product market fluidity scores can be accessed through the “Hoberg-Philips Data Library” (https://hobergphillips.tuck.dartmouth.edu/). 23 which takes a value of “1” if a customer recently complained to the focal brand and “0” otherwise in ACSI’s private sector study, as a proxy for a product/service failure. This proxy measure is validated through customer descriptions of their complaints, most of which center around a product or service failure situation (see Table B1, Appendix B). Drawing on extant literature from exit-voice-loyalty theory (e.g., Hirschman 1970), we recognize that not all customers who experience a failure with a brand complain (Morgeson et al. 2020). Thus, in one of our robustness checks, we use an alternate measure of service failure obtained from an independent secondary source (Ravenpack Edge News Analytics). Additionally, to further enhance the internal validity of our findings concerning H4, we experimentally manipulate customer experience (service failure vs. control) in Study 4. Control variables. Drawing on research on customer-brand relationships (e.g., Morgeson et al. 2020; Watson et al. 2015) and political identity in marketing (e.g., Jung and Mittal 2020; Fernandes et al. 2022), we include a diverse set of relevant controls in our analysis. These include demographic segmentation factors of customer age, gender, education, income, and race. We also control for alternative fundamental drivers of customers’ relationships with brands, including customer satisfaction and price perceptions of the goods consumed, as both are strong predictors of customers’ repurchase decisions (Watson et al. 2015). Furthermore, to control for the direction of political party affiliation (i.e., “left” vs. “right”), we include the midpoint- centered (i.e., independent-centered) measure of political party affiliation as a covariate in our analyses (see Brandt, Evans, and Crawford 2015).8 Doing so is critical for estimating the unique 8In auxiliary analyses, we control for the direction of political affiliation in two alternate ways (Zwicker, van Prooijen, and Krouwel 2020), 1) by including two dummy variables to indicate customers as having either Democratic affiliation or Republican affiliation, with independents serving as the comparison group, and 2) by including an affiliation variable coded -1 (Democratic), +1 (Republican), and 0 (independents). We conduct these auxiliary analyses for 24 effect of political intensity beyond that of the direction of political affiliation (Democrat vs. Republican). Table B2 (Appendix B) provides details on the inclusion rationale and operationalization of all control variables used in Study 1A. We deploy similar and additional appropriate control variables in subsequent experimental studies as well. Next, we include a variety of fixed effects to account for potential biases arising from omitted variables. Our dataset spans a period of six years (2016-2021). This period includes a variety of time-varying factors, such as three changes in the political party in power in the United States and the COVID-19 pandemic crisis, as well as other unobserved time-varying events of potential significance. Such factors undeniably affected the political discourse in the U.S. as well as brands’ ability to invest in customer experiences (Bloom, Fetcher, and Yeh 2021), thus potentially affecting customers’ post-purchase loyalty and advocacy behaviors towards brands. As such, we include year and month-fixed effects to account for these potential biases and any other idiosyncratic time-varying factors of importance omitted from our models.9 Moreover, a large stream of research on customer-brand relationships (e.g., Watson et al. 2015) and a growing stream of research on political identity in marketing (Table A1, Appendix A) has documented that brands’ strategies, especially in the political domain, interact with customers’ political dispositions and affect the latter’s relationships with brands. As such, we include brand fixed effects to account for unobserved differences in brand strategies and characteristics (including brands’ actual political affiliation) omitted from our models. Finally, we also incorporate industry-fixed effects to account for industry-level unobservable factors in our Study 1A and all the subsequent studies (2A, 2B, 2BW, 3, 4, 5, and 5W) and find results similar to those from our main models, continuing to lend support to our hypotheses. 9In an auxiliary analysis, we alternatively include more granular time-fixed effects (year, month, and week) to account for more frequently occurring time-varying idiosyncratic factors omitted from our models and find similar results. 25 model.10 Model Specification. As outlined above, our dataset includes cross-sectional observations at the customer level spanning multiple years, brands, and industries. As such, we estimate regression models of customers’ loyalty to and advocacy for their chosen brands as a function of their political intensity with relevant control variables and fixed effects discussed above. The model specification is as follows: Yi = αi + β1 Political Intensityi + β2 Political Intensityi × External Threati + β3 Political Intensityi × Internal Threati + β4 External Threati + β5 Internal Threati + β ∑ Controlsi + β ∑ Education Level Fixed Effectsi + β ∑ Income Level Fixed Effectsi + β ∑ Industry Fixed Effectsd + β ∑ Brand Fixed Effectsi + β ∑ Year Fixed Effectsi + β ∑ Month Fixed Effectsi + εi (1) where i denotes the customer, and the outcome variable Yi is either customer loyalty or advocacy. Our parameters of interest are β1, which is the estimate of the effect of political intensity to test H1, and β2 and β3, which are the estimates of the moderating effects of the external threat of product launches by competitors (measured as product market fluidity) and the internal threat of service failure (proxied by customer complaint) to test H3 and H4, respectively. Controlsi is the vector of control variables, which includes customer age, gender, race, political affiliation, satisfaction, and price perceptions. The measures of education and income are categorical and thus are included as fixed effects. Across all models, we use heteroskedasticity 10Our dataset includes customers residing in all 50 U.S. states and Washington, D.C. Evidence suggests that geographic differences explain significant variance in both individuals’ evaluations of firms as customers (e.g., Mittal, Kamakura, and Govind 2004) and their political dispositions (Scala and Johnson 2017). Hence, in auxiliary models, we also incorporate state-of-residence fixed effects to account for such unobserved geographic differences in our model and find similar results. However, since prior research in marketing has utilized geographic units (e.g., county or state) as distant proxies of the political affiliations or ideologies of residents, we do not incorporate them in our main models (e.g., Jung et al. 2017). 26 robust standard errors.11 Preliminary Evidence. We start with preliminary evidence of the effect of political intensity on customer loyalty (advocacy) while accounting for customers’ political orientation. To do so, we regress loyalty (advocacy) on political affiliation and its quadratic term to simultaneously show how the level of customer loyalty (advocacy) varies by both political affiliation and political intensity. Figures C1 and C2 (Appendix C) show the non-linear effects for Study 1A (and studies 2A, 2B, 2BW, 3, 4, 5, and 5W) and show that more politically intense customers on both sides of the political spectrum exhibit higher loyalty to and advocacy for their chosen brands compared to weaker partisans and independents. Estimation Results. We report the descriptive statistics and correlations of all variables used in the models in Table D1 (Appendix D). The results from our specification (equation 1) are summarized in Table 2, where columns 1-2 and 3-4 show results for customer loyalty and advocacy as outcome variables, respectively. To begin, we note that several of the control variables (for example, customer satisfaction (.276, p < .01; Table 2, column 1)) significantly affect our dependent variables (for example, customer loyalty) and in the expected direction. We also note that several of the fixed effects are statistically significant across the models (not reported for the sake of parsimony). Taken together, these results indicate the validity of our model specification. Next, in support of H1, we find that political intensity has a positive effect on customer loyalty (.105, p < .01; Table 2, model 1) and advocacy (.112, p < .01; Table 2, model 3). This means, on average, that a consumer’s level of political intensity is positively associated with their loyalty and advocacy for their chosen brands. 11We note that the results are robust to clustering of standard errors by brands, industries, or time (year-month) (Abadie et al. 2017). 27 Moreover, the interaction models show strong support for our moderation hypotheses. We observe that the main effects of political intensity are stronger at higher levels of product market fluidity (external threat) (on loyalty (.014, p < .01, Table 2, model 2); on advocacy (.013, p < .01, Table 2, model 4)), providing support for H3. The results provide evidence in support of H4 as well, such that the main effects of political intensity are stronger in the presence of an internal threat (product/service failure) (on loyalty (.174, p < .01, Table 2, model 2); on advocacy (.170, p < .01, Table 2, model 4)). Robustness Checks (Study 1A). We undertake a variety of additional analyses to assess the robustness of our findings from Study 1A. These include 1) using an alternate measure of internal threat computed from data on comprehensive media coverage of product/service failure- related news of the brands in our dataset (2015-2020) (H4); 2) using an alternate quadratic/curvilinear specification (H1, H3, and H4); and 3) addressing potential endogeneity concerns using a quasi-experimental instrumental variable approach. We leverage the geographic coverage of our dataset (3a) and comprehensive data on political advertisements aired in the U.S. (2015-2020) (3b) to craft arguably relevant and valid instrumental variables for these analyses (H1, H3, and H4). The results from these robustness checks continue to provide consistent support for our hypotheses. We provide a summary of the objectives, strategies adopted, and conclusions drawn within these robustness checks in Table 3, with complete details presented in the respective appendices. Study 1B: Experimental Manipulation of Political Intensity (H1) Design and Procedure. While Study 1A infers the causal effect of political intensity using a quasi-experimental instrumental variable approach (Table 3, #3), Study 1B alternatively does so by experimentally manipulating political intensity in a controlled context regarding web- 28 based photo enhancement services. A total of 149 U.S.-based respondents were recruited from Cloud Research and provided with appropriate compensation. Participants read about and chose from a list of three photo-editing brands and subsequently engaged in a writing task intended to temporarily manipulate their level of political intensity. Manipulation of Political Intensity. We took the manipulation of political orientation from prior research (Han et al. 2019) and modified it to manipulate political intensity. In the manipulation condition, participants were asked to select from two topics (gay/transgender marriage and detention/deportation of illegal immigrants12) and describe why they were in favor of or opposed to the selected issue. Since Democrats and Republicans tend to hold diametrically opposite views on these two issues (Jung and Mittal 2020), describing their views in favor of/opposed to them in detail should lead participants to feel more intensely or more strongly about their position on the political spectrum.13 In the control condition, participants were instead asked to describe the first meal they had during the day. Tables I1, I2, and I3 (Appendix I) provide complete details on the manipulation, brand choices, and the stimuli administered to participants, respectively. Pretest of Manipulation. We first conducted a pretest to assess the effectiveness of our manipulation of political intensity. For the manipulation check, we measured respondents’ political ideology (1=Strongly liberal; 7=Strongly conservative), and similar to Study 1A, we 12In a pre-test, from a longer list of socio-political issues, these two issues were chosen most frequently by respondents as the issues of most importance to them. We use the term “illegal immigrants” to also include “undocumented immigrants.” 13The process of articulating reasons for a stance involves engaging in argumentation, which requires individuals to think critically about their position, anticipate counterarguments, and reinforce their own arguments. This engagement can lead to a deeper conviction in one’s stance, as the effort to persuade others (or even oneself) solidifies the individual’s position and makes their political views more intense (Xu and Wyer 2012). 29 then created a measure of political intensity by subtracting the midpoint of the scale and taking its absolute value (Brandt, Evans, and Crawford 2015; Fernbach et al. 2013; Zwicker, van Prooijen, and Krouwel 2020). The pretest (N= 153) shows that the manipulation effectively increased respondents’ level of political intensity (Mmanipulaion=1.780, SD=.100; Mcontrol=1.417, SD =.113; t=-2.363. p=.019). Measures. Customer loyalty is captured using the same single-item measure used in Study 1A, but instead on a 7-point scale. For advocacy, we adopt an alternate 3-item measure from Kemp et al. (2012) (Table I4, Appendix I). Respondent demographic characteristics (gender, age, education, income, and race) were included in the survey instrument to serve as controls. Similar to Study 1A, we also measure and control for respondents’ satisfaction with their chosen brand (1 = “Highly dissatisfied”, 7 = “Highly satisfied”). Estimation and Results. Political intensity (1=manipulation condition; 0=control condition) is the key independent variable. To test the main effect of political intensity on customers’ loyalty and advocacy (H1), we use a multiple regression model with brand-fixed effects and robust standard errors (results are consistent with clustering of standard errors by brands). We also ask participants about their perceptions of the chosen brand’s political affiliation (“Democratic Party,” “Republican Party,” “other party,” “no affiliation” or “I don’t know”) and include the perceived political affiliation as fixed effects. Detailed results from this analysis are summarized in Table I5 (Appendix I). Validating results from Study 1A and in support of H1, we find that political intensity has a strong positive effect on customers’ loyalty towards (.665, p < .01) and advocacy for (.636, p < .01) their chosen firms.14 14The effect size reflects the mean difference in outcome variables between the manipulation condition and the control condition after partialling out the effects of control variables and fixed effects. We note that the difference in raw means between the two conditions is also statistically 30 Study 2: Mechanisms (H2) The following series of studies – 2A, 2B, and 2BW – are aimed primarily at testing the mechanisms proposed under H2 in a variety of controlled and recall contexts of products and services. Study 2A: Mediation (Belief Superiority of Consumption Choice) Design and Procedure. In Study 2A, we replicate the main effects of political intensity on customer loyalty and advocacy (H1) and test the mediating mechanism of believed superiority of consumption choices (H2) using a service-embedded non-durable product in a controlled context experiment, minimizing the effects of extraneous factors. A total of 303 U.S.-based respondents were recruited from Prolific and provided appropriate compensation for their participation. Similar to our approach in Study 1A, participants were asked to read about and choose from five local home window sales and installation businesses (see Table J1, Appendix J for descriptions). Since participants had no prior knowledge about the context used in the study, providing options is critical to mimic real-world decision-making contexts and allows respondents to compare and choose from among the options freely. We similarly provide options in each of our eight primary data studies. Next, participants were asked to read an experiential scenario (Table J2, Appendix J) about their experience with window purchase and installation from the chosen brand. Participants then responded to questions that captured their evaluation of the experience described in the scenario, measures of belief superiority of their consumption choices, political affiliation, and other relevant covariates. Measures. Similar to Study 1A, we construct our political intensity measure by transforming the raw ratings of political party affiliation by subtracting the midpoint of the scale significant based on a t-test. 31 (4) and taking the absolute value. We also account for the direction of political affiliation by including the mid-point-centered (4) party affiliation as a covariate (Brandt, Evans, and Crawford 2015). Belief superiority of consumption choices is measured as the customer-level unweighted mean of responses to four items measured on a 7-point scale (α = .77), a measure adapted from Ruvio et al. (2020) (see Table J3, Appendix J). We use the same measures of customer loyalty and advocacy and the same control variables from Study 1B. Table J4 (Appendix J) provides correlations among all the variables included in Study 2A along with descriptive statistics. Estimation and Results. Similar to Study 1A, to test H1 regarding the main effects of political intensity on customers’ loyalty and advocacy, we use a multiple regression model with brand fixed effects, perceived brand political affiliation fixed effects and robust standard errors (results are consistent with clustering of standard errors by brands), along with relevant control variables (age, gender, education, income, race, satisfaction, and political affiliation). Results are summarized in Panel A of Table J5 (Appendix J). Replicating and in support of H1, we find that political intensity has a strong positive effect on customer loyalty (.106, p < .05) and advocacy (.138, p < .01). To test the mediating role of belief superiority of consumption choice, we run additional models (Hayes 2009, Model 4, with 5,000 bootstrap samples) with political intensity as the key independent variable, including all controls and fixed effects mentioned above, and with robust standard errors. Results from these models are summarized in Panel B of Table J5 (Appendix J). We find that customers’ belief in the superiority of their consumption choices mediates the positive effect of political intensity on customer loyalty (.038, 95% CI: .010 to .072) and advocacy (.072, 95% CI: .023 to .126). 32 Study 2B: Serial Mediation (Belief Superiority of Consumption Choice and Post-Purchase Involvement) Design and Procedure. In this study, we replicate the main effect (H1) and test the mediating mechanisms of customers’ belief in the superiority of their consumption choices and their subsequent or post-purchase involvement with the choice (H2). We do so in a controlled context of a durable product (kitchen blenders). Similar to the procedure adopted in Study 2A, 195 participants (recruited from Cloud Research) read about and chose from three blender options (see Table K1, Appendix K for option descriptions). Next, they read about their experience with the blender (Table K2, Appendix K) and answered a series of questions intended to capture the constructs of interest in this study. Measures. Using the same procedure adopted in Study 1A and 2A, we capture political party affiliation and transform it into a measure of political intensity. Customer loyalty, advocacy, and belief superiority of consumption choice (α=.93) and other covariates (satisfaction, age, gender, race, income, and education) are measured using the same approaches as in Study 2A. Customers’ post-purchase involvement with the chosen product (α=.77) is captured using a multi-item measure adapted from Zaichkowsky (1985) and Behe et al. (2015) (see Table K3, Appendix K). Table K4 (Appendix K) provides correlations among the variables and their descriptive statistics. Estimation and Results. To validate H1, we use the same set of covariates and fixed effects and apply the same estimation procedure used in Study 2A and once again validate H1, evidencing that political intensity has a strong positive effect on loyalty (.261, p < .01) and advocacy (.159, p < .01). Results from this analysis are summarized in Panel A of Table K5 (Appendix K). To test H2 regarding the mediating roles of customers’ belief superiority of consumption choice and their subsequent post-purchase involvement with the choice, we run 33 additional models (Hayes 2009, Model 6, with 5,000 bootstrap samples) with political intensity as the key independent variable and all controls, including brand fixed effects and robust standard errors. In support of H2, the results (Panel B of Table K5, Appendix K) show that belief superiority of consumption choice and post-purchase involvement mediate the positive effect of customers’ political intensity on their loyalty (.024, 95% CI: .006 to .054) and advocacy (.032, 95% CI: .009 to .069). Complementary Study 2BW. While study 2B tests H1 and H2 in a controlled context, driven by recommendations for enhancing realism in research (Morales, Amir, and Lee 2017), we conduct a complementary study (2BW) by utilizing customers’ actual experiences with real cell phone brands. In this study, we adopt a similar approach (covariates, fixed effects, and estimation procedure, as in Study 2B) and in addition, take into consideration customers’ previous experiences with the brands by controlling their overall attitude towards the chosen brand (in addition to their satisfaction with the cell phone) and their perceptions of the brand’s political affiliation. Findings from this study continue to provide evidence in support of H1 and H2. We provide a complete description of this study and the results in Appendix L. Study 3: Moderation by External Threat (via Competitor Action of Price Promotion) (H3) Design and Procedure. In this study, we again replicate the main effect (H1) and test the moderating role of an external threat via competitor price promotion (H3) using a controlled context of a non-durable product (soft drinks). We experimentally manipulate the presence of competitor price promotion using scenarios of real brands of soft drinks – Pepsi and Coca-Cola (Coke) – that are the market share leaders in this industry. Both brands offer very similar products, and previous studies indicate that consumers of either brand comparably evaluate their brand experiences, rendering them suitable for investing loyalty for one in the presence of a price 34 promotion threat undertaken by the other, i.e., testing the heterogeneity of the main effect of political intensity under the threat of competitor actions (Pratt et al. 2022). Manipulation of Competitor Action of Price Promotion. We adopt a manipulation of the external threat (competitor price promotion: present, absent/control) from Pratt et al. (2022), which identifies price promotion by a competitor as a threat to the customer-firm relationship through the lowering of switching costs. Participants were recruited from Amazon Mechanical Turk and selected based on being a customer who primarily drinks either Coke or Pepsi. Participants were then asked to imagine a shopping experience at a convenience store during which they felt like purchasing a soft drink. They were then randomly assigned to one of two conditions: a shelf containing bottles of regular Coke and Pepsi, with the non-primary brand’s (i.e., the brand the participant did not choose as their primary brand) price reduced from $2.09 to $0.99 (with the other brand at $2.09) due to a promotion (competitor price promotion condition), or a shelf containing bottles of Coke and Pepsi both priced at $2.09 (control condition). Participants then completed items on loyalty, advocacy, political ideology, and other relevant covariates. See Figure M1 (Appendix M) for details of the scenarios used in manipulation and control conditions. Measures. In this study, we use an alternate multi-item measure of loyalty adapted from Pratt et al. (2022) (Table M1, Appendix M) and use the same measures for all other key variables (advocacy and political intensity) and covariates (satisfaction, gender, education, income, race, and age) as in Study 2B. Similar to Study 2BW, and given that this is a study of customers’ experiences with real brands with which they have experience and a preference, we also control for customers’ overall attitude towards the brand and their perceptions of the chosen brand’s political affiliation and account for them in our empirical analysis. Table M2 (Appendix M) 35 provides correlations among all the variables included in this study along with their descriptive statistics. Estimation and Results. We include data from only those participants (N = 279) who correctly answered the attention check question (i.e., “was the price of Coke and Pepsi the same in the scenario that you just read?”). Similar to Study 2B, we run regression models of customer loyalty and advocacy as a function of customers’ political intensity and all other covariates. Consistent with findings from the previous studies and in support of H1, we find that political intensity has a positive effect on loyalty (.399, p < .01) and advocacy (.365, p < .01). Further, to test the moderating role of competitor action, we add an interaction term between political intensity (mean-centered) and competitor action (1=competitor’s price promotion present, 0=control). In support of H3, we find that the effect of political intensity is stronger in the presence of external threat via competitor’s price promotion for both loyalty (.355, p < .05) and advocacy (.384, p < .05). These results are summarized in Table M3 (Appendix M). Study 4: Moderation by Internal Threat (via Service Failure) (H4) Design and Procedure. Study 4 replicates the main effect (H1) and tests the moderating role of internal threat via a brand’s service failure (H4) using a controlled context of a service- embedded non-durable product (restaurant dining experience). We recruited 603 respondents from Amazon Mechanical Turk and experimentally manipulated their experiences (service failure versus control) using scenarios. Manipulation of Customer Experience. Participants were asked to choose from one of five restaurant options to celebrate a special occasion with a significant other (see Table N1, Appendix N for restaurant descriptions). After making the choice, participants were randomly assigned to one of two conditions (adapted from Ruvio et al. 2020) where they read about a 36 hypothetical lunch experience at the chosen restaurant (a service failure scenario, where the experience was described as very displeasing; and a control scenario, where the experience was described as average and typical) (Table N2, Appendix N). Following this manipulation, participants completed evaluations of their experience (loyalty and advocacy) and items about their views on political issues and other covariates. Measures. With the exception of Study 1B, where we experimentally manipulated political intensity, in all other studies thus far we have utilized direct measures of customers’ political party affiliation or political ideology to compute measures of political intensity. In this study, we alternatively adopt an indirect measure of political ideology based on customers’ views on various socio-political issues (Kidwell, Farmer, and Hardesty 2013; Nail et al. 2009).15 We then calculate the political intensity index by centering this ideology measure at the mid-point of the scale (4) and taking the absolute value (Frimer et al. 2019; Zwicker, van Prooijen, and Krouwel 2020). We employ identical measures for our dependent variables and control variables (satisfaction, gender, education, income, race, and age) as utilized in Study 1B. We also include brand/restaurant-fixed effects and fixed effects for respondents’ perceptions of the chosen restaurant’s political affiliation (Piacenza 2018). Table N3 (Appendix N) provides correlations among all variables included in Study 4, along with descriptive statistics. Estimation and Results. For the manipulation check, participants were asked whether they agreed (or disagreed) that they had a bad experience at the restaurant (1= “strongly disagree”, 7 = “strongly agree”). As intended, in the internal threat (service failure) condition (M 15This measure assesses customers’ stances on capital punishment (reverse coded), abortion (pro- life), gun control, socialized healthcare, same-sex marriage, illegal immigration, and Democrats (α = .70). These seven items were measured on a 7-point scale: 1 = “strongly against,” 7 = “strongly favor,” where lower numbers represent a liberal ideology, and higher numbers represent a conservative ideology and were combined into an ideology measure. 37 = 5.63), participants evaluated their experience significantly more negatively than those in the control condition (M = 4.99, t-test of difference from the mean = 5.59, p < .01). Adopting the same estimation procedure as in previous studies, in support of H1, the results (Table N4, Appendix N) indicate that political intensity has a strong positive effect on loyalty (.357, p < .01) and on advocacy (.272, p < .01). To test the moderating role of internal threat via service failure (H4), we add an interaction term between political intensity (mean-centered) and service failure (1=service failure condition, 0=control condition). Results from the interaction models (Table N4, Appendix N) show a positive moderating effect of the internal threat of service failure on loyalty (.293, p < .01) and on advocacy (.238, p < .05), validating results from Study 1A and in support of H4. Study 5: Moderation by Customer-Brand Political Misalignment (H5) Design and Procedure. This study provides another substantiation of the main effect (H1) and tests the moderating role of customer-firm political misalignment (H5) in a controlled context of a service experience (a budget hotel). A total of 299 participants recruited from Cloud Research were presented with information about and given the option to choose from a list of three budget hotels for a weekend trip to a major city (see Table O1, Appendix O, for details). After reading an experiential description of their stay at the hotel (Table O2, Appendix O), participants were randomly allocated to one of three scenarios, where they read an email from the hotel manager after their stay: 1) a liberal activism scenario, wherein the hotel was portrayed as endorsing gun control; 2) a conservative activism scenario, wherein the hotel was portrayed as endorsing gun freedom; and 3) a control scenario, wherein no information about the hotel’s activism was presented. (See Panel A of Table O3, Appendix O for complete information about these scenarios). Participants then answered a sequence of questions about other constructs 38 of interest to this study. Manipulation and Measures. While our objective in this study is to test the moderating role of customer-firm political misalignment (H5), our study set-up allows us to test the moderating role of alignment as well. Participants were divided into three groups – misaligned, aligned, and control – based on the extent of misalignment or alignment between their own revealed political ideology and the activism scenario presented randomly to them. Detailed information regarding this matching process and the manipulation check is provided in Panel B of Table O3, Appendix O. We utilize the same measures to capture political intensity and other relevant factors such as satisfaction, gender, age, race, income, and education as employed in Study 3, and adopt the same measures of customer loyalty and advocacy as utilized in Study 4. Table O4, Appendix O provides details of the correlations among the variables and descriptive statistics. Estimation and results. Adopting the same approach (covariates, fixed effects, and estimation procedure) as in previous studies, and once again in support of H1, the results (Table O5, Appendix O) indicate that political intensity has a strong positive effect on loyalty (.291, p < .01) and on advocacy (.436, p < .01). To test H5, we introduce an interaction(s) term between the political intensity variable (mean-centered) and a misalignment dummy variable (as well as an alignment dummy variable). The results from models with these interactions (Table O5, Appendix O) provide evidence in support of H5. Specifically, results indicate that the positive effect of political intensity is significantly weaker in the case of customer-firm political misalignment for both loyalty (-.375, p < .05) and advocacy (-.454, p < .05). However, interactions with customer-firm political alignment do not reach statistical significance (p >.10), suggesting that alignment does not alter the strength of these relationships. 39 Complementary Study 5W. Firms primarily engage in political activities in two ways: socio-political activism and corporate political activity (Bhagwat et al. 2020). While Study 5 focuses on the former, this complementary study employs the latter to manipulate the political orientation of firms to test H5 – customer-firm political misalignment. In line with our continued effort to enhance the realism of our research, we alternatively utilize customers’ actual experiences with real fast-food brands (versus hypothetical brands in Study 5). Additionally, since our goal is to experimentally manipulate the political affiliation of brands, we pre-screen respondents and select only those who possess no prior knowledge about the political affiliation of their chosen brand to participate in this study. To test the hypothesis, we adopt a similar methodology as used in Study 5 in terms of variables and model specification and take into account the customers’ previous experiences by controlling their overall attitude towards the chosen brand (in addition to their satisfaction with their experience). We find support for both H1 and H5 and no significant interaction on the main effect from customer-brand alignment, replicating the results from Study 5. We provide a complete description of this study and the results in Appendix P. 40 DISCUSSION AND IMPLICATIONS In an age marked by increasing political polarization, the heightened attention among marketers to the relationship between individuals’ political identity and their attitudes and behaviors as consumers is not only understandable, but justified and essential (Boxell, Gentzkow, and Shapiro 2022). Prior research in marketing has predominantly adopted an asymmetric perspective on political orientation focused on key differences between those on the left versus the right of the political spectrum and the resulting consequences for brands. However, notwithstanding the important insights that have resulted from this work on political orientation, this body of research provides an incomplete picture of the implications of political polarization for marketers (see Figure 1). This is because it overlooks the other dimension of political polarization – the strengthening of political identities, or political intensity – focused on how those nearer to the ends of the political spectrum are similar to each other and different from those with weaker or moderate political identities, and how such differences inform their emotions, cognition, and behaviors as customers (van Prooijen and Krouwel 2019). In this paper, adopting a symmetric perspective on political identity, we demonstrate that relative to those with weaker and moderate political identities politically intense customers (on both sides of the political spectrum) are more loyal to and more likely to advocate on behalf of the brands they chose to purchase from (Studies 1, 1A, 2A, 2B, 2BW, 3, 4, 5, and 5W). This relationship is explained by their beliefs about the superiority of their consumption choices and their post-purchase involvement with those choices (Studies 2A, 2B, and 2BW). Moreover, because customer loyalty and brand advocacy are difficult to cultivate and sustain in competitive markets, we show that the positive effects of political intensity are accentuated in the presence of (or at higher levels of) competitor-related external (Studies 1 and 3) and brand-related internal 41 (Studies 1 and 4) contexts characterized by threats to the continuity of customer-brand relationships. Conversely, since brands also engage in nonmarket activities such as socio- political activism and corporate political activities that signal their political values to the public, we demonstrate that the positive effects of political intensity are attenuated in the case of customer-brand political misalignment (Studies 5 and 5W). These findings suggest that it is necessary to consider a complementary perspective on how political identity relates to marketing strategy. Until now, virtually all insights available to marketing managers regarding the political identity of their customers have focused on the differences between “liberals” and “conservatives” in their attitudes and behaviors, in a sense reflecting the divisive ethos of the current era of increased political polarization. On the other hand, few insights exist concerning the commonalities of more politically intense individuals on both sides, whom we demonstrate to be more loyal brand advocates. Given that many companies meticulously analyze customer data to identify potential high-value segments of customers, this new perspective on political identity should prove highly valuable to companies. Collectively, these findings have important managerial and theoretical implications and motivate several avenues for future research, as we discuss next. Implications for Marketing Practitioners Marketers have long sought customers likely to remain committed, loyal purchasers of their brands and to recommend them to other consumers; these customers are often the most profitable (Watson et al. 2015). Prior research in marketing suggests that practitioners can effectively segment and target customers based on their political orientation, for instance, liberals with company-designed products and conservatives with user-designed products (Paharia and Swaminathan 2019), and conservatives for offerings with higher financial risk and liberals for 42 products in other high-risk domains such as recreation and health (Choma et al. 2014). In keeping with this stream of research and its focus on segmentation, the results from our work help point marketing managers to a newly identified segment of customers more likely to be loyal to and to recommend brands to others – politically intense customers on both sides of the political spectrum. Notably, marketing managers and research practitioners can apply these findings and locate this more politically intense, potentially more valuable segment of customers. While including survey items measuring political identity (i.e., political affiliation/partisanship) in post- purchase surveys and observing the social media behaviors of customers are options (Jung and Mittal 2017), others exist as well. For example, it has long been noted that as partisanship strength increases, so too does the likelihood that an individual will turn out and vote in elections (e.g., Rau 2022).16 Data for voter turnout (e.g., the “U.S. Elections Project”) and voting results (e.g., Federal Election Commission data or the N.Y. Times’ “Extremely Detailed Maps”) are available at the national, state, county, district, and zip-code levels. Controlling for relevant factors such as the competitiveness of the election, mid-term vs. presidential elections, and the demographics of the local population, estimates of the percentage of strong partisans in a particular area (relative to others), as well as the composition of these partisans (i.e., Democrats vs. Republicans), can be inferred from a combination of election results and turnout data (Miller and Conover 2015). Simply put, areas with higher relative voter turnout among the voting- eligible population also likely contain a larger proportion of politically intense individuals. Furthermore, the relevance of our findings for marketing strategists vis-à-vis the 16According to 2020 American National Elections Study data, for example, strong partisans (both Democrats and Republicans) were 1.21 times more likely to have voted in the 2016 presidential election than weak partisans, and 1.27 times more likely to have done so than non-partisans. 43 moderators tested in this study deserves particular attention. Concerning marketplace contexts, we find that the political intensity-loyalty/advocacy relationships are stronger in the presence of both competitor-related external and brand-related internal threats, suggesting that politically intense customers are a segment that can provide a “cushion” for brands in the face of such product market threats. As brands are regularly affected by product-market competition and product/service failures, identifying customers likely to serve as a reservoir of goodwill and lessen the blow from such threats is inherently valuable. On the contrary, we find that the political intensity-loyalty/advocacy relationships are weaker in a nonmarket context – i.e., when the customer’s and brand’s political orientations are not aligned – but are not stronger when they are aligned. Two factors likely explain the latter. First, the higher baseline loyalty/advocacy expressed by politically intense customers may result in a ceiling effect, precluding further increase in loyalty/advocacy when information about alignment is received. Second, politically intense customers, characterized by a stronger belief in the superiority of their consumption choices, may already assume that they are politically aligned with the brands they purchase. Nevertheless, this is an important cautionary note for brands that engage in non-market activities (such as socio-political activism) which signal brands’ political orientations to customers (Hydock, Paharia, and Blair 2020). Overall, our findings suggest that politically intense customers (on both sides of the political spectrum) are an easily identifiable segment of customers that are intrinsically more inclined to remain loyal to the brand they chose and willingly endorse that brand to others.17 Not 17Related to the above, we investigate whether the effects of political intensity observed in our work vary by political orientation. Adopting the methodological approach utilized in previous research (Zwicker, van Prooijen, and Krouwel 2020), and utilizing data from studies 1A, 2A, 2B, 2BW, 3, 5, and 5W, we estimate regression models of loyalty and advocacy as a function of political intensity, political orientation, and the interaction between the two. On the whole, 44 only that, but they are also more likely to continue exhibiting such behaviors in the face of marketplace threats wherein customers’ loyalty has typically been shown to dwindle (Pick and Eisend 2014). However, it may be prudent for brands to employ non-political marketing strategies to identify, target, and acquire such customers since there seems to be a downside of brands’ engaging in political activities, i.e., resulting customer-brand misalignment diminishes the loyalty/advocacy benefits of having politically intense individuals as customers, while customer-brand political alignment provides no upside to such benefits. Implications for Theory and Future Research By presenting a novel conceptual framework that is grounded in theory and subjecting it to rigorous testing, this work expands the existing body of research on political identity in marketing. Previous studies have primarily focused on examining the implications of the distinctions between left- and right-wing orientations among customers. However, our work goes beyond this by shedding light on how political intensity influences customers’ loyalty and advocacy toward brands. Our theoretical perspective, emphasizing the psychological similarities among politically intense customers on both ends of the political spectrum, offers a fresh and complementary viewpoint for future research at the intersection of marketing and political identity. Intuitively, while the content of politically polarized individuals’ beliefs may vary widely, thus explaining the predominant focus on asymmetries in orientation, it has nevertheless long been observed that “all movements, however different in doctrine and aspiration, draw their early adherents from the same types of humanity; they all appeal to the same types of mind” results from these models weakly suggest (due to the lack of consistent statistical significance of the interaction term) that the effects of political intensity, although strong on both sides of the political spectrum, are slightly more pronounced on the “right” side. Refer to Appendix R for complete details of this analysis. 45 (Hoffer 1951). Our results suggest that this shared zealotry of politically intense individuals transcends their doctrinal beliefs and drives behaviors of significance to brands and should thus inspire greater theoretical efforts to identify mechanisms that unite politically intense individuals in their behavior as consumers. We next share some theoretically driven extensions of our work that can guide future research, followed by an empirical limitation suggesting future research avenues. First, the findings from our work show that, in post-purchase contexts, politically intense customers exhibit a greater inclination to remain loyal to and advocate for their chosen brands. As such, politically intense customers are likely to be more challenging to attract away from rival brands, and thus the identification of marketing strategies effective at acquiring such customers is a promising area for future research. Second, one of the proposed reasons why individuals initially become politically intense corresponds to significant-quest theory, which says that the quest for significance (i.e., the need to feel important, valued, and respected) drives individuals toward the political extremes (Kruglanski et al. 2014; van Prooijen and Krouwel 2019). This motivation of politically intense individuals could complement a recent finding in the marketing literature that conservatives have a stronger preference for luxury products because of their greater need for status maintenance (Kim, Park, and Dubois 2018). Specifically, political intensity in general likely has a positive association with a preference for luxury products, which cater to the need to feel important, but this association may simply be stronger among customers with a more conservative political orientation. Next, since most brands in competitive markets face the challenge of retaining customers in contexts that threaten the continuity of customer-brand relationships, we chose to examine the moderating roles of market (competitor-related; brand-related) and nonmarket (customer-brand 46 political misalignment) threats as moderators. However, there could be additional heterogeneity in the effects of political intensity on customer loyalty and advocacy arising from other contextual factors. For instance, while we situated our studies in a variety of product categories and consumption contexts such as CPG goods, restaurant services, and durable good purchases, the effects of political intensity on customer loyalty and brand advocacy may vary in more complex product settings characterized by a lower frequency of purchase and higher risk. As such, we conducted a sensitivity study (Study 6W, Appendix S) to investigate whether the effects of political intensity differ based on the complexity of product categories. While we did not find the effects to vary in highly complex product settings (i.e., the repurchase of an automobile), this may be because our research is focused on examining customers’ post-purchase evaluations of brands that they have already chosen to establish relationships with (hence, the perceived risk is lower). This pattern might not carry over to the context of new customer acquisition, where a consumer’s pre-purchase evaluation can vary significantly by category complexity or level of risk. Future research can delve deeper into this question. Finally, while our work is situated within the U.S. political ecosystem and utilizes samples of U.S. customer data, we recommend research among customers in a more diverse set of national markets, as it is possible that our findings may not apply equally elsewhere (Brader and Tucker 2012). For example, the U.S. is a two-party democracy, while many other democracies are multi-party systems with a more diverse assortment of politically intense consumers. Do politically intense individuals from smaller parties with marginalized ideological platforms exhibit similar beliefs about the superiority of choices and stronger loyalty/advocacy found in centralized two-party systems? 47 Figure 1: Representative Structure of Political Identity Research in Marketing (2011-2024)1 FIGURES 1 Includes relevant articles published from 2011 and 2024 (through March 31, 2024) in the Journal of Marketing, Journal of Marketing Research, Marketing Science, Journal of the Academy of Marketing Science, Marketing Letters, Journal of Consumer Research, Journal of Consumer Psychology, Journal of Service Research, International Journal of Research in Marketing, Journal of Public Policy and Marketing, Journal of Retailing, and Journal of Services Marketing. See Table A1 (Appendix A) for additional details on articles included in this figure. *Examines how liberals and conservatives react differently compared to moderates when watching political speech or debate but does not examine political intensity and its implications for customer- or marketing-related outcomes. 48 Table 1: Summary of Research Designs and Empirical Approaches Adopted in Studies TABLES Study 1A Main Effect# Mechanisms H2 H1 Political Intensity (based on party affiliation, single item) -- Moderators Market-Based Threats H3 Competitor New Product Launches H4 Product/ Service Failure 1B 2A 2B 2BW 3 Political Intensity (manipulated) -- Political Intensity (based on party affiliation, single item) Belief Superiority of Consumption Choice Belief Superiority of Consumption Choice and Post- Purchase Choice Involvement Belief Superiority of Consumption Choice and Post- Purchase Choice Involvement -- Political Intensity (based on political ideology, single item) Political Intensity (based on political ideology, single item) Political Intensity (based on political ideology, single item) -- -- -- -- Competitor Price Promotion (manipulated) -- -- -- -- -- Nonmarket Threat H5 -- -- -- -- -- -- Data Source(s) Context Type Actual experiences with real brands Secondary: ACSI, Ravenpack News Analytics, Hoberg & Philips Data Library, Wesleyan Media Project – Neilson Primary: Prolific Controlled Primary: Cloud Research Primary: Cloud Research Primary: Amazon Mechanical Turk context (hypothetical brands) Controlled context (hypothetical brands) Controlled context (hypothetical brands) Actual experiences with real brands Primary: Amazon Mechanical Turk Controlled context (real brands) Offering Services (airlines, internet service providers, insurance providers, retail stores) Online photo enhancement service Service embedded durable product – window installation Durable product - blenders Durable product – cell phones Non-durable product – soft drinks– restaurants Estimation* OLS, Quasi- Experimental Instrumental Variable (IV) Approach using two unique IVs OLS OLS, SPSS Process Macro Model 4 OLS, SPSS Process Macro Model 6 OLS, SPSS Process Macro Model 6 OLS 49 Table 1 (cont’d) 4 5 5W -- -- -- Political Intensity (based on political ideology, multi- item) Political Intensity (based on political ideology, single item) Political Intensity (based on party affiliation, single item) -- -- -- -- Service Failure (manipulated) -- -- Customer-Firm Political Misalignment (manipulated via brand’s activism) Customer-Firm Political Misalignment (manipulated via the brand’s party affiliation) Primary: Cloud Research Controlled context (hypothetical brands) Service embedded non-durable product OLS Primary: Cloud Research Primary: Cloud Research Controlled context (hypothetical Brands) Actual experiences with real brands Service – budget hotels OLS OLS Service embedded non-durable product – fast food chains Notes: ACSI=American Customer Satisfaction Index. #All studies test the effect of political intensity on both customer loyalty and brand advocacy. Customer loyalty and brand advocacy are measured using a variety of established measures across studies. *All studies include relevant covariates and fixed effects. 50 Table 2: Results for Main Effect (H1) and Moderation (H3 and H4) - Study 1A Customer Loyalty Brand Advocacy Predictor Variables Political Intensity (H1) Political Intensity*External Threat (Product Market Fluidity) (H3) Main Effect Model (1) 0.105*** (0.013) -- Political Intensity*Internal Threat (Product/Service Failure) (H4) -- External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Political Affiliation Customer Satisfaction Price Perception Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) R-squared -0.007 (0.011) -1.007*** (0.045) 0.009 (0.006) 0.276*** (0.013) 0.261*** (0.017) -0.001 (0.001) 0.018 (0.029) -0.241*** (0.036) 7.434*** (0.500) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 0.247 Interaction Model (2) 0.073*** (0.013) 0.014*** (0.004) 0.174*** (0.041) -0.006 (0.011) -1.029*** (0.046) 0.007 (0.006) 0.274*** (0.013) 0.262*** (0.017) -0.001 (0.001) 0.017 (0.029) -0.225*** (0.036) 7.462*** (0.499) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 0.249 Main Effect Model (3) 0.112*** (0.013) -- -- -0.003 (0.011) -1.131*** (0.047) 0.001 (0.006) 0.277*** (0.013) 0.288*** (0.018) -0.003*** (0.001) -0.001 (0.030) -0.244*** (0.037) 7.871*** (0.467) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 0.274 Interaction Model (4) 0.080*** (0.014) 0.013*** (0.004) 0.170*** (0.041) -0.003 (0.011) -1.152*** (0.047) -0.001 (0.006) 0.275*** (0.013) 0.288*** (0.018) -0.003*** (0.001) -0.003 (0.030) -0.229*** (0.037) 7.899*** (0.466) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 0.276 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political intensity and external threat (product market fluidity) variables were mean-centered in the interaction models before creating the interactions (2 and 4). Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. 51 Table 3: Summary of Robustness Checks (Complete Details in Appendices) Robustness Check 1: Alternate measure of internal threat of product/service failure (h4) 2: Alternate – quadratic/curvilinear estimation Objective or Potential Concern In the main models (Study 1A), we use customers’ self-reported complaint behavior to operationalize service failure (a firm-related internal threat). This could introduce bias because outcome variables (loyalty and advocacy) and moderator (complaints) are captured using the same measurement instrument, i.e., ACSI’s private sector study. In study 1A and all subsequent studies, drawing on research from political science literature, we adopt an alternate quadratic curvilinear approach to testing the effect of political intensity. Strategy Adopted and Conclusion We alternatively gather information on product and service failures concerning firms in our data set from the Ravenpack (Edge) News Analytics. This alternate measure of internal threat (via product/service failure) in this analysis is the aggregate confidence-weighted sentiment score of comprehensive news coverage (over three months before the month in which the customer response was captured by the ACSI) of the focal firm’s product and service failures. We follow the same model specification in this analysis as in the main models. As additional controls, we include the aggregate confidence-weighted sentiment scores of all the other positive and negative news concerning the focal firms in our sample for the same period. Complete details of this analysis and results are summarized in Appendix E. We continue to find strong support for H4. We estimate regression models of loyalty and advocacy as a function of customers’ linear party/ideology affiliation variable, which provides an estimate of the effect of political party/ideology affiliation and its quadratic term, which provides an estimate of the effect of political intensity controlling for party affiliation (H1). Further, for Study 1A, we include interactions of internal threat and external threat with both the linear and quadratic terms of party affiliation, with the latter providing estimates for moderating effects of external threat (H3) and internal threat (H4). Complete details of this analysis and results are summarized in Appendix F. We adopt this alternate specification for the remaining studies (1A, 2A, 2B, 2BW, 3, 4, 5, and 5W) as well and present results in Appendix Q. We continue to find support for all tested hypotheses using this alternate estimation. 3 Addressing Potentially Endogenous Nature of Political Intensity In study 1A, while our inclusion of conceptually and empirically relevant covariates and a rich set of fixed effects account for customer, time, firm, and industry-specific potential omitted variable biases (Wooldridge 2010), there are still potential identification concerns arising from additional omitted factors such as household-level norms and values that could drive both an individual’s political intensity as well as their selection of products and services and subsequently their post-purchase relationships with firms in the customer domain. Given these concerns, as a part of our identification strategy, we adopt a control function approach using two sets of appropriate instrumental variables for the potentially endogenous predictor variable of political intensity. 3a: Neighbors’ political intensity as the instrumental variable We first use data from all U.S. counties from which we have at least two customer observations in a given year to construct our instrument variable of neighbors’ aggregate political intensity. We then focus on sub- sets of counties with a larger number of observations to enhance the representativeness of neighbors. For this, we apply additional restrictions limiting our analysis to data from those counties with an increasingly larger number of customer observations (i.e., >=5 and >=10 neighbors). Using an increasingly larger number of neighbors increases the likelihood that they are representative of the political orientation and the intensity of such orientation of the residents in a given county. Complete details of this analysis and results (with and without these restrictions) from the first stage models and the full control function estimation are summarized in Appendix G. We find continued support for our hypotheses. For this first instrumental variable approach, drawing on contagion effects literature in political science, we use the aggregate political intensity of a customer’s “neighbors,” defined as those living in the same county as the focal customer in a given year, as the instrument. 3b: Political advertising-based alternate instrumental variables and estimation strategy The quadratic estimation utilized in robustness check 2 allows us to better leverage the institutional context of our study, i.e., the political ecosystem in the U.S., to craft two alternate instrumental variables for addressing the potential endogeneity concerning our predictor variable of political intensity. For this control function estimation, we adopt the curvilinear (quadratic) model specification and add two alternate instrumental variables of Democratic Party advertising intensity and Republican Party advertising intensity, which we craft using data on more than 3.4 million TV political advertisements aired during the period 2015-2020. Complete details of this analysis, including a detailed discussion on how these instruments satisfy the relevance and validity requirements, and results thereof (first stage and full models) are presented in Appendix H. We continue to find consistent support for our hypotheses (H1, H3, and H4). 52 REFERENCES Abadie, Alberto, Susan Athey, Guido W. Imbens, and Jeffrey Wooldridge (2017), “When Should You Adjust Standard Errors for Clustering?,” National Bureau of Economic Research, w24003. Ahluwalia, Rohini, Robert E. Burnkrant, and H. Rao Unnava (2000), “Consumer Response to Negative Publicity: The Moderating Role of Commitment,” Journal of Marketing Research, 37(2), 203-214. Allard, Tomas and Brent McFerran (2022), “Ethical Branding in A Divided World: How Political Orientation Motivates Reactions to Marketplace Transgressions,” Journal of Consumer Psychology, 32(4), 551-572. Altemeyer, B. (1988), Enemies of Freedom: Understanding Right-Wing Authoritarianism. Jossey-Bass. Angle, Justin W., Sokiente W. Dagogo-Jack, Mark R. Forehand, and Andrew W. Perkins (2017), “Activating Stereotypes with Brand Imagery: The Role of Viewer Political Identity,” Journal of Consumer Psychology, 27(1), 84-90. Arkes, Hal R. and Catherine Blumer (1985), “The Psychology of Sunk Cost,” Organizational Behavior and Human Decision Processes, 35(1), 124-140. Ascarza, Eva, Oded Netzer, and Bruce G. S. Hardie (2018), “Some Customers Would Rather Leave Without Saying Goodbye,” Marketing Science, 37(1), 54-77. Antil, John H. (1984), “Conceptualization and Operationalization of Involvement,” Advances in Consumer Research, 11(1). Baltas, George, Paraskevas Argouslidis, and Dionysis Skarmeas (2010), “The Role of Customer Factors in Multiple Store Patronage: A Cost-Benefit Approach,” Journal of Retailing, 86(1), 37-50. Baron, David P. (1995), “The Nonmarket Strategy System,” MIT Sloan Management Review. Behe, Bridget, Mikyeubg Bae, P. Huddleston, and Lynnell Sage (2015), “The Effect of Involvement on Visual Attention and Product Choice,” Journal of Retailing and Consumer Services, 24, 10-21. Belaid, Samy and Azza Temessek Behi (2011), “The Role of Attachment in Building Consumer‐ brand Relationships: An Empirical Investigation in the Utilitarian Consumption Context,” Journal of Product & Brand Management, 20(1), 37-47. Bennett, W. Lance (1998). “The Uncivic Culture: Communication, Identity, and the Rise of Lifestyle Politics,” PS: Political Science and Politics, 31 (December), 41-61. 53 Bhagwat, Yashoda, Nooshin L. Warren, Joshua T. Beck, and George F. Watson IV (2020), “Corporate Sociopolitical Activism and Firm Value,” Journal of Marketing, 84(5), 1-21. Bhattacharya, C. B. and Sankar Sen (2003), “Consumer–company Identification: A Framework for Understanding Consumers’ Relationships with Companies,” Journal of Marketing, 67(2), 76-88. Bhattacharya, Abhi, Neil A. Morgan, and Lopo L. Rego (2022), “Examining Why and When Market Share Drives Firm Profit,” Journal of Marketing, 86(4), 73-94. Bloom, Nicholas, Robert S. Fletcher, and Ethan Yeh (2021), “The Impact of COVID-19 on US firms,” National Bureau of Economic Research, w28314. Boxell, Levi, Matthew Gentzkow, and Jesse M. Shapiro (2022), “Cross-Country Trends in Affective Polarization,” Review of Economics and Statistics, 1-60. Brader, Ted and Joshua A. Tucker (2012), “Following the Party’s Lead: Party Cues, Policy Opinion, and the Power of Partisanship in Three Multiparty Systems,” Comparative Politics, 44(4), 403-403. Brandt, Mark J., Christine Reyna, John R. Chambers, Jarret T. Crawford, and Geoffrey Wetherell (2014), “The Ideological-Conflict Hypothesis: Intolerance Among Both Liberals and Conservatives,” Current Directions in Psychological Science, 23(1), 27-34. Brandt, Mark J., A.M. Evans, and Jarret T. Crawford (2015). “The Unthinking or Confident Extremist? Political Extremists are More Likely than Moderates to Reject Experimenter- Generated Anchors,” Psychological Science, 26, 189–202. Bryant, Fred B. (2021), “Current Progress and Future Directions for Theory and Research on Savoring,” Frontiers in Psychology, 12, 771698. Burnham, Tomas, Judy Frels, and Vijay Mahajan (2003), “Consumer Switching Costs: A Typology, Antecedents, and Consequences,” Journal of the Academy of marketing Science, 31(2), 109-126. Center for Political Accountability (2022), “CPA-Zicklin Index of Corporate Political Disclosure and Accountability,” Center for Political Accountability, https://bit.ly/3xek1fy Chan, Eugene Y. (2019), “Political Conservatism and Anthropomorphism: An Investigation,” Journal of Consumer Psychology, 30(3), 515-524. Chan, Eugene Y. and Jasmina Ilicic (2019), “Political Ideology and Brand Attachment,” International Journal of Research in Marketing, 36(4), 630-646. Cho, Jaeho (2008), “Political Ads & Citizen Communication,” Communication Research, 35(4). 54 Choma, B. L., Yaniv Hanoch, Gordon Hodson, and Michaela Gummerum (2014), “Risk Propensity Among Liberals and Conservatives: The Effect of Risk Perception, Expected Benefits, and Risk Domain,” Social Psychological and Personality Science, 5(6), 713- 721. Churchill Jr, Gilbert A., and Carol Surprenant (1982), “An investigation into the determinants of customer satisfaction,” Journal of Marketing Research, 19(4), 491-504. Coppock, Alexander, Donald P. Green, and Ethan Porter (2022), “Does Digital Advertising Affect Vote Choice? Evidence From a Randomized Field Experiment,” Research & Politics, 9(1), 20531680221076901 Dawar, Niraj and Madan M. Pillutla (2000), “Impact of Product-Harm Crises on Brand Equity: The Moderating Role of Consumer Expectations,” Journal of Marketing Research, 37(2), 215-226. Dodds, William B., Kent B. Monroe, and Dhruv Grewal (1991), “Effects of Price, Brand, and Store Information on Buyers’ Product Evaluations,” Journal of Marketing Research, 28(3), 307-319. Emich, Kyle J. and Jin S. Pyone (2018), “Let It Go: Positive Affect Attenuates Sunk Cost Bias by Enhancing Cognitive Flexibility,” Journal of Consumer Psychology, 28(4), 578-596. Fernandes, Daniel, Nailya Ordabayeva, Kyuhong Han, Jihye Jung, and Vikas Mittal (2022), “How Political Identity Shapes Customer Satisfaction,” Journal of Marketing, 86(6), 116-134. Fernandes, Daniel, and Naomi Mandel (2014), “Political Conservatism and Variety-Seeking,” Journal of Consumer Psychology, 24(1), 79-86. Farmer, Adam, Blair Kidwell, and David M. Hardesty (2020a), “Helping a Few a Lot or Many a Little: Political Ideology and Charitable Giving,” Journal of Consumer Psychology, 30(4), 614-630. Farmer, Adam, Blair Kidwell, and David M. Hardesty (2020b), “The Politics of Choice: Political Ideology and Intolerance of Ambiguity,” Journal of Consumer Psychology, 31(1), 6-21. Fernbach, Philip M., Todd Rogers, Craig R. Fox, and Steven A. Sloman (2013), “Political Extremism is Supported by an Illusion of Understanding,” Psychological Science, 24(6). Fowler, Erika F., Michael M. Franz, and Travis N. Ridout (2021), “Political advertising in the United States,” Routledge. Frennea, Carly, K. Han, and V. Mittal (2019), “Value Appropriation and Firm Shareholder Value: Role of Advertising and Receivables Management,” Journal of Marketing Research, 56(2), 291-309. 55 Frimer, Jeremy, Mark Brandt, Zachary Melton, and Matt Motyl (2019), “Extremists on the Left and Right Use Angry, Negative Language,” Personality and Social Psychology Bulletin, 45(8), 1216-1231. Friston, Karl, Thomas FitzGerald, Francesco Rigoli, Phillip Schwartenbeck, and Giovanni Pezzulo, G. (2016), “Active Inference and Learning,” Neuroscience & Biobehavioral Reviews, 68, 862-879. Gedenk, Karen and Scott A. Neslin (1999), “The Role of Retail Promotion in Determining Future Brand Loyalty: Its Effect on Purchase Event Feedback,” Journal of Retailing, 75(4), 433-459. Gilbert, G. Ronald, and Cleopatra Veloutsou (2006), “A Cross‐Industry Comparison of Customer Satisfaction,” Journal of Services Marketing. Goenka, Shreyans, and Stijn M.J. van Osselaer (2023), “Why Is It Wrong to Sell Your Body? Understanding Liberals’ Versus Conservatives’ Moral Objections to Bodily Markets,” Journal of Marketing, 87(1), 64-80. Goldenberg, Amit, Joseph M. Abruzzo, Zi Huang, Jonas Schöne, David Bailey, Robb Willer, Eran Halperin, and Jmes J. Gross (2023), “Homophily and Acrophily as Drivers of Political Segregation,” Nature Human Behaviour, 7(2), 219-230. Goldstein, Ken and Paul Freedman (2002), “Campaign Advertising and Voter Turnout: New Evidence for a Stimulation Effect,” Journal of Politics, 64(3), 721-740. Haans, Richard F., Constant Pieters, and Zi-Lin He (2016), “Thinking About U: Theorizing and Testing U‐and Inverted U‐shaped Relationships in Strategy Research,” Strategic Management Journal, 37(7), 1177-1195. Hall, Michael P., and Kaitlin T. Raimi (2018), “Is Belief Superiority Justified by Superior Knowledge?” Journal of Experimental Social Psychology, 76, 290-306. Han, Kyuhong, Jihye Jung, Vikas Mittal, Jinyong Daniel Zyung, and Hajo Adam (2019), “Political Identity and Financial Risk Taking: Insights from Social Dominance Orientation,” Journal of Marketing Research, 56(4), 581-601. Hardin, Russel (2002), “The Crippled Epistemology of Extremism,” Political Extremism and Rationality, 3-22. Havitz, Mark E. and Roger C. Mannell (2005), “Enduring Involvement, Situational Involvement, and Flow in Leisure and Non-leisure Activities,” Journal of Leisure Research, 37 (2), 152–77. Hayes, Andrew F. (2009), “Beyond Baron and Kenny: Statistical Mediation Analysis in the New Millennium,” Communication Monographs, 76(4), 408-420. 56 Hirschman, Albert O. (1970), Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Harvard University Press. Hoberg, Gerald, Gordon Phillips, and Nagpurnanand Prabhala (2014), “Product Market Threats, Payouts, and Financial Flexibility,” The Journal of Finance, 69(1), 293-324. Hoffer, Eric. (1951), The True Believer: Thoughts on the Nature of Mass Movements. Harper & Brothers: N.Y. Homburg, Christian and Annette Giering (2001), “Personal Characteristics as Moderators of the Relationship Between Customer Satisfaction and Loyalty—An Empirical Analysis,” Psychology & Marketing, 18(1), 43-66. Homburg, Christian, Nicole Koschate, and Wayne Hoyer (2005), “Do Satisfied Customers Really Pay More? A Study of the Relationship Between Customer Satisfaction and Willingness to Pay,” Journal of Marketing, 69(2), 84-96. Hult, G. Tomas M., Forrest V. Morgeson III, Udit Sharma, and Claes Fornell (2022), “Customer Satisfaction and International Business: A Multidisciplinary Review and Avenues for Research,” Journal of International Business Studies, 53(8), 1695-1733. Hydock, Chris, Neeru Paharia, and Sean Blair (2020). “Should Your Brand Pick a Side? How Market Share Determines the Impact of Corporate Political Advocacy,” Journal of Marketing Research, 57(6) Irmak, Caglar, Mitchel R. Murdock, and Vamsi K. Kanuri (2020), “When Consumption Regulations Backfire: The Role of Political Ideology,” Journal of Marketing Research, 57(5), 966-984. Iyengar, Shanto and Sean J. Westwood (2015), “Fear and Loathing Across Party Lines: New Evidence on Group Polarization,” American Journal of Political Science, 59(3), 690-707. Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. Sulloway (2003), “Political Conservatism as Motivated Social Cognition,” Psychological Bulletin, 129(3), 339. Jost, John T., (2017a), “The Marketplace of Ideology: “Elective Affinities” in Political Psychology and their Implications for Customer Behavior,” Journal of Customer Psychology, 27(4), 502-520. Jost, John T., (2017b), “Ideological Asymmetries and the Essence of Political Psychology,” Political Psychology, 38(2), 167-208. Jost, John T., Melanie Langer, and Vishal Singh (2017), “The Politics of Buying, Boycotting, Complaining, and Disputing: An Extension of the Research Program by Jung, Garbarino, Briley, and Wynhausen,” Journal of Consumer Research, 44(3), 503-510. 57 Jung, Jihye and Vikas Mittal (2020), “Political Identity and the Customer Journey: A Research Review,” Journal of Retailing, 96(1), 55-73. Jung, Jihye, and Vikas Mittal (2021), “Political Identity and Preference for Supplemental Educational Programs,” Journal of Market Research, 58(3), 559-578. Jung, Kiju, Ellen Garbarino, Donnel A Briley, and Jesse Wynhausen (2017a), “Blue and Red Voices: Effects of Political Ideology on Consumers’ Complaining and Disputing Behavior,” Journal of Consumer Research, 44(3), 477-499. Jung, Kiju, Ellen Garbarino, Donnel A Briley, and Jesse Wynhausen (2017b), “Political Ideology and Consumer Research beyond Complaining Behavior: A Response to the Commentaries,” Journal of Consumer Research, 44(3), 511-518. Kaikati, Andrew M., Carlos J. Torelli, Karen Page Winterich, and María A.Rodas (2017), “Conforming Conservatives: How Salient Social Identities Can Increase Donations,” Journal of Consumer Psychology, 27(4), 422-434. Kashmiri, Saim and Vijay Mahajan (2017), “Values That Shape Marketing Decisions: Influence of Chief Executive Officers’ Political Ideologies on Innovation Propensity, Shareholder Value, and Risk,” Journal of Marketing Research, 54(2), 260-278. Kemp, Elyria, Carla Y. Childers, and Kim H. Williams (2012), “Place Branding: Creating Self‐ Brand Connections and Brand Advocacy,” Journal of Product & Brand Management. 21(7), 508-515. Kendall, Chad, Tommaso Nannicini, and Francesco Trebbi (2015), “How Do Voters Respond To Information? Evidence From A Randomized Campaign,” American Economic Review, 105(1), 322-53. Kerwin, William (2023), “Apple’s Ecosystem of Integrated Hardware and Software Is Differentiated and Merits a Wide Moat,” Morning Star (November 30), https://bit.ly/4ayHp62 Khamitov, Mansur, Yany Grégoire, and Anshu Suri (2020), “A Systematic Review of Brand Transgression, Service Failure Recovery, and Product-Harm Crisis: Integration and Guiding Insights,” Journal of the Academy of Marketing Science, 48, 519-542. Kidwell, Blair, Adam Farmer, and David M. Hardesty (2013), “Getting Liberals and Conservatives to Go Green: Political Ideology and Congruent Appeals,” Journal of Consumer Research, 40(2), 350-367. Kim, C., B. Park, and D. Dubois (2018), “How Consumers’ Political Ideology and Status- Maintenance Goals Interact to Shape Their Desire for Luxury Goods,” Journal of Marketing, 82(6), 132-149. 58 Kim, Jeehye Christine, Brian Park, and David Dubois (2018), “How Consumers' Political Ideology and Status-Maintenance Goals Interact to Shape Their Desire for Luxury Goods,” Journal of Marketing, 82(6), 132-149. Kruglanski, A. W., A. Pierro, L. Mannetti, and E. Grada (2006), “Groups as Epistemic Providers: Need for Closure and the Unfolding of Group-Centrism,” Psychological Review, 113(1), 84. Kruglanski, A. W., Michele J. Gelfand, Jocelyn J. Bélanger, Anna Sheveland, Malkanthi Hetiarachchi, and Rohan Gunaratna (2014), “The Psychology of Radicalization and Deradicalization: How Significance Quest Impacts Violent Extremism,” Political Psychology, 35, 69-93. Kwak, Nojin, Ann E. Williams, Xiaoru Wang, and Hoon Lee (2005), “Talking Politics and Engaging Politics: An Examination of the Interactive Relationships Between Structural Features of Political Talk and Discussion Engagement,” Communication Research, 32(1), 87-111. Lammers, Joris, Alex Koch, Paul Conway, and Mark J. Brandt (2017), “The Political Domain Appears Simpler to the Politically Extreme Than to Political Moderates,” Social Psychological and Personality Science, 8(6), 612-622. Larivière B., T. L. Keiningham, L. Aksoy, A. Yalçin, F. Morgeson III, and S. Mithas (2016), “Modeling Heterogeneity in the Satisfaction, Loyalty Intention, and Shareholder Value Linkage: A Cross-Industry Analysis at the Customer and Firm Levels,” Journal of Marketing Research, 53(1), 91-109. Lee, Younghwa, Sukki Yoon, Young Woo Lee, and Marla B. Royne (2018), “How Liberals and Conservatives Respond to Equality-Based and Proportionality-Based Rewards in Charity Advertising,” Journal of Public Policy and Marketing, 37(1), 108-118. Lemon, Katherine N., and Peter C. Verhoef (2016), “Understanding Customer Experience Throughout the Customer Journey,” Journal of Marketing, 80(6), 69-96. Lisjak, Monika and Nailya Ordabayeva (2023), “How Political Ideology Shapes Preferences For Observably Inferior Products,” Journal of Consumer Research, 49(6), 1014-1031. McConnell, C., Y. Margalit, N. Malhotra, and M. Levendusky (2018), “The Economic Consequences of Partisanship in a Polarized Era,” American Journal of Political Science, 62(1), 5-18. McGregor, Ian, Mike Prentice, and Kyle Nash (2013), “Anxious Uncertainty and Reactive Approach Motivation (RAM) for Religious, Idealistic, and Lifestyle Extremes,” Journal of Social Issues, 69(3) Miller, Patrick R., and Pamela Johnston Conover (2015), “Red and Blue States of Mind: Partisan 59 Hostility and Voting in the United States,” Political Research Quarterly, 68(2), 225-239. Mittal, Vikas and Wagner A. Kamakura (2001), “Satisfaction, Repurchase Intent, and Repurchase Behavior: Investigating the Moderating Effect of Customer Characteristics,” Journal of Marketing Research, 38(1), 131-142. Mittal, Vikas, Wagner A. Kamakura, and Rahul Govind (2004), “Geographic Patterns in Customer Service and Satisfaction: An Empirical Investigation,” Journal of Marketing, 68(3) Mintz, Ofer, Imran S. Currim, Jan-Benedict E.M. Steenkamp, and Martijn de Jong (2019), “Managerial Metric Use in Marketing Decisions Across 16 Countries: A Cultural Perspective,” Journal of International Business Studies, 1-27. Mohan, Bhavya, Tobias Schlager, Rohit Deshpandé, and Michael I. Norton (2018), “Consumers Avoid Buying from Firms with Higher CEO-to-Worker Pay Ratios,” Journal of Consumer Psychology, 28(2), 344-352. Morales, Andrea C., On Amir, and Leonard Lee (2017), “Keeping It Real in Experimental Research—Understanding When, Where, and How to Enhance Realism and Measure Consumer Behavior,” Journal of Consumer Research, 44(2), 465-476. Morgeson III, Forrest V., G. Tomas M. Hult, Sunil Mithas, Timothy Keiningham, and Claes Fornell (2020), “Turning Complaining Customers into Loyal Customers: Moderators of the Complaint Handling-Customer Loyalty Relationship,” Journal of Marketing, 84(5), 79-99. Morgeson III, F. V., P. Sharma, U. Sharma, and G. T. M. Hult (2022). “Partisan Bias and Citizen Satisfaction, Confidence, and Trust in the US Federal Government,” Public Management Review, 1-24. Morgeson III, Forrest V., Udit Sharma, Xiaoxu W. Schultz, Anita Pansari, Ayalla Ruvio, and Tomas G. M. Hult (2023), “Weathering the Crash: Do Customer-company Relationships Pay off During Economic Crises?” Journal of the Academy of Marketing Science, 1-23. Oliver, Richard L. (1980), “A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions,” Journal of Marketing Research, 17(4), 460-469. Otto, Ashley S., David M. Szymanski, and Rajan Varadarajan (2020), “Customer Satisfaction and Firm Performance: Insights From Over a Quarter Century of Empirical Research,” Journal of the Academy of Marketing Science, 48(3), 543-564. Ordabayeva, Nailya and Daniel Fernandes (2018), “Better or Different? How Political Ideology Shapes Preferences for Differentiation in the Social Hierarchy,” Journal of Consumer Research, 45(2), 227-250. 60 Panagopoulos, Costas, Donald P. Green, Jonathan Krasno, Michael Schwam-Baird, and Kyle Endres (2020), “Partisan Consumerism: Experimental Tests of Customer Reactions to Corporate Political Activity,” Journal of Politics, 82(3), 996-1007. Paharia, Neeru and Vanitha Swaminathan (2019), “Who Is Wary of User Design? The Role of Power-Distance Beliefs in Preference for User-Designed Products,” Journal of Marketing, 83(3), 91-107. Papies, Dominik, Peter Ebbes, and Harald J. van Heerde (2017), “Addressing endogeneity in marketing models,” In Advanced Methods for Modeling Markets (pp. 581-627). Springer, Cham. Peterson, Mark and Robert W. Godby (2020), “Citizen Participation in Political Markets: Extending Service-Dominant Logic to Public Policy,” Journal of Public Policy & Marketing, 39(4), 412-429. Pew Research Center (2017), “Key Takeaways on Americans’ Growing Partisan Divide over Political Values,” Pew Research Center (October 5), https://pewrsr.ch/3I6aJUt. Pew Research Center (2022). “The Polarization in Today’s Congress Has Roots That Go Back Decades,” Pew Research Center (March 10), https://pewrsr.ch/3A7a3Mm. Piacenza, Joanna (2018), “Customers Want Brands to Get Political, but Companies Need to Tread Carefully,” Morning Consult (February 6), https://bit.ly/3uoKser. Pick, Doreen, and Martin Eisend (2014), “Buyers’ Perceived Switching Costs and Switching: A Meta-Analytic Assessment of Their Antecedents,” Journal of the Academy of Marketing Science, 42, 186-204. Pizzutti, C., R. Gonçalves, and M. Ferreira (2022), “Information Search Behavior at the Post- purchase Stage of the Customer Journey,” Journal of the Academy of Marketing Science, 50(5), 981-1010. Pratt, Alexander B., Stacey Robinson, Clay Voorhees, Joyce (Feng) Wang, and Michael Giebelhausen (2022), “Unintended Effects of Price Promotions: Forgoing Competitors’ Price Promotions Strengthens Incumbent Brand Loyalty,” Journal of the Academy of Marketing Science, 1-22. Raimi, Kaitlin T. and Mark R. Leary (2014), “Belief Superiority in the Environmental Domain: Attitude Extremity and Reactions to Fracking,” Journal of Environmental Psychology, 40, 76-85. Rau, Eli G. (2022), “Partisanship as Cause, Not Consequence, of Participation,” Comparative Political Studies, 55(6), 1021-1058. Rutz, Oliver J., and George F. Watson F. (2019), “Endogeneity and Marketing Strategy 61 Research: An Overview,” Journal of the Academy of Marketing Science, 47(3), 479-498. Ruvio, Ayalla, Richard P. Bagozzi, G. Tomas M. Hult, and Richard Spreng (2020), “Consumer Arrogance and Word-of-mouth,” Journal of the Academy of Marketing Science, 48(6), 1116-1137. Samadi, Mehrdad (2016). Intermarket Competition: Evidence from Short Sales. Scala, Dante J., and Kenneth M. Johnson (2017), “Political Polarization Along the Rural-urban Continuum? The Geography of the Presidential Vote, 2000–2016,” The ANNALS of the American Academy of Political and Social Science, 672(1), 162-184. Schoenmueller, Verena, Oded Netzer, and Florian Stahl (2023), “Frontiers: Polarized America: From Political Polarization to Preference Polarization,” Marketing Science, 42(1), 48-60. Sharot, Tali, Cristina M.Velasquez, and Raymond J. Dolan (2010), “Do Decisions Shape Preference? Evidence From Blind Choice,” Psychological Science, 21(9), 1231-1235. Sharot, Tali, Stephen M. Fleming, Xiaoyu Yu, Raphael Koster, and Raymond J. Dolan (2012), “Is Choice-induced Preference Change Long Lasting? Psychological Science, 23(10) Shepherd, Steven, Hesam Teymouri Athar and Sahel Zaboli (2024), “On the Political Right, The Customer is Always Right: Political Ideology, Entitlement, and Complaining,” Journal of Consumer Psychology, 34(1), 83-91. Spenkuch, Jorg L. and David Toniatti (2018), “Political Advertising and Election Results,” The Quarterly Journal of Economics, 133(4), 1981-2036. Stäbler, Samuel and Marc Fischer (2020), “When Does Corporate Social Irresponsibility Become News? Evidence From More Than 1,000 Brand Transgressions Across Five Countries,” Journal of Marketing, 84(3), 46-67. Staiger, Douglas O. and James H. Stock (1994), IV Regression with Weak Instruments. Sussman, Abigail B. and Christopher Y. Olivola (2011), “Axe the Tax: Taxes are Disliked More than Equivalent Costs,” Journal of Marketing Research, 48(SPL), S91-S101. Tanner, J., & Raymond, M. A. (2010), “Low-involvement versus high-involvement buying decisions and the consumer’s decision-making process,” Marketing Principles. Thomson, M., D. MacInnis, and C. Park (2005), “The Ties That Bind: Measuring the Strength of Consumers’ Emotional Attachments to Brands,” Journal of Consumer Psychology, 15(1), 77-91. Toner, Kaitlin, Mark R. Leary, Michael W. Asher, and Katrina P. Jongman-Sereno (2013). “Feeling Superior Is a Bipartisan Issue: Extremity (Not Direction) of Political Views 62 Predicts Perceived Belief Superiority,” Psychological Science, 24(12), 2454–2462. Ulver, Sofia, and Christofer Laurell (2020), “Political Ideology in Consumer Resistance: Analyzing Far-Right Opposition to Multicultural Marketing,” Journal of Public Policy & Marketing, 39(4), 477-493. Umashankar, Nita, Morgan K. Ward, and Darren W. Dahl (2017), “The Benefit of Becoming Friend: Complaining After Service Failures Leads Customers With Strong Ties to Increase Loyalty,” Journal of Marketing, 81(6), 79-98. Varma, Rajita, Raghu Bommaraju, and Siddharth S. Singh (2023), “EXPRESS: Female Chief Marketing Officers: When and Why Their Marketing Decisions Differ From Their Male Counterparts?,” Journal of Marketing Research, 00222437231156902. van Bavel, J. Jay and Andrea Pereira (2018), “The Partisan Brain: An Identity-Based Model of Political Belief,” Trends in Cognitive Sciences, 22(3), 213-224. van Prooijen, Jan-Willem, Andre P. M. Krouwel, and Thomas V. Pollet (2015), “Political Extremism Predicts Belief in Conspiracy Theories,” Social Psychological and Personality Science, 6(5), 570-578. van Prooijen, Jan-Willem and Andre P. M. Krouwel (2019), “Psychological Features of Extreme Political Ideologies,” Current Directions in Psychological Science, 28(2), 159-163 Villanueva, Julian, Shijin Yoo, and Dominiq M. Hanssens (2008), “The Impact of Marketing- Induced Versus Word-of-Mouth Customer Acquisition on Customer Equity Growth,” Journal of Marketing Research, 45(1), 48-59. Voigt, Katherina, Carsten Murawski, and Stefan Bode (2017), “Endogenous Formation of Preferences: Choices Systematically Change Willingness-to-Pay for Goods,” Journal of Experimental Psychology: Learning, Memory, and Cognition, 43(12), 1872. Walsh, Gianfranco, Heiner Evanschitzky, and Maren Wunderlich (2008), “Identification and Analysis of Moderator Variables: Investigating the Customer Satisfaction‐Loyalty Link,” European Journal of Marketing. Wang, Yang, Marco Shaojun Qin, Xueming Luo, and Yu (Eric) Kou (2022), “Frontiers: How support for black lives matter impacts consumer responses on social media,” Marketing Science, 41 (6), 1029–44. Warren, Nooshin L. and Alena Sorescu (2017a), “Interpreting the Stock Returns to New Product Announcements: How the Past Shapes Investors’ Expectations of The Future,” Journal of Marketing Research, 54(5), 799-815. Warren, Nooshin L. and Alena Sorescu (2017b), “When 1+ 1> 2: How Investors React to New Product Releases Announced Concurrently With Other Corporate News,” Journal of 63 Marketing, 81(2), 64-82. Watson, George, Joshua Beck, Conor Henderson, and Robert Palmatier (2015), “Building, Measuring, and Profiting From Customer Loyalty,” Journal of the Academy of Marketing Science, 43, 790-825. White, Katherine, Jennifer J. Argo, and Jaideep Sengupta (2012), “Dissociative Versus Associative Responses to Social Identity Threat: The Role of Consumer Self-construal,” Journal of Consumer Research, 39(4), 704-719. Winterich, Karen Page, Yinlong Zhang, and Vikas Mittal (2012), “How Political Identity and Charity Positioning Increase Donations: Insights from Moral Foundations Theory,” International Journal of Research in Marketing, 29(4), 346-354. Xu, Alison J. and Robert S. Wyer J. (2012), “The Role of Bolstering and Counterarguing Mind- sets in Persuasion,” Journal of Consumer Research, 38(5), 920-932. You, Ya, Gautham G. Vadakkepatt, and Amit M. Joshi (2015), “A Meta-Analysis of Electronic Word-Of-Mouth Elasticity,” Journal of Marketing, 79(2), 19-39. Zaichkowsky, L. Judith (1985), “Measuring the Involvement Construct, “Journal of Consumer Research, 12(3), 341-352. Zmigrod, Leor, Peter Jason Rentfrow, and Trevor W. Robbins (2019), “Cognitive Inflexibility Predicts Extremist Attitudes,” Frontiers in Psychology, 10, 989. Zwicker, Maria V., Jan-Wille van Prooijen, and Andre P. M. Krouwel (2020), “Persistent Beliefs: Political Extremism Predicts Ideological Stability Over Time,” Group Processes & Intergroup Relations, 23(8), 1137-1149. 64 Table A1: A Review of Literature on Political Identity in Marketing (2011-2024)1 APPENDIX A: A REVIEW OF LITERATURE Article Sussman and Olivola (2011) Winterich, Zhang, and Mittal (2012) Form of Polarization Political Orientation Political Orientation Study Focus Tax aversion Donation behavior Xu and Wyer (2012)* Political Intensity Consumer mindset Kidwell, Farmer, and Hardesty (2013) Fernandes and Mandel (2014) Angle, Dagogo-Jack, Forehand, and Perkins (2016) Jost, Langer, and Singh (2017) Jung et al. (2017a) Jung et al. (2017b) Kaikati et al. (2017) Political Orientation Political Orientation Political Orientation Political Orientation Political Orientation Political Orientation Political Orientation Kashmiri and Mahajan (2017) Kim, Park, and Dubois (2018) Political Orientation Political Orientation Lee et al. (2018) Mohan et al. (2018) Political Orientation Political Orientation Sustainable behavior Variety-seeking Stereotype activation Buying and boycotting behavior Complaint and dispute behavior Prosocial behavior Donation behavior Firm innovation propensity Luxury goods consumption Charity advertising Purchase intention Key Findings Members of anti-tax parties (Republicans and Libertarians) show a stronger tendency to avoid tax-related costs than members of pro-tax parties (Democrats, Communists, and Socialists). When moral identity internalization is high, liberals have higher donation intentions to government-managed charities whereas conservatives have higher donation intentions to privately managed charities. Democrats and Republicans when watching same-party speech/debate activate bolstering mindsets whereas when watching opposite-party speech/debate activate counter-arguing mindsets. Independents watching either party’s speech activate counter-arguing mindsets, whereas watching debate activates bolstering mindsets. Liberals have higher recycling intentions when exposed to an individualizing moral appeal; conservatives have higher recycling intentions when exposed to a binding moral appeal. Conservatives are more likely to engage in brand variety-seeking behavior because the positive effect on variety-seeking from their motivation to follow social norms is stronger than the negative effect on variety-seeking from their desire for control. Political ideology moderates the effect of ethnic brand imagery on implicit stereotype activation. When exposed to stereotypical brand imagery, implicit stereotypes are activated only among liberals but not among conservatives. Liberal consumers are more likely than conservative consumers to engage in politically motivated buying and boycotting behaviors. Conservatives are less likely to report complaints or engage in complaint dispute behaviors than liberals. U.S. counties that are more liberal are more likely to engage in pro-social and pro-environmental behaviors. This effect is more pronounced when the counties have higher socioeconomic status. Motivated by their need for social approval, conservatives donate more if they are accounted for by liberals who share a salient identity, but not when the shared identity is not salient or when political identity is activated in the donation context. CEOs’ political liberalism positively impacts firms' innovation propensity (rate of new product development). Compared with liberals, conservatives show a greater desire for luxury goods. This effect occurs with the activation of the status maintenance goal but not with the activation of the status advancement goal or in the absence of a status goal. In both the U.S. and South Korea, liberals are more favorable to equality-based rewards, while conservatives are more favorable to proportionality-based rewards. When high CEO-to-Worker pay ratios are revealed, Democrats and Independents show decreased purchase intention for that firm's products, but the purchase intention of Republicans is not affected. 65 Table A1 (cont’d) Article Ordabayeva and Fernandes (2018) Chan (2019) Form of Polarization Political Orientation Political Orientation Chan and Ilicic (2019) Political Han et al. (2019) Orientation Political Orientation Study Focus Preference for differentiation Product anthropomorphism Brand attachment Financial risk-taking behavior Paharia and Swaminathan (2019) Political Orientation Product preference Farmer, Kidwell, and Hardesty (2020a) Political Orientation Donation behavior Farmer, Kidwell, and Hardesty (2020b) Political Orientation Hedonic and utilitarian consumption Irmak, Murdock, and Kanuri (2020) Political Orientation Reaction to consumption regulations Peterson and Godby (2020) Political Orientation Citizen participation in political markets Ulver and Laurell (2020) Political Orientation Resistance to multicultural marketing Jung and Mittal (2021) Political Orientation Parents’ choice of educational program Allard and McFerran (2022) Political Orientation Attitude toward brand transgression Key Findings Conservative ideology leads consumers to pursue products that signal they are better than others. In contrast, liberal ideology leads consumers to pursue products that signal they are unique from others. Motivated by their need to avoid uncertainty and need for order, conservatives are more likely to anthropomorphize consumer products than liberals. Motivated by their need to avoid uncertainty, conservatives are more likely to have greater attachment bonds to brands and less price sensitivity than liberals. This effect only arises when the brands have a domestic country of origin. Conservatives are greater on social dominance orientation and are more likely to take financial risks because of their focus on the upside potential of a decision. Self-efficacy moderates this relationship in that it positively influences financial risk-taking among conservative consumers but not among liberals. Liberals prefer user-designed products for the sense of empowerment, whereas conservatives prefer company-designed products as they are perceived as having higher quality and value. Liberals (motivated by social justice) tend to donate to a greater number of causes with a smaller amount to each, whereas conservatives (motivated by social order) tend to donate to a smaller number of causes with a greater amount to each. Driven by a higher intolerance of ambiguity, conservatives are more likely to prefer utilitarian options than liberals. This can be reversed if the utilitarian option is framed more ambiguously and the hedonic option is framed less ambiguously. Conservatives tend to view government regulations as freedom restrictions and are more likely than liberals to act reactively towards those regulations to engage more in regulated behaviors. Democrats have higher budget knowledge in electoral markets than Republicans, independents, or non-affiliates. Not-affiliated individuals are the most aversive to tax increases but are also the most positive about Medicaid expansion. Far-right consumers view brands that engage in multicultural marketing efforts as ideological adversaries that prioritize liberal political propaganda, betray the brand’s national roots, and discriminate against far-right consumers. Conservative parents prefer supplemental educational programs with a conformance- oriented pedagogy, whereas liberal parents prefer programs with an independence-oriented pedagogy. When there is marketplace brand transgression, liberals punish ethical brand users less than conventional brand users, whereas conservatives punish ethical brand users more than conventional brand users. 66 Table A1 (cont’d) Fernandes et al. (2022) Wang et al. (2022) Goenka and Osselaer (2023) Political Orientation Political Orientation Political Orientation Lisjak and Ordabayeva (2023) Political Orientation Schoenmueller, Netzer, and Stahl (2023) Shephered, Athar, and Zaboli (2024) Political Orientation Political Orientation Current Study Political Intensity (while accounting for Political Orientation) Customer satisfaction Consumer response to brand activism Attitude on commercial bodily markets Product preference Brand preference Complaint Customer Loyalty and Advocacy Conservatives express stronger customer satisfaction relative to liberals, and this relationship is mediated by conservatives’ stronger belief in free will, which drives stronger trust in their consumption decisions. Compared to liberal consumers, conservative consumers response more negatively on social media to brands supporting the BLM (Black Lives Matter) movement. Both liberals and conservatives object to commercial bodily markets but for different moral reasons. Liberals object due to concerns about body exploitation, whereas conservatives object due to concerns about body sanctity. Conservatives tend to be more interested in observably inferior products than liberals because conservatives are more likely to engage in compensatory reasoning due to their belief that positives can offset negatives. The increased political polarization post-2016 U.S. election is associated with increased polarization in consumer brand preference. This effect is stronger for liberals relative to conservatives due to their need for compensatory consumption. Conservatives exhibit a higher level of entitlement and hence are more likely to complain to firms. Politically intense customers (on either end of the political spectrum) are both more loyal to and more likely to advocate for firms than less intense and politically independent customers, and this relationship is mediated by their stronger belief superiority of consumption choice and higher subsequent post-purchase involvement with the choice. Additionally, the effect of political intensity on loyalty and advocacy is accentuated in the presence of both external (competitor actions such as product launches and price promotions) and internal threats (such as product/service failure), strategic factors that originate from the actions (or inactions) of the focal firm (or its competitors) and that threaten the continuity of customers’ relationships with firms. However, these effects are weaker in the case of an internal firm-related threat of customer-firm political misalignment. Notes. We searched the following journals for articles published between 2011 and 2024 (2011 through March 31, 2024) containing the keywords “political ideology,” “political partisan” (or partisanship), “party affiliation,” “political affiliation,” or “political identity”: Journal of Marketing; Journal of Marketing Research, Marketing Science, Journal of the Academy of Marketing Science, Marketing Letters, Journal of Consumer Research, Journal of Consumer Psychology, Journal of Service Research, International Journal of Research in Marketing, Journal of Public Policy and Marketing, Journal of Retailing, and Journal of Services Marketing. After removing articles that (1) only mention the topic (without making it a central focus of the research), (2) only cite the keywords in the references, and (3) did not contain original research and findings, these 32 papers remain. Reflecting the increased interest in this research stream in marketing, 20 of these 32 papers were published in the last five years (in or after 2018). * Examines how liberals and conservatives react differently compared to moderates when watching political speech or debate but does not examine political intensity and its implications for customer- or marketing-related outcomes. 67 APPENDIX B: DESCRIPTION OF THE CONTROL VARIABLES In the main models for Study 1A, as a measure of product/service failure, we use instances of a customer complaint, which takes a value of “1” if a customer recently complained upon experiencing a product/service failure to the focal brand and “0” otherwise in ACSI’s private sector study. If a customer indicates that they have recently complained, the ACSI study further probed the nature of the complaint via an open-ended question. Our adoption of the instances of complaint as a proxy for product/service failure is validated through these customer verbatims describing the details of their complaint – the majority of which centered around a product or service failure situation. Below are sample verbatim responses from the dataset. 68 Table B1: Verbatim Examples of the Proxy Measure of Internal Threat (Product/Service Failure) Industry Airlines Airlines Airlines Retail Retail Retail Internet Service Internet Service Internet Service Insurance Insurance Insurance Brand Southwest Jet Blue Delta Walmart Kroger Best Buy Verizon Failure Type Service Service Verbatim “Both flights were canceled for mechanical reasons and they offered no help.” “They lost my baggage on a flight that made a stop and didn’t get my baggage till the day after I arrived at my destination.” Product/service “The food was old. The people that help you take their time and you can't get any sleep.” Product Product Service Product/service “I called to report my internet speed as well as a charge on my bill that should no longer be there. Let's just say the “Bought a pair of 25.00 standard tennis shoes and the next day at work the bottom of the right shoe came off.” “The product was opened and tampered with.” “Long waiting time both on the floor and at customer service.” Frontier Product/service “I could not stream anything. Internet kept cutting out.” problems have not been resolved though they were pretty sure it was.” AT&T Service “Request for a technician was not accomplished quickly.” Prudential Financial State Farm Humana Service “My policy lapsed due to the company making an error entering my new payment information.” Product Product/service “Had a problem with renewal, getting my new card, and the cost of the rate increase. Changing my primary care “The premiums continue to increase excessively.” doctor.” 69 Table B2: Description, Rationale for Inclusion, and Literature Source for the Control Variables - Study 1A Variable Age Gender Education Income Race Customer Satisfaction Description [Data Source] Customer’s age in years [Source: ACSI] Male=1, Female=0 [Source: ACSI] 1=Less than high school, 2=High school graduate, 3=Some college or associate degree, 4=College graduate, 5=Post- graduate.” [ACSI] 1= Under $20,000, 2=$20,000 but less than $30,000, 3=$30,000 but less than $40,000, 4=$40,000 but less than $60,000, 5=$60,000 but less than $80,000, 6=$80,000 but less than $100,000, 7=$100,000 or more.” [Source: ACSI] White=1, All others=0 [Source: ACSI] The difference between measures that capture customers’ pre- purchase expectations of the overall quality and post-purchase evaluation of the overall quality of the product/service consumed. [Source: ACSI] Perceived Price Residuals from regressing the overall perceived quality measure on two measures that capture “price given quality” and “quality given price” of the product or service (Morgan and Rego 2009). Higher values indicate increasing positive price perceptions [Source: ACSI] The Rationale for Inclusion and Supporting Literature Heterogeneity in customer characteristics drives the nature of their relationships with firms (Mittal and Kamakura 2001). Customers’ gender, income, education, race, and age affect their relationships with firms (Baltas et al. 2010; Homburg and Giering 2001; Gilbert and Veloutsou 2006; Walsh, Evanschitzky, and Wunderlich 2008). According to the expectations-disconfirmation paradigm (Oliver 1980; Otto, Szymanski, and Varadarajan 2020), satisfaction is a function of consumers’ comparison of outcomes to expectations (e.g., Churchill and Surprenant 1982) and is a strong predictor of both loyalty and advocacy for the firms (Hult et al. 2022), requiring that we control for it in our models. Further, recent work shows that it is associated with the political orientation of customers (Fernandes et al. 2022). Customers’ price perceptions are strong determinants of their post-purchase attitudes and behaviors, including loyalty towards the products and services offered by firms (e.g., Dodds, Monroe, and Grewal 1991, Homburg, Koschate, and Hoyer 2005). For certain customers, sensitivity to perceived price may override the impact of their partisanship strength on customer-firm relationships, suggesting that it be controlled for in our models. 70 Figure C1: Political Intensity and Customer Loyalty – Preliminary Evidence (Quadratic Effects) APPENDIX C: PRELIMINARY EVIDENCE 71 Figure C2: Political Intensity and Brand Advocacy – Preliminary Evidence (Quadratic Effects) 72 APPENDIX D: CORRELATIONS AND DESCRIPTIVE STATISTICS (STUDY 1A) Table D1: Correlations and Descriptive Statistics Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) External Threat (Product Market Fluidity) (5) Internal Threat (Product/Service Failure) (6) Political Affiliation (7) Customer Satisfaction (8) Price Perception (9) Age (10) Gender (11) Education (12) Income (13) Race Mean S. D. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) 1 0.845 0.032 -0.264 -0.142 -0.002 0.326 0.226 -0.001 0.004 0.026 0.08 -0.053 8.110 2.186 1 0.034 -0.283 -0.157 -0.015 0.324 0.231 -0.022 -0.006 0.02 0.079 -0.062 7.959 2.289 1 0.076 0.032 -0.086 -0.001 0.053 0.01 0.03 0.033 -0.017 0.068 2.028 1.092 1 0.159 0.02 -0.238 -0.064 -0.113 0.095 0.084 0.092 0.017 0.182 0.386 1 0.017 -0.102 -0.031 0.017 0.044 -0.029 -0.021 0.025 3.643 3.450 1 -0.013 -0.007 0.096 0.091 -0.043 0.088 0.194 3.746 2.289 1 0.391 0.029 0.004 -0.004 -0.032 0.008 -0.055 1.537 1 0.04 -0.017 0.022 -0.001 0.055 0.028 1.157 1 -0.048 0.011 -0.053 0.184 47.942 15.342 1 0.078 0.166 0.017 0.437 0.496 1 0.436 0.028 3.411 1.018 1 -0.002 3.864 2.080 1 0.817 0.387 Notes: All correlations above 0.02 are statistically significant at p <0.05 or less. N = 22,444. 73 APPENDIX E: ALTERNATE MEASURE OF INTERNAL THREAT In the main models, we use customers’ self-reported complaint behavior to operationalize service failure (a firm-related internal threat). This could introduce bias because our outcome variables (customer loyalty and advocacy) and moderator (complaints) are captured using the same measurement instrument, i.e., ACSI’s private sector study. In this analysis, we alternatively gather data on product and service failures for brands from the Ravenpack (Edge) News Analytics. Ravenpack compiles information concerning a variety of events (e.g., new product releases, product/service failures) for thousands of entities (including companies, people, events, and commodities) from all major news wires and other internet sources and collates this into structured data using text analytics tools. This data has been used in prior research to gather marketing-related information such as new product introductions by brands (Warren and Sorescu 2017a; 2017b, Varma, Bommaraju, and Singh 2023) and inter-market competition (Samadi 2016). Specifically, we obtained Document Sentiment Score (DSS) and Document Sentiment Confidence (DSC) variables of all news items under the “product failure” category for the period of 2015 to 2021 for all the brands included in our sample.18 This category captures news covering failures encountered by brands concerning their products/services. To ensure that the news items can be accurately assigned to the focal firm, we use only those with an entity relevance score of 100 (Warren and Sorescu 2017a; 2017b). Our alternate measure of internal threat (product/service failures) is, therefore, the sum of 18DSS represents “the sentiment for the news document by analyzing the language in each and every sentence. The score is calculated by averaging the sentiment across all sentences in the document.” It ranges from -1.00 to +1.00. DSC is “a measure of the confidence when determining the DSS. This confidence score is measured from the same aggregated sentiment distribution used to compute the DSS.” It ranges from 0.00 to +1.00 with values under 0.20 and above 0.80 representing very low confidence and very high confidence, respectively. 74 the product of DSS and DSC of all “product failure” news items of the focal brand from the quarter previous to the one in which an interview of the focal customer was undertaken by the ACSI. Taking a sum of all the news items covering a given product or service failure allows us to better capture the media coverage and hence potential exposure to information about such failures by customers (Stabler and Fischer 2020). Increasingly negative values of this measure indicate high levels of customer exposure to news about a brand’s service failures. We follow the same model specification (equation 1) and estimation approach for this analysis as in the main model. As additional controls, we include the normalized sum of sentiment scores (DSS*DSC) of all the positive and negative news items (excluding those from the “product failure” category) concerning the focal brands in our sample. The results from this estimation, summarized in Table E1, continue to provide strong support for H4. 75 Table E1: Results for Main Effect (H1) and Moderation (H3 and H4) Using Alternate Measure of Internal Threat (Product/Service Failure) - Study 1A Customer Loyalty Brand Advocacy Predictor Variables Political Intensity (H1) Political Intensity*External Threat (Product Market Fluidity) (H3) Main Effect Model (1) 0.084*** (0.013) -- Political Intensity*Internal Threat (Product/Service Failure) (H4) -- External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Political Affiliation Positive News Negative News Customer Satisfaction Price Perception Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) -0.004 (0.011) -0.001 (0.003) 0.007 (0.006) 0.000 (0.000) 0.000 (0.000) 0.324*** (0.013) 0.258*** (0.018) 0.001 (0.001) -0.018 (0.029) -0.248*** (0.037) 7.210*** (0.517) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Interaction Model (2) 0.081*** (0.013) 0.014*** (0.004) 0.005*** (0.002) -0.004 (0.011) -0.001 (0.003) 0.006 (0.006) 0.000 (0.000) 0.000 (0.000) 0.324*** (0.013) 0.259*** (0.018) 0.001 (0.001) -0.017 (0.029) -0.234*** (0.037) 7.235*** (0.516) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Main Effect Model (3) 0.088*** (0.013) -- -- -0.001 (0.012) -0.003 (0.003) -0.001 (0.006) 0.000 (0.000) 0.000 (0.000) 0.330*** (0.014) 0.285*** (0.018) -0.001 (0.001) -0.041 (0.030) -0.250*** (0.038) 7.626*** (0.490) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Interaction Model (4) 0.086*** (0.013) 0.014*** (0.004) 0.005** (0.002) -0.000 (0.012) -0.003 (0.003) -0.002 (0.006) 0.000 (0.000) 0.000 (0.000) 0.331*** (0.013) 0.286*** (0.018) -0.001 (0.001) -0.040 (0.030) -0.237*** (0.038) 7.651*** (0.490) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political intensity, external threat (product market fluidity and internal threat (product/service failure) variables were mean-centered in the interaction models before creating the interactions (2 and 4)). Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. 76 APPENDIX F: ALTERNATE (CURVILINEAR) ESTIMATION STUDY 1A In the main models, we adopted a widely used linear approach to estimating the effects of political intensity (Brandt, Evans, and Crawford 2015; Fernbach et al. 2013; Zwicker, van Prooijen, and Krouwel 2020). In this analysis, drawing on studies from the political science literature, we adopt an alternate curvilinear (quadratic) approach to examine the effect of political intensity in this analysis (e.g., Toner et al. 2013; Zwicker, van Prooijen, and Krouwel 2020). Specifically, we extend our approach used to provide preliminary evidence and estimate regression models of loyalty and advocacy as a function of customers’ linear party affiliation variable,19 which provides an estimate of the effect of political party affiliation and its quadratic term (mean-centered before creation), which provides an estimate of political intensity controlling for party affiliation. Our modified model specification with this approach is as below: Yi = αi + β1 Party Affiliationi + β3 Party Affiliation2 + β2 Party Affiliation2 × External Threat + β4 Party Affiliation2 × Internal Threat + β5 Party Affiliation × External Threat + β6 Party Affiliation × Internal Threat + β7 External Threat + β8 Internal Threat + β ∑ Controlsi + β ∑ Education Level Fixed Effectsi + β ∑ Income Level Fixed Effectsi + β ∑ Industry Fixed Effectsi + β ∑ Brand Fixed Effectsi + β ∑ Year Fixed Effectsi + β ∑ Month Fixed Effectsi + εi (1w) In this specification, our parameter of interest for the main effect of political intensity is β2 (H1). Further, following the recommended approach for testing interactions on non-linear effects (Haans, Pieters, and He 2016), we include interactions of an external threat (product market fluidity) and internal threat (product/service failure) with both the linear and quadratic 19As noted earlier, having values from 1 through 7, where; 1= “Strong Democrat,” 2= “Not Very Strong Democrat,” 3= “Leaning Democrat,” 4= “Independent,” 5= “Leaning Republican,” 6= “Not Very Strong Republican,” and 7= “Strong Republican.” 77 terms of party affiliation. With this, β3 and β4 provide the estimates for the moderating effects of an external threat (H3) and internal threat (H4), respectively. We continue to include all the control variables and fixed effects as included in our main models and use heteroskedasticity robust standard errors. The results from this estimation, which continue to lend support to the tested hypotheses, are summarized in Table F1. 78 Table F1: Results for Main Effect (H1) and Moderation (H3 and H4) - Study 1A Predictor Variables Political Affiliation Political Affiliation2 (H1) Political Affiliation2 * External Threat (Product Market Fluidity) (H3) Political Affiliation2 * Internal Threat (Product/Service Failure) (H4) Political Affiliation * External Threat (Product Market Fluidity) Political Affiliation * Internal Threat (Product/Service Failure) External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Customer Satisfaction Price Perception Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) Customer Loyalty Brand Advocacy Main Effect Model (1) -0.004 (0.006) 0.034*** (0.004) -- -- -- -- -0.006 (0.011) - 1.011*** (0.045) 0.277*** (0.013) 0.260*** (0.017) -0.001 (0.001) 0.017 (0.029) - 0.238*** (0.036) 7.259*** (0.500) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Interaction Model (2) 0.001 (0.006) 0.026*** (0.004) 0.003** (0.001) 0.045*** (0.012) -0.001 (0.002) -0.032* (0.018) -0.022* (0.013) -1.270*** (0.084) 0.275*** (0.013) 0.261*** (0.017) -0.001 (0.001) 0.015 (0.029) -0.228*** (0.036) 7.319*** (0.499) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Main Effect Model (3) -0.013** (0.006) 0.037*** (0.004) -- -- -- -- -0.003 (0.011) - 1.137*** (0.047) 0.277*** (0.013) 0.287*** (0.018) - 0.003*** (0.001) -0.002 (0.030) - 0.241*** (0.037) 7.681*** (0.466) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Interaction Model (4) -0.010 (0.006) 0.028*** (0.004) 0.003*** (0.001) 0.047*** (0.012) -0.001 (0.002) -0.025 (0.018) -0.020 (0.013) -1.407*** (0.085) 0.275*** (0.013) 0.288*** (0.018) -0.003*** (0.001) -0.004 (0.030) -0.230*** (0.037) 7.746*** (0.465) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political intensity and external threat (product market fluidity) variables were mean-centered in the interaction models before creating the interactions (2 and 4). Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. 79 APPENDIX G: ADDRESSING ENDOGENEITY (NEIGHBOR POLITICAL INTENSITY) For this analysis, we use the aggregate political intensity of a customer’s “neighbors,” defined as those living in the same county as the focal customer in a given year as the instrumental variable. Valid instruments should satisfy both the relevance criterion and the exclusion criterion. First, in terms of relevance, arguably the level of political intensity of residents living in the same county partially determines the intensity of political discourse in that county as reflected in factors such as the intensity of political campaigning (e.g., geo-targeted political campaigning by political parties on online platforms) and the frequency of potentially conflicting discussions around political issues (e.g., climate change, gun control). These factors increase both the salience and the strength of political affiliation for an individual (Cho 2008; Kwak et al. 2005). Importantly, recent research on political segregation finds direct evidence for such effects of politically intense peers or neighbors on focal individuals in the form of “political acrophily,” defined as “the tendency of people to prefer to affiliate with others who have more extreme political views (rather than more moderate)” (Goldenberg et al. 2023, p. 219). Therefore, it is reasonable to argue that an individual’s neighbors’ aggregate political intensity is positively associated with the focal individual’s political intensity (i.e., it satisfies the first condition as an instrument). Second, in terms of the validity criterion, it is unlikely that the neighbors’ aggregate political intensity would correlate with the error term – i.e., it significantly affects the focal individual’s relationships with firms as a customer beyond the effects on the focal individual’s political intensity after accounting for other covariates and fixed effects in our models. For this to occur, the neighbors would need to cooperate and jointly adjust their political intensity in anticipation of factors omitted from our models, such as the focal customer’s household norms 80 and values. This is unlikely to happen because neighbors only rarely (if ever) observe such factors. Taken together, neighbors’ political intensity arguably meets both the relevance and the exclusion criteria for valid instruments. Regarding the diagnostic statistics on the relevance of this instrument, the F-value of the regression model with the customer’s political intensity as the dependent variable and only the instrument as the independent variable is significantly higher than 10. Further, in the first-stage regression with the instrument and all other covariates and fixed effects from equation (1), presented in Table G1, we find the instrument to be statistically significant (p < .01) and the F- test value of its stepwise inclusion greater than 10 (Staiger and Stock 1994). These statistics help us rule out the potential weak instrument problem. Following the standard control function approach (Papies, Ebbes, and van Heerde 2017), we first estimate the first-stage regression model with customers’ political intensity as the dependent variable and our instrumental variable as the key predictor, along with all other control variables and fixed effects from equation (1). We then predict the residuals of this model and incorporate them into our full model as the endogeneity correction term (control function). We follow the recommended control function approach of simultaneously estimating both equations with bootstrapping (500 replications with replacement) to get correct standard errors, as the control function is an estimated quantity (Papies, Ebbes, and van Heerde 2017). Next, we apply additional restrictions to this analysis, limiting our analysis to data from the counties with an increasingly larger number of customer observations (i.e., >=5 and >=10 neighbors). Using an increasingly larger number of neighbors increases the likelihood that they are representative of the political dispositions and intensity of such dispositions of the residents in a given county. Results from the full control function estimation with and without these 81 restrictions are summarized in Tables G2 and G3 for customer loyalty and brand advocacy, respectively. To summarize, our results from this instrumental variable analysis continue to provide strong support for H1, H3, and H4. 82 Table G1: Results for First Stage Models for Neighbors’ Political Intensity Instrumental Variable – Study 1A Predictor Variables Neighbors’ Aggregate Political Intensity External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Customer Satisfaction Price Perception Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year-Month Fixed Effects Month Fixed Effects N (Customers) R-squared Customer Political Intensity Number of Neighbors >=1 (1) 0.548*** (0.012) 0.009 (0.006) 0.171*** (0.020) -0.001 (0.006) 0.055*** (0.007) 0.001 (0.001) 0.011 (0.015) 0.120*** (0.020) 0.476 (0.382) Incl. Incl. Incl. Incl. Incl. Incl. 20,027 0.189 Customer Loyalty Number of Neighbors >=10 (3) 0.413*** (0.032) 0.015 (0.010) 0.202*** (0.030) -0.011 (0.009) 0.080*** (0.011) 0.001 (0.001) 0.050** (0.025) 0.165*** (0.029) 1.487*** (0.546) Incl. Incl. Incl. Incl. Incl. Incl. 9,564 0.215 Number of Neighbors >=5 (2) 0.468*** (0.020) 0.016* (0.008) 0.196*** (0.025) -0.006 (0.007) 0.066*** (0.009) 0.001 (0.001) 0.041** (0.020) 0.126*** (0.024) 1.023** (0.478) Incl. Incl. Incl. Incl. Incl. Incl. 13,687 0.170 Number of Neighbors >=1 (4) 0.549*** (0.011) 0.009 (0.006) 0.171*** (0.020) -0.002 (0.006) 0.056*** (0.007) 0.001 (0.001) 0.010 (0.015) 0.117*** (0.020) 0.457 (0.381) Incl. Incl. Incl. Incl. Incl. Incl. 20,076 0.189 Brand Advocacy Number of Neighbors >=10 (6) 0.415*** (0.032) 0.015 (0.010) 0.206*** (0.030) -0.011 (0.009) 0.081*** (0.011) 0.001 (0.001) 0.048* (0.025) 0.163*** (0.029) 1.487*** (0.549) Incl. Incl. Incl. Incl. Incl. Incl. 9,571 0.214 Number of Neighbors >=5 (5) 0.471*** (0.020) 0.014* (0.008) 0.199*** (0.025) -0.008 (0.007) 0.067*** (0.009) 0.000 (0.001) 0.040** (0.020) 0.124*** (0.024) 0.998** (0.476) Incl. Incl. Incl. Incl. Incl. Incl. 13,707 0.170 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. 83 Table G2: Results for Main Effect (H1) and Moderation (H3 and H4) (Customer Loyalty) - (IV-Neighbors’ Political Intensity) Predictor Variables Political Intensity (H1) Political Intensity* External Threat (Product Market Fluidity) (H3) Political Intensity* Internal Threat (Product/Service Failure) (H4) External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Political Affiliation Customer Satisfaction Price Perception Age Gender Race Control Function Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) Main Effect Model (1) 0.185*** (0.041) -- -- Number of Neighbors>=1 Interaction Model (2) 0.143*** (0.040) 0.014*** (0.004) 0.190*** (0.044) 0.001 (0.013) -1.037*** (0.048) 0.010* (0.006) 0.279*** (0.014) 0.258*** (0.018) -0.003*** (0.001) 0.012 (0.033) -0.227*** (0.036) -0.085** (0.041) 6.932*** (0.883) Incl. Incl. Incl. Incl. Incl. Incl. 20,027 0.000 (0.013) -1.013*** (0.048) 0.011* (0.006) 0.281*** (0.014) 0.256*** (0.018) -0.002** (0.001) 0.013 (0.033) -0.245*** (0.036) -0.093** (0.041) 6.750*** (0.889) Incl. Incl. Incl. Incl. Incl. Incl. 20,027 Main Effect Model (3) 0.385*** (0.088) -- -- Customer Loyalty Number of Neighbors>=5 Interaction Model (4) 0.318*** (0.087) 0.019*** (0.005) 0.244*** (0.050) -0.009 (0.016) -0.956*** (0.059) 0.001 (0.007) 0.271*** (0.018) 0.224*** (0.023) -0.003** (0.001) 0.037 (0.042) -0.216*** (0.045) -0.269*** (0.087) 6.130*** (1.206) Incl. Incl. Incl. Incl. Incl. Incl. 13,687 -0.010 (0.016) -0.918*** (0.058) 0.004 (0.007) 0.274*** (0.018) 0.222*** (0.023) -0.002* (0.001) 0.037 (0.042) -0.245*** (0.045) -0.289*** (0.087) 5.797*** (1.221) Incl. Incl. Incl. Incl. Incl. Incl. 13,687 Main Effect Model (5) 0.644*** (0.155) -- -- Number of Neighbors>=10 Interaction Model (6) 0.540*** (0.155) 0.023*** (0.007) 0.228*** (0.058) -0.012 (0.020) -0.933*** (0.072) 0.003 (0.008) 0.248*** (0.024) 0.243*** (0.031) -0.004** (0.002) 0.017 (0.050) -0.246*** (0.063) -0.495*** (0.157) 6.416*** (1.580) Incl. Incl. Incl. Incl. Incl. Incl. 9,564 -0.014 (0.021) -0.902*** (0.071) 0.005 (0.008) 0.250*** (0.024) 0.237*** (0.031) -0.003** (0.002) 0.013 (0.050) -0.285*** (0.063) -0.553*** (0.155) 5.837*** (1.613) Incl. Incl. Incl. Incl. Incl. Incl. 9,564 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political intensity and external threat variables were mean-centered in the interaction models before creating the interactions (2, 4, and 6). Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. Estimated using bootstrapping (500 replications) 84 Table G3: Results for Main Effect (H1) and Moderation (H3 and H4) (Brand Advocacy) - (IV-Neighbors’ Political Intensity) Brand Advocacy Predictor Variables Political Intensity (H1) Political Intensity* External Threat (Product Market Fluidity) (H3) Political Intensity* Internal Threat (Product/Service Failure) (H4) External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Political Affiliation Customer Satisfaction Price Perception Age Gender Race Control Function Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) Main Effect Model (1) 0.213*** (0.043) -- -- Number of Neighbors>=1 Interaction Model (2) 0.172*** (0.042) 0.013*** (0.004) 0.181*** (0.044) 0.001 (0.013) -1.160*** (0.050) 0.000 (0.006) 0.277*** (0.014) 0.278*** (0.019) -0.004*** (0.001) 0.010 (0.033) -0.205*** (0.038) -0.109*** (0.044) 7.241*** (0.892) Incl. Incl. Incl. Incl. Incl. Incl. 20,076 0.000 (0.013) -1.138*** (0.049) 0.001 (0.006) 0.279*** (0.014) 0.276*** (0.019) -0.004*** (0.001) 0.011 (0.033) -0.222*** (0.038) -0.117*** (0.044) 7.068*** (0.899) Incl. Incl. Incl. Incl. Incl. Incl. 20,076 Main Effect Model (3) 0.508*** (0.088) -- -- Number of Neighbors>=5 Interaction Model (4) 0.447*** (0.087) 0.014*** (0.005) 0.240** (0.050) -0.008 (0.017) -1.096*** (0.062) -0.006 (0.008) 0.267*** (0.017) 0.239*** (0.024) -0.004*** (0.001) 0.050 (0.043) -0.197*** (0.046) -0.390*** (0.087) 6.308*** (1.152) Incl. Incl. Incl. Incl. Incl. Incl. 13,707 -0.010 (0.017) -1.057*** (0.060) -0.003 (0.008) 0.270*** (0.017) 0.238*** (0.024) -0.004*** (0.001) 0.051 (0.043) -0.222*** (0.046) -0.406*** (0.088) 6.015*** (1.168) Incl. Incl. Incl. Incl. Incl. Incl. 13,707 Main Effect Model (5) 0.860*** (0.156) -- -- Number of Neighbors>=10 Interaction Model (6) 0.775*** (0.156) 0.014*** (0.007) 0.224** (0.053) -0.026 (0.073) -1.098*** (0.073) -0.004 (0.008) 0.246*** (0.023) 0.227*** (0.031) -0.006*** (0.002) 0.044 (0.055) -0.255*** (0.061) -0.713*** (0.156) 5.479*** (1.917) Incl. Incl. Incl. Incl. Incl. Incl. 9,571 -0.028 (0.022) -1.062*** (0.072) -0.002 (0.008) 0.249*** (0.023) 0.223*** (0.031) -0.006*** (0.002) 0.043 (0.056) -0.283*** (0.061) -0.755*** (0.155) 5.012*** (1.941) Incl. Incl. Incl. Incl. Incl. Incl. 9,571 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political intensity and external threat variables were mean-centered in the interaction models before creating the interactions (2, 4, and 6). Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. Estimated using bootstrapping (500 replications) 85 APPENDIX H: ADDRESSING ENDOGENEITY (POLITICAL ADVERTISING) For this second control function analysis, we build on the alternate curvilinear or quadratic specification (equation 1w, reproduced below for reference) that we used in robustness check 2: Yi = αi + β1 Party Affiliationi + β3 Party Affiliation2 + β2 Party Affiliation2 × External Threat + β4 Party Affiliation2 × Internal Threat + β5 Party Affiliation × External Threat + β6 Party Affiliation × Internal Threat + β7 External Threat + β8 Internal Threat + β ∑ Controlsi + β ∑ Education Level Fixed Effectsi + β ∑ Income Level Fixed Effectsi + β ∑ Industry Fixed Effectsi + β ∑ Brand Fixed Effectsi + β ∑ Year Fixed Effectsi + β ∑ Month Fixed Effectsi + εi (1w) In this specification, our parameter of interest for the main effect of political intensity is β2 (H1). Further, following the recommended approach for testing interactions on non-linear effects (Haans, Pieters, and He 2016), we include interactions of an external threat (product market fluidity) and internal threat (product/service failure) with both the linear and quadratic terms of party affiliation. With this, β3 and β4 provide the estimates for the moderating effects of an external threat (H3) and internal threat (H4), respectively. We continue to include all the control variables and fixed effects and use heteroskedasticity robust standard errors. This alternate approach allows us to further leverage the institutional context of our study, i.e., the political ecosystem in the U.S., to craft two alternate instrumental variables for party affiliation and hence political intensity (since it is estimated using the quadratic term of party affiliation). Drawing on the literature on the effect of political advertising on voters’ political dispositions during elections (e.g., Fowler, Franz, and Ridout 2021), we construct two instrumental variables based on the cost of political advertising. Specifically, we use the total cost of political advertisements aired during the previous year (from the year in which the data 86 was collected by ACSI from the customer) at the Designated Marketing Area (DMA) level for both the Democratic and Republican party candidates to construct our instrumental variables of Democratic political advertising intensity and Republican political advertising intensity, respectively. To create these instrumental variables, we obtained comprehensive data on more than 3.4 million TV political advertisements aired during the period 2015-2020 from the Wesleyan Media Project (WMP) – Kantar Media/CMAG.20 The relevance of these instrumental variables stems from the well-identified effect of political advertisements on voters’ political dispositions. Evidence from both randomized field experiments (Kendall, Nannicini, and Trebbi 2015) and quasi-experiments (Spenkuch and Toniatti 2018) shows a meaningful change in voters’ political dispositions caused by exposure to political advertisements, especially for those voters who are either independent or lean slightly in favor of a political party. In the first-stage regression (Table H1), we find that both instruments are statistically significant (Democratic political advertising intensity: p < .01; Republican political advertising intensity: p < .01) and the F-test value (21.17, p < .01) for their stepwise inclusion is greater than 10 (Staiger and Stock 1994). In the first-stage regression, the instruments also affect customer political affiliations in the expected directions, such that Democratic political advertising intensity and Republican political advertising intensity have negative and positive signs, respectively. Since we use two instruments, we conduct the Hansen J 20The Wesleyan Media Project tracks all broadcast advertisements aired by or on behalf of federal and state election candidates in every Demographic Market Area (DMA) or media market in the country. In the U.S., almost all the TV advertising is strategized and purchased at the DMA level (Goldstein and Freedman 2002; Spenkuch and Toniati 2018). TV advertising accounts for about 73.4% of the campaign budgets of political parties (Rideout, Fowler, and Franz 2021) and is possibly more effective than digital media advertising (Coppock, Green, and Poter 2022), making it appropriate to construct our instrumental variables. We use the cost of advertising because it closely reflects actual exposure to advertisements. 87 statistic of over-identification. The J statistic value (.74) is statistically non-significant (p = .62), indicating that overidentification is not a concern. Concerning the exclusion restriction, the three core aspects of broadcast political advertisements, namely their timing, volume, and content are directly unrelated to customers’ relationships with private sector firms. First, the timing of political advertisements is driven entirely by the timing of local, county, state, and federal elections which is mandated by law. For example, the years 2017 (a non-election year) and 2018 (a midterm election year) saw 122,067 and 1,209,430 political advertisement television broadcasts in the U.S., respectively. Second, the intensity of political advertising is driven by factors such as campaign budgets, the competitiveness of race, and external endorsement of candidates none of which correlate with customer relationships with private sector firms. Third, the content of political advertisement concerns appeals for votes or other forms of support (e.g., donations) with no specific reference to private sector brands. This is because the Federal Election Commission (FEC) closely regulates and oversees campaign finance and advertising, imposing specific regulations that prevent corporations and other entities from directly contributing money to political campaigns (the Federal Election Campaign Act 197l and amended several times subsequently). When a political advertisement mentions a particular company, it runs the risk of being perceived as an in-kind contribution to a campaign, which is likewise forbidden by these regulations. Taken together, it is unlikely that political advertisements affect customer behaviors and attitudes toward private sector brands directly over and above the effects of the focal customer’s political dispositions and other covariates and fixed effects in our models. Results from these full models, summarized in Table H2, continue to provide support for H1, H3, and H4. 88 Table H1: Results for First Stage Models for IV Estimation for Political Advertising-Based Instrumental Variables – Study 1A Political Affiliation Predictor Variables Democratic Political Advertising Intensity Republican Political Advertising Intensity External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Customer Satisfaction Price Perception Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) Customer Loyalty (1) -0.000*** (0.000) 0.000*** (0.000) 0.015 (0.012) 0.039 (0.045) -0.025** (0.011) 0.003 (0.015) 0.010*** (0.001) 0.267*** (0.032) 1.070*** (0.038) 2.603*** (0.778) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Brand Advocacy (2) -0.000*** (0.000) 0.000*** (0.000) 0.016 (0.012) 0.028 (0.045) -0.023** (0.011) 0.005 (0.015) 0.010*** (0.001) 0.266*** (0.032) 1.074*** (0.038) 2.601*** (0.777) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. 89 Table H2: Results for Main Effect (H1) and Moderation (H3 and H4) [Alternate IV Models] - Study 1A Predictor Variables Political Affiliation Political Affiliation2 (H1) Political Affiliation2 * External Threat (Product Market Fluidity) (H3) Political Affiliation2 * Internal Threat (Product/Service Failure) (H4) Political Affiliation * External Threat (Product Market Fluidity) Political Affiliation * Internal Threat (Product/Service Failure) External Threat (Product Market Fluidity) Internal Threat (Product/Service Failure) Customer Satisfaction Price Perception Age Gender Race Control Function Term Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) Customer Loyalty IV Estimation Brand Advocacy IV Estimation Main Effect Model (1) 0.022 (0.112) 0.034*** (0.004) -- -- -- -- -0.007 (0.012) -1.012*** (0.046) 0.277*** (0.014) 0.260*** (0.017) -0.002 (0.001) 0.010 (0.042) -0.267** (0.128) -0.027 (0.112) 6.895*** (0.878) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Interaction Model (2) 0.052 (0.119) 0.021*** (0.004) 0.005*** (0.001) 0.034*** (0.013) -0.003 (0.002) -0.014 (0.018) -0.033*** (0.013) -1.103*** (0.093) 0.253*** (0.015) 0.292*** (0.019) -0.002 (0.002) 0.003 (0.047) -0.292** (0.131) -0.049 (0.120) 6.962*** (0.529) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 Main Effect Model (3) 0.014 (0.123) 0.035*** (0.004) -- -- -- -- -0.006 (0.012) - 1.014*** (0.049) 0.263*** (0.015) 0.305*** (0.019) - 0.004*** (0.002) 0.004 (0.047) -0.295** (0.138) -0.020 (0.124) 7.798*** (0.542) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Interaction Model (4) 0.038 (0.126) 0.023*** (0.004) 0.005*** (0.001) 0.044*** (0.013) -0.005** (0.002) 0.001 (0.018) -0.029** (0.014) -1.268*** (0.094) 0.262*** (0.015) 0.307*** (0.019) -0.005*** (0.002) -0.001 (0.048) -0.298** (0.139) -0.042 (0.126) 7.913*** (0.547) Incl. Incl. Incl. Incl. Incl. Incl. 22,489 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political intensity and external threat (product market fluidity) variables were mean-centered in the interaction models before creating the interactions (2 and 4). Several education, income, brand, industry, year, and month fixed effects are statistically significant at least at p < .05 across all models. The models reported in this table are estimated using bootstrapping (500 replications) following the standard procedure to correct SE since the Control Function Term is an estimated quantity (Papies, Ebbes, and van Heerde 2017). 90 APPENDIX I: TABLES AND FIGURES FOR STUDY 1B Table I1: Details of Manipulation of Political Intensity Question(s) Answer Choice(s) Manipulation Condition From the list of topics below, please select one that you are passionate about. Topic 1: gay/transgender marriage Topic 2: detention/deportation of illegal immigrants You selected “selected topic” as the topic you are passionate speaking about. Are you in favor of or opposed to the “selected topic”? Option 1: In favor of Option 2: Opposed to Now, please describe in detail why you are “in favor of/opposed to the “selected topic.” Open-ended question Please describe below what you had for breakfast today in a sentence or two. Open-ended question Control Condition Table I2: Online Photo Enhancement Services Options and Their Descriptions Instructions Imagine that you want to edit some photographs for your friends for a professional website (e.g., LinkedIn). Read the descriptions of the three highly rated online photo editing applications to select one. Option 1 Option 2 Option 3 Quick Picture Enhance: Instantly upgrade your photos with QuickPix Enhance – the swift and user- friendly online editing solution. Correct colors, eliminate blemishes, and enhance details in a few clicks, ensuring your images look their best effortlessly. Picture Perfector Pro: Elevate your visuals with PicPerfector Pro, a streamlined online photo editing service. Our professionals apply precision edits, delivering polished results for photographers and individuals alike. Fast, efficient, and tailored to your needs. Snap Sculpt Studio: Sculpt your memories with SnapSculpt Studio's online editing expertise. From basic touch-ups to creative enhancements, our platform offers a simple yet powerful toolkit for transforming your photos into captivating masterpieces. 91 Table I3: Scenario Used Please read the scenario below carefully Next, imagine that you have downloaded the app and used the “selected option” to enhance the pictures for your friends Below are the outputs that the “selected option” produced. Notes. The images used in this study are stock images available for public use. “Pictures” Table I4: Alternate Measure of Brand Advocacy Brand Advocacy (Adopted from Kemp et al. 2012) Items 1. 2. 3. Please indicate the extent to which you agree with the following statements. [1=Strongly Disagree to 7=Strongly Agree] I would recommend to other people that they support this business. I would talk to other people about my experience with it positively. I would suggest to others that they use its services. Table I5: Results for Manipulated Political Intensity (H1) - Study 1B Predictor Variables Political Intensity (H1) Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Brand’s Perceived Political Affiliation Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.665*** (0.215) 0.624*** (0.072) 0.012 (0.011) -0.639*** (0.219) -0.071 (0.253) 2.269*** (0.808) Incl. Incl. Incl. Incl. 149 0.622 Brand Advocacy (2) 0.636*** (0.241) 0.608*** (0.072) 0.003 (0.011) -0.268 (0.224) -0.425 (0.316) 1.938** (0.793) Incl. Incl. Incl. Incl. 149 0.563 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. 92 APPENDIX J: TABLES AND FIGURES FOR STUDY 2A Table J1: Window Sales and Installation Business Options and Their Descriptions Study 2A Instructions Imagine that you want to purchase and install a new window at your home to replace a broken one. Read the descriptions of the five highly rated window sales and installation businesses to select one. Option 1 Option 2 Option 3 Option 4 Option 5 Window Pros: Window Pros is a small window purchase and installation service business that specializes in residential window installation. They offer a wide selection of high-quality windows, including energy- efficient options, and provide professional installation services to ensure a perfect fit. Quick Glass: Quick Glass is a mobile window repair and replacement service that offers fast and affordable solutions for small-scale window projects. They specialize in fixing cracked or foggy glass in homes and small commercial buildings and can often complete the job on the same day. Window Wise: Window Wise is a small window purchase and installation service business that provides cost-effective solutions for residential window installations. They offer a range of affordable window options, including energy-efficient windows, and provide expert installation services to ensure a long- lasting installation. Window Wizards: Window Wizards is a small window purchase and installation service business that specializes in custom window installations. They offer a wide range of custom window options, including stained glass and decorative windows, and provide expert installation services to ensure a beautiful, high- quality installation. Window Express: Window Express is a small window purchase and installation service business that focuses on providing quick and efficient window installations for residential customers. They offer a range of standard window options and provide fast installation services, often completing projects within a single day. 93 Table J2: Scenario Used - Study 2A Please read the scenario below carefully Next, imagine that you contact “Chosen Firm Name” to schedule a consultation. A representative visits your home to measure the window and provide recommendations for replacement options. You have some questions about the different types of windows and the installation process, and the representative answers all your questions patiently. After the consultation, you receive a quote for the work, which you find reasonable. You schedule the installation for the following week. On the day of the installation, the “Chosen Firm Name” team arrives on time and begins work right away. They carefully remove the old window and install the new one, taking care to minimize any disruption to the household. The installation process takes several hours, but you can continue with your daily activities without any major issues. The team finishes up the installations, cleans up the work area, and walks you through the features of your new window, including the improved insulation and ease of use. Table J3: Details of the Belief Superiority of Consumption Choice Measure Belief Superiority of Consumption Choice (Adopted from Ruvio et al. 2020) Items 1. 2. 3. 4. Please indicate the extent to which you agree with the following statements [1=Strongly Disagree to 7=Strongly Agree] Compared to others, I usually know the best product or service to buy. Not many people know the best products or services to buy as well as I do. I tend to buy better products or services than most people I know. I usually know where to get the best deals better than others. 94 Table J4: Correlations and Descriptive Statistics - Study 2A Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Political Affiliation (5) Customer Satisfaction (6) Gender (7) Education (8) Income (9) Age (10) Race Mean S. D. (1) 1 0.493 0.201 0.016 (2) (3) (4) (5) (6) (7) (8) (9) (10) 1 0.212 -0.061 1 -0.082 1 0.228 0.233 0.14 -0.22 1 -0.049 0.036 0.085 0.06 -0.07 5.894 0.782 -0.152 -0.02 0.12 0.065 0.029 5.747 0.787 0.156 0.018 0.01 0.207 0 1.848 1.244 0.036 0.022 0.036 0.041 -0.125 0.073 2.229 -0.1 0.098 0.121 0.068 0.061 5.386 1.607 1 -0.077 -0.025 0.058 -0.038 0.604 0.490 1 -0.034 0.085 -0.068 4.729 1.092 1 -0.056 -0.081 6.683 2.433 1 0.118 35.406 10.726 1 0.957 0.203 Notes: All correlations above 0.11 are statistically significant at p <0.05 or less. N = 303. 95 Table J5: Results for Main Effect (H1 - Panel A) and Mediation (H2 – Panel B) - Study 2A A. B: Variable Political Intensity Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Brand’s Perceived Political Affiliation Fixed Effects Variable Controlled Effects Political Intensity Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Brand’s Perceived Political Affiliation Fixed Effects Brand Advocacy S.E. 0.036 0.025 0.033 0.088 0.003 0.245 0.434 95% C.I. 0.068 to 0.208 -0.032 to 0.066 0.041 to 0.170 -0.237 to- 0.110 -0.004 to 0.010 -0.335 to 0.630 4.685 to 6.392 Hypothesis 1 (Main Effect) Customer Loyalty S.E. 0.043 0.016 0.035 95% C.I. Estimate 0.138*** 0.017 0.105*** 0.021 to 0.191 0.012 to 0.111 0.026 to 0.162 Estimate 0.106** 0.061** 0.094*** -0.005 0.002 -0.283 6.544*** Incl. Incl. Incl. Incl. 0.108 0.004 0.280 0.456 -0.217 to 0.207 -0.006 to 0.010 -0.835 to 0.269 5.646 to 7.441 -0.064** 0.003 0.147 5.538*** Incl. Incl. Incl. Incl. Hypothesis 2 (Mediation) Belief Superiority of Consumption Choice Estimate S.E. 95% C.I. 0.036 0.023 0.035 0.094 0.004 0.241 0.464 0.032 to 0.173 -0.012 to 0.079 -0.002 to 0.135 -0.209 to 0.160 -0.007 to 0.008 -0.086 to 0.862 4.981 to 6.808 0.102*** 0.033 0.067* -0.025 0.000 0.388 5.894*** Incl. Incl. Incl. Incl. Customer Loyalty S.E. 95% C.I. Estimate S.E. 95% C.I. Brand Advocacy 0.039 0.068 Estimate Variable Controlled Effects Political Intensity Belief Superiority of Consumption Choice Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Brand’s Perceived Political Affiliation Fixed Effects Indirect Effects via Belief Superiority of Consumption Choice Political Intensity 0.371*** 0.049** 0.070** 0.004 0.002 -0.426** 4.077*** Incl. Incl. Incl. 0.066 0.021 0.031 0.095 0.004 0.204 0.504 0.038** 0.016 Incl. -0.009 to 0.145 0.234 to 0.501 0.007 to 0.091 0.008 to 0.131 -0.184 to 0.192 -0.006 to 0.009 -0.828 to 0.025 3.084 to 5.070 0.010 to 0.072 0.067** 0.023 0.020 to 0.113 0.050 0.017 0.020 0.055 0.003 0.157 0.340 0.603 to 0.800 -0.039 to 0.027 0.020 to 0.097 -0.155 to 0.062 -0.002 to 0.008 -0.433 to 0.184 0.916 to 2.256 0.702*** -0.006 0.059** -0.046 0.003 -0.125 15861*** Incl. Incl. Incl. Incl. 0.072** 0.027 0.023 to 0.126 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. C.I. = Confidence Interval. Heteroscedasticity robust standard errors (S.E.). N=303. Models reported in Panel B are estimated using Hayes Process Model 4 with 10,000 bootstrap replications. 96 APPENDIX K: TABLES AND FIGURES FOR STUDY 2B Table K1: Blender Options and Their Descriptions – Study 2B Instructions Imagine that you are given the opportunity to purchase a high-end blender. Read the descriptions below to help you select one to purchase. Read the descriptions of the three blenders below and select one to purchase. Option 1 Option 2 Option 3 Velvet BlendWorks: Velvet BlendWorks is the epitome of luxury in kitchen appliances, offering unparalleled smooth blending with its advanced vortex technology. It features intuitive touch controls and a sleek, elegant design that complements any modern kitchen, perfect for those who seek both style and substance. ElixirEdge Pro: ElixirEdge Pro is crafted for the health enthusiast who demands the best. With its powerful motor and precision blades, it effortlessly pulverizes fruits, vegetables, and superfoods, extracting maximum nutrients and flavor. Its professional-grade performance makes it a favorite among fitness enthusiasts and chefs alike. SilkSpin Craft: Merging state-of-the-art blending technology with sustainable materials, SilkSpin Craft offers an eco-friendly solution without compromising on power. Ideal for the environmentally conscious consumer, SilkSpin Craft delivers silky-smooth textures, from frozen fruits to nut butter, while maintaining a low carbon footprint. Table K2: Scenario Used – Study 2B Please read the scenario below carefully Imagine that it has been a few years since you first purchased the “selected choice” blender. You have enjoyed using the blender for both making smoothies and cooking. It’s a typical experience with a blender overall. Table K3: Details of the Post-Purchase Involvement Measure – Study 2B Post-Purchase Involvement (Adapted from Zaichkowsky 1985 and Behe et al. 2015) Items 1. 2. 3. 4. Please indicate the extent to which you agree with the following statements [1=Strongly Disagree to 7=Strongly Agree] I would be interested in more information about this blender. I would be interested in reading consumer reviews about this blender. I would enjoy learning more about this blender. I would want to know how to get the most use out of this blender. 97 Table K4: Correlations and Descriptive Statistics - Study 2B Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Belief Sup. of Choice (5) Post-purchase Involvement (6) Political Affiliation (7) Customer Satisfaction (8) Gender (190) Age (10) Race (11) Education (12) Income Mean S. D. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) 1 0.6 0.277 0.212 0.445 0.013 0.29 0.026 0.104 -0.067 -0.044 0.075 5.826 1.089 1 0.204 0.305 0.543 -0.053 0.275 -0.173 0.097 -0.103 0.04 0.083 5.530 1.036 1 0.197 0.235 -0.214 0.170 -0.021 0.139 0.039 0.083 0.015 1.826 1.131 1 0.295 0.057 -0.023 0.07 -0.213 -0.14 0.101 0.151 4.542 1.333 1 -0.002 0.239 0.027 -0.009 -0.072 -0.098 -0.015 5.365 1.238 1 -0.029 0.062 0.127 -0.019 -0.79 0.039 3.343 2.048 1 -0.11 0.138 0.014 -0.046 -0.051 5.672 1.416 1 -0.129 -0.052 -0.124 -0.067 0.579 0.495 1 0.006 0.027 -0.006 40.877 11.022 1 -0.026 -0.089 0.805 0.397 1 0.343 4.436 1.197 1 6.667 3.017 Notes: All correlations above 0.14 are statistically significant at p <0.05 or less. N = 195. 98 Table K5: Results for Main Effect (H1 - Panel A) and Serial Mediation (H2 – Panel B) A. B: Variable Political Intensity Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Variable Controlled Effects Political Intensity Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Hypothesis 1 (Main Effect) Customer Loyalty S.E. 0.073 0.035 0.065 0.146 0.007 0.219 0.707 95% C.I. 0.1160 to 0.406 -0.032 to 0.106 0.099 to 0.356 -0.177 to 0.401 -0.012 to 0.034 -0.118 to 0.219 2.007 to 4.799 Estimate 0.261*** 0.037 0.228*** 0.112 0.001 -0.118 3.403*** Incl. Incl. Incl. Hypothesis 2 (Serial Mediation) Belief Superiority of Consumption Choice S.E. Estimate 95% C.I. 0.614 0.047 0.063 0.167 0.009 0.210 0.614 0.152 to 0.456 -0.018 to 0.167 -0.112 to 0.138 -0.161 to 0.496 -0.048 to -0.014 -0.819 to 0.010 3.847 to 6.272 0.304*** 0.075 0.013 0.168 -0.031*** -0.404* 5.059*** Incl. Incl. Incl. Brand Advocacy S.E. 0.074 0.040 0.061 0.137 0.006 0.180 0.825 95% C.I. 0.013 to 0.320 -0.074 to 0.085 0.074 to 0.314 -0.485 to 0.058 -0.017 to 0.007 -0.722 to -0.011 2.130 to 5.387 Estimate 0.159** 0.006 0.194*** -0.214 -0.005 -0.366** 3.758*** Incl. Incl. Incl. 99 Table K5 (cont’d) Variable Controlled Effects Political Intensity Political Affiliation Belief Superiority of Consumption Choice Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Post-Purchase Involvement S.E. Estimate 95% C.I. 0.079 0.035 0.078 0.061 0.165 0.008 0.192 0.712 -0.020 to 0.293 -0.043 to 0.093 0.124 to 0.432 0.124 to 0.365 -0.231to 0.421 -0.021 to 0.009 -0.591 to 0.166 1.597 to 4.409 0.136* 0.025 0.278*** 0.244*** 0.095*** -0.006 -0.212 3.003*** Incl. Incl. Incl. Customer Loyalty S.E. Estimate 95% C.I. Variable Controlled Effects Political Intensity Post-Purchase Involvement with Choice Belief Superiority of Consumption Choice Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Indirect Effects via Belief Superiority of Consumption Choice and Post-Purchase Involvement with Choice Political Intensity 0.047 to 0.293 0.151 to 0.421 -0.018 to 0.200 -0.039 to 0.073 0.042 to 0.270 -0.199 to 0.311 -0.005 to 0.020 -0.334 to 0.362 1.544 to 3.764 0.170*** 0.286*** 0.091 0.017 0.156*** 0.056 0.008 0.012 2.654*** Incl. Incl. Incl. 0.062 0.068 0.055 0.029 0.058 0.129 0.006 0.177 0.562 0.006 to 0.054 0.024** 0.013 Brand Advocacy S.E. 0.057 0.059 0.046 0.030 0.044 0.113 0.006 0.136 0.461 95% C.I. -0.084 to 0.145 0.262 to 0.495 0.056 to 0.237 -0.082 to 0.037 0.011 to 0.186 -0.516 to -0.068 -0.006 to 0.016 -0.452 to 0.085 1.019 to 2.839 Estimate 0.031 0.378*** 0.147*** -0.023 0.098** -0.292** 0.005 -0.184 1.929** Incl. Incl. Incl. 0.032** 0.016 0.009 to 0.069 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. C.I. = Confidence Interval. Heteroscedasticity robust standard errors (S.E.). N=195. Models reported in Panel B are estimated using Hayes Process Model 6 (5,000 bootstrap replications). 100 APPENDIX L: DETAILS OF COMPLEMENTARY STUDY 2BW In this study, we validate the main effects of political intensity on customers’ loyalty and advocacy (H1) and the mechanisms of belief superiority of consumption choice and subsequent post-purchase involvement with the choice (H2) from study 2B utilizing customers’ actual experiences with real brands in the context of durable products (cell phones, a high-involvement product category). Design and Procedure. We recruited a total of 294 U.S.-based respondents from Amazon Mechanical Turk who were provided with appropriate compensation. Respondents were screened before participation in the study through an eligibility screening question.21 They were then asked to think about their experience with their cell phones and answered questions evaluating their experiences, belief superiority of consumption choice, post-purchase involvement, political ideology, and other relevant covariates. Measures. We use the same measures of all the key variables (political intensity, customer loyalty, belief superiority of consumption choice, and post-purchase involvement with choice) and covariates (customer demographics; gender, age, race, education, and income) as in study 2B except for advocacy, which is captured using the same single-item measure used in study 1B. Given that this is a recall study, we measure and account for respondents’ overall attitude towards the brand (1=Very negative; 7=Very positive), satisfaction with the product (cell phone), and perceptions about the brand’s political affiliation. Table L1 provides correlations among all the variables included in this study along with their descriptive statistics. 21Participants were asked to select the wireless phone brand (from a list of five – Apple, Samsung, Google, Motorola, OnePlus) which they had purchased in the past 12 months (including a “none of the above” option which terminated the survey). The order of presentation of brands in the list was randomized to avoid ordering bias. 101 Estimation and Results. To test H1 regarding the main effects of political intensity, we use a multiple regression model with (cell phone) brand fixed effects to account for brand-level unobserved factors and robust standard errors (results are consistent with clustering of standard errors by brands). Customer loyalty and advocacy serve as dependent variables and political intensity is the key independent variable, along with all the relevant control variables. The results from this analysis are summarized in Panel A of Table L2. Validating results from studies 1A, 1B, 2A, and 2B and in support of H1, we find that political intensity has a strong positive effect on loyalty (.339, p < .01) and on advocacy (.298, p < .01). To test H2 regarding the mediating role of customers’ belief superiority of consumption choice and post-purchase involvement, we run additional models (Hayes 2009, Model 6, with 5,000 bootstrap samples) with political intensity as the key independent variable and all controls, including brand fixed effects and heteroskedasticity robust standard errors. Results are summarized in Panel B of Table L2. Supporting H2, we find that customers’ belief superiority of the consumption choice and post-purchase involvement mediate the positive effect of political intensity on loyalty (.048, 95% CI: .005 to .112) and advocacy (.064, 95% CI: .011 to .123). 102 Table L1: Correlations and Descriptive Statistics - Study 2BW Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Believed Sup. Of Choice (5) Post Purchase Involvement (6) Political Affiliation (7) Customer Satisfaction (8) Attitude Towards Company (9) Gender (10) Age (11) Race (12) Education (13) Income Mean S. D. (1) 1 0.554 0.403 0.377 0.465 0.121 0.29 0.401 -0.229 0.006 -0.078 0.082 0.083 6.037 1.166 (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) 1 0.407 0.42 0.566 0.2 0.299 0.399 -0.106 0.014 -0.038 0.066 0.071 6.099 0.946 1 0.373 0.421 0.338 0.248 0.259 -0.127 0.015 0.007 0.15 0.104 2.034 0.904 1 0.773 0.265 0.188 0.265 -0.013 -0.053 -0.029 0.03 0.179 5.683 0.832 1 0.247 0.334 0.372 -0.08 -0.001 0.034 0.089 0.191 5.872 0.790 1 0.145 0.01 0.095 -0.218 -0.171 0.026 0.044 5.204 1.875 1 0.62 0.007 0.013 -0.031 0.083 0.202 5.884 1.417 1 -0.141 0.016 -0.025 0.075 0.167 5.827 1.393 1 -0.023 -0.017 0.067 0.034 0.629 0.484 1 0.284 -0.039 0.04 34.388 8.889 1 0.114 0.045 0.827 0.379 1 0.143 4.878 1.011 6.857 2.549 Notes: All correlations above 0.08 are statistically significant at p <0.05 or less. N = 294. 103 Table L2: Results for Main Effect (H1 – Panel A) and Mediation (H2 – Panel B) – Study 2BW A. B: Variable Political Intensity Political Affiliation Attitude Towards the Company Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand’s Perceived Pol.Affiliation Fixed Effects Brand Fixed Effects Variable Controlled Effects Political Intensity Political Affiliation Attitude Towards the Company Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand’s Perceived Pol. Affiliation Fixed Effects Brand Fixed Effects Hypothesis 1 (Main Effect) Customer Loyalty S.E. 0.103 0.041 0.057 0.064 0.142 0.007 0.141 0.559 95% C.I. 0.135 to 0.543 -0.078 to 0.083 0.098 to 0.323 -0.084 to 0.166 -0.725 to 0.168 -0.013 to 0.013 -0.473 to 0.082 2.978 to 5.180 Estimate 0.339*** 0.003 0.211*** 0.041 -0.447*** 0.000 -0.195 4.079*** Incl. Incl. Incl. Incl. Hypothesis 2 (Serial Mediation) Brand Advocacy S.E. 0.085 0.035 0.053 0.060 0.105 0.006 0.161 0.464 95% C.I. 0.131 to 0.465 -0.014 to 0.123 0.076 to 0.286 -0.105 to 0.132 -0.201 to 0.211 -0.011 to 0.012 -0.408 to 0.227 3.353 to 5.180 Estimate 0.298*** 0.054 0.181*** 0.013 0.005 0.000 -0.091 4.267*** Incl. Incl. Incl. Incl. Belief Superiority of Consumption Choice S.E. Estimate 95% C.I. 0.073 0.032 0.044 0.048 0.097 0.006 0.137 0.434 0.043 to 0.331 0.031 to 0.156 0.042 to 0.215 -0.089 to 0.101 -0.099 to 0.283 -0.017 to 0.006 -0.050 to 0.490 3.855 to 5.564 0.187** 0.093*** 0.128 0.006 0.092 -0.005 0.219 4.710*** Incl. Incl. Incl. Incl. 104 Table L2 (cont’d) Variable Controlled Effects Political Intensity Political Affiliation Belief Superiority of Consumption Choice Attitude Towards the Company Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand’s Perceived Pol. Affiliation Fixed Effects Brand Fixed Effects Post-Purchase Involvement with Choice S.E. Estimate 95% C.I. 0.044 0.018 0.061 0.024 0.027 0.059 0.003 0.072 0.418 0.003 to 0.177 -0.021 to 0.050 0.517 to 0.757 0.003 to 0.098 -0.009 to 0.096 -0.196 to 0.036 -0.005 to 0.007 -0.027 to 0.257 0.220 to 1.867 0.090** 0.014 0.637*** 0.050** 0.044 -0.080 0.001 0.115 1.044** Incl. Incl. Incl. Incl. Customer Loyalty S.E. Estimate 95% C.I. Variable Controlled Effects Political Intensity Post-Purchase Involvement with Choice Belief Superiority of Consumption Choice Political Affiliation Customer Satisfaction Gender Age Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand’s Perceived Pol. Affiliation Fixed Effects Brand Fixed Effects Indirect Effects via Belief Superiority of Consumption Choice and Post-Purchase Involvement with Choice Political Intensity 0.247*** 0.406*** 0.041 -0.031 0.022 -0.442*** 0.001 -0.308 2.642*** Incl. Incl. Incl. Incl. 0.060 to 0.433 0.121 to 0.691 -0.174 to 0.255 -0.107 to 0.045 -0.081 to 0.125 -0.701 to -0.183 -0.011 to 0.014 -0.549 to -0.067 1.119 to 4.166 0.095 0.145 0.109 0.039 0.052 0.132 0.006 0.122 0.774 0.005 to 0.112 0.048** 0.027 Estimate Brand Advocacy S.E. 95% C.I. 0.072 0.113 0.086 0.029 0.048 0.095 0.005 0.127 0.656 0.058 to 0.342 0.313 to 0.757 -0.244 to 0.94 -0.036 to 0.079 -0.106 to 0.082 -0.106 to 0.082 -0.009 to 0.012 -0.461 to 0.040 1.415 to 3.998 0.200*** 0.535*** -0.075 0.022 -0.012 0.024 0.001 -0.211* 2.706*** Incl. Incl. Incl. Incl. 0.064** 0.028 0.011 to 0.123 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. C.I. = Confidence Interval. Heteroscedasticity robust standard errors (S.E.). N=294. Models reported in Panel B are estimated using Hayes Process Model 6 with 5,000 bootstrap replications. 105 APPENDIX M: TABLES AND FIGURES FOR STUDY 3 Figure M1: Scenarios Used for Manipulation (adopted from Pratt et al. 2022) Competitor's Price Promotion (Control Condition) “Picture containing Coke and Pepsi bottles placed on a supermarket shelf both priced at $2.09.” Competitor’s Price Promotion Present (Manipulation Condition) “Picture containing Coke and Pepsi bottles placed on a supermarket shelf with Coke priced at $2.09 and Pepsi priced at $1.09.” “Picture containing Coke and Pepsi bottles placed on a supermarket shelf with Pepsi priced at $2.09 and Coke priced at $1.09.” OR Table M1: Details of the Multi-Item Customer Loyalty Measure Customer Loyalty (Adopted from Pratt et al. 2022) Items 1. 2. 3. Please indicate the extent to which you agree with the following statements [1=Strongly Disagree to 7=Strongly Agree] I consider myself loyal to this [brand]. I really love [brand]. I would really miss [brand] if it went away. 106 Table M2: Correlations and Descriptive Statistics - Study 3 Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Political Affiliation (5) External Threat (Competitor Price Promotion) (6) Customer Satisfaction (7) Attitude Towards the company (8) Age (9) Gender (10) Race (11) Education (12) Income Mean S. D. (1) 1 0.649 0.483 0.206 -0.092 0.191 0.384 -0.029 0.065 -0.154 -0.042 0.073 5.873 0.780 (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) 1 0.362 0.151 -0.047 0.208 0.372 -0.066 0.090 -0.092 0.000 0.048 6.007 0.925 1 0.200 -0.156 0.107 0.187 0.031 0.002 -0.185 -0.069 -0.052 2.065 0.761 1 0.181 -0.098 0.101 -0.028 0.168 -0.125 -0.076 0.051 5.168 1.867 1 -0.196 -0.105 -0.013 0.019 0.082 -0.013 -0.117 0.717 0.451 1 0.224 -0.106 -0.138 -0.165 0.120 0.114 5.491 1.679 1 -0.041 0.003 -0.219 0.056 0.134 23.735 1.206 1 -0.009 0.187 -0.028 0.004 35.133 8.067 1 -0.19 -0.235 0.064 0.710 0.455 1 -0.077 -0.03 0.878 0.328 1 -0.011 4.566 1.387 1 7.122 2.468 Notes: All correlations above 0.12 are statistically significant at p <0.05 or less. N = 279. 107 Table M3: Results for Main Effect (H1) and Moderation (H4) – Study 3 Predictor Variables Political Intensity (H1) Political Intensity *External Threat (Competitor Price Promotion) (H3) External Threat (Competitor Price Promotion) Political Affiliation Customer Satisfaction Attitude Towards the Company Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Perceived Political Affiliation Fixed Effects Brand Fixed Effects N (Customers) R-squared Brand Advocacy Interaction Model Customer Loyalty Main Effect Model (1) 0.399*** (0.065) -- Main Effect Model (3) 0.134 0.365*** (0.080) -- (2) Interaction Model (4) 0.079 (0.173) 0.384** (0.189) 0.023 (0.123) 0.006 (0.031) 0.072** (0.031) 0.240*** (0.049) -0.009 (0.010) 0.160 (0.131) 0.124 (0.197) -1.048 (1.177) Incl. Incl. Incl. Incl. 279 0.352 (0.127) 0.355** (0.141) -0.059 (0.096) 0.003 (0.026) 0.044* (0.026) 0.069 (0.118) 0.017 (0.031) 0.067** (0.031) 0.174*** 0.239*** (0.050) -0.009 (0.010) 0.143 (0.131) 0.169 (0.200) -0.975 (1.183) Incl. Incl. Incl. Incl. 279 0.336 (0.040) -0.002 (0.006) 0.079 (0.101) -0.064 (0.141) -0.021 (0.946) Incl. Incl. Incl. Incl. 279 0.474 -0.016 (0.093) 0.013 (0.027) 0.040 (0.026) 0.173*** (0.040) -0.002 (0.006) 0.064 (0.101) -0.022 (0.139) 0.047 (0.953) Incl. Incl. Incl. Incl. 279 0.454 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Intensity was mean-centered in the interaction models (2 and 4) before creating the interactions. 108 APPENDIX N: TABLES AND FIGURES FOR STUDY 4 Table N1: Restaurant Options and Their Descriptions - Study 4 Instructions Imagine that you want to celebrate a special occasion with your significant other and you need to select a restaurant for the occasion. Read the descriptions of the five highly rated restaurants below to help you select one for the occasion. Option 1 Option 2 Option 3 Option 4 Option 5 Truly Asian: An Asian restaurant founded on the concept of making food from scratch every day. Its chefs have a tradition of hand-rolling dim sums, hand chopping and slicing all vegetables and meats, scratch- cooking every sauce, and wok-cooking each dish. El Centro: At El Centro, it is all about authentic Mexican cuisine. Walking into El Centro is like walking into your grandmother’s kitchen, that is if your grandmother’s kitchen is in Hermosillo, Mexico. Mighty Barbeque: It brings together the great BBQ techniques to create something uniquely its own. The process begins with all-natural meats and poultry, seasoned with spice blends, and smoked with wood for many hours until the harmony of smoke, flavor and time emerges. Keeva Indian: At Keeva, chefs believe good food is all about patience and doing small things right. They prepare their dishes using traditional Indian techniques that were developed over 5,000 years ago and beautifully blended with current culinary techniques. Celeste Italian: The restaurant pays homage to the essence of the great Italian restaurants of the mid-20th century, where delicious, exceptionally well-prepared food was served in settings that were simultaneously elegant, comfortable, and unpretentious. Notes. To mimic real-world options available to customers, we selected five real New York City-based restaurants with four-star ratings on Google reviews and used their actual descriptions (with fictitious names). Table N2: Scenarios Used in Service Failure and Control Conditions - Study 4 Scenario Service Failure Condition Description After exploring and evaluating the five restaurants, you decide to go for lunch to the highly rated, well-known “Chosen Restaurant Name” restaurant. However, at the restaurant you receive very poor service, the waiter is unpleasant, you waste a lot of time waiting for your order and the food is served cold. A complete disaster! You are very disappointed that you wasted money on lunch and did not enjoy it at all. It’s just a very poor experience overall. Control Condition After exploring and evaluating the five restaurants, you decide to go for lunch to the highly rated, well-known “Chosen Restaurant Name” restaurant. At the restaurant, you receive average service. The waiter is fine. The food is decent but not outstanding. It’s just a typical, average experience overall. 109 Table N3: Correlations and Descriptive Statistics - Study 4 Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Political Affiliation (5) Customer Satisfaction (6) Age (7) Gender (8) Education (9) Income (10) Race Mean S. D. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) 1 0.854 0.373 0.302 0.661 0 0.071 0.028 0.02 -0.082 5.323 1.541 1 0.363 0.343 0.603 -0.039 0.11 0.027 0.063 -0.075 5.471 1.532 1 0.563 0.263 -0.005 0.185 -0.064 -0.13 -0.181 1.316 0.835 1 0.212 0.016 0.199 0.004 -0.125 -0.125 4.816 1.328 1 -0.056 0.016 0.092 0.042 -0.096 5.204 1.652 1 -0.012 0.07 -0.11 0.084 34.794 9.188 1 -0.133 -0.098 -0.048 0.638 0.481 1 0.13 -0.123 4.615 1.326 1 0.104 6.159 2.469 1 0.934 0.249 Notes: All correlations above 0.08 are statistically significant at p <0.05 or less. N = 603. Table N4: Results for Main Effect (H1) and Moderation (H3) - Study 4 Predictor Variables Political Intensity (H1) Political Intensity*Internal Threat (Service Failure) (H4) Internal Threat (Service Failure) Political Affiliation Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Restaurant Brand Fixed Effects Restaurant Brand Perceived Political Affiliation Fixed Effects N (Customers) R-squared Main Effect Model (1) 0.357*** (0.074) -- Customer Loyalty Interaction Model (2) 0.204*** (0.078) 0.293*** (0.102) -0.351*** (0.092) 0.046 (0.042) 0.494*** (0.042) -0.010* (0.005) 0.119 (0.104) -0.006 (0.187) 2.544*** (0.615) Incl. Incl. Incl. Incl. 603 0.550 -0.343*** (0.091) 0.042 (0.041) 0.505*** (0.042) -0.010* (0.005) 0.122 (0.104) 0.011 (0.186) 2.474*** (0.577) Incl. Incl. Incl. Incl. 603 0.544 Main Effect Model (3) 0.291*** (0.077) -- Brand Advocacy Interaction Model (4) 0.167** (0.081) 0.238** (0.107) -0.233** (0.098) 0.119*** (0.044) 0.430*** (0.048) -0.004 (0.005) 0.180* (0.107) -0.019 (0.209) 2.956*** (1.069) Incl. Incl. Incl. Incl. 603 0.498 -0.227** (0.097) 0.116*** (0.044) 0.438*** (0.047) 0.005 (0.006) 0.182* (0.107) -0.006 (0.210) 2.899*** (0.991) Incl. Incl. Incl. Incl. 603 0.494 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Intensity was mean-centered in the interaction models (2 and 4) before creating the interaction 110 APPENDIX O: TABLES AND FIGURES FOR STUDY 5 Table O1: Hotel Options and Their Descriptions - Study 5 Instructions Imagine that you are planning to visit a big city for a weekend leisure trip. Read the descriptions of the three highly-rated budget hotels and select one Option 1 Option 2 Option 3 Urban Hub Hotel: This modern hotel, situated in the bustling city center, offers a sleek design with glass and steel accents. The minimalist rooms are perfectly optimized for both comfort and functionality, ideal for travelers seeking convenience. Its prime location near major public transport hubs makes it an excellent choice for exploring the city's attractions. Central City Value Stay: A unique budget option, this hotel stands out with its vibrant, artsy exterior featuring colorful murals. The interior is equally stylish, with compact rooms that offer essential amenities for a comfortable stay. Surrounded by lively cafes, shops, and cultural attractions, it's perfect for guests wanting to dive into the local urban experience. Metropolitan Budget Hotel: Positioned in the dynamic core of the city, this hotel offers a harmonious blend of affordability and strategic location. It's classic urban architecture and straightforward design appeal to both business travelers and tourists. The hotel provides functional, cozy rooms and is conveniently located near key business districts, shopping centers, and entertainment venues, making it an ideal base for city exploration. Table O2: Description of the Stay at the Hotel - Study 5 Below is a summary of your stay at the “name of chosen” hotel. Morning Check-In: You arrive at the hotel around 10 AM. The check-in process is routine, and you're assigned a standard room with basic amenities. Utilizing Hotel Facilities: In the afternoon, you spend a short time in the hotel's small lounge area. The hotel lacks additional amenities like a pool or gym. Overnight Stay: Your night at the hotel is uneventful, with the room providing basic comfort for a night's sleep. Morning Departure: The following morning, after a standard continental breakfast at the hotel, you check out by 9 AM to continue your travels. 111 Table O3: Scenarios Used in Manipulation and Control Conditions - Study 5 A: Scenarios Politically Conservative Hotel Scenarios Used to Manipulate Hotel’s Political Affiliation Through Activism Description Dear guest, I would like to extend my heartfelt thanks for choosing to stay with us during your recent visit to the city. It was truly a pleasure to host you. We’d like to let you know our stance that we support gun freedom. Accordingly, our hotel will donate a portion of the proceeds from your booking directly to organizations that publicly support gun rights through campaigns, candidates, and activism. Should you have any special requests or need further assistance, please don't hesitate to contact us. Warm regards, Manager “Name of Chosen Hotel” Politically Liberal Hotel Dear guest, I would like to extend my heartfelt thanks for choosing to stay with us during your recent visit to the city. It was truly a pleasure to host you. We’d like to let you know our stance that we support gun control. Accordingly, our hotel will donate a portion of the proceeds from your booking directly to organizations that publicly support gun control through campaigns, candidates, and activism. Should you have any special requests or need further assistance, please don't hesitate to contact us. Warm regards, Manager “Name of Chosen Hotel” Dear guest, I would like to extend my heartfelt thanks for choosing to stay with us during your recent visit to the city. It was truly a pleasure to host you. Should you have any special requests or need further assistance, please don't hesitate to contact us. Warm regards, Manager “Name of Chosen Hotel” Construction of the Misalignment/Alignment Conditions No Political Affiliation Information About Hotel B: Political Affiliation of Respondent Liberal Conservative Liberal Conservative Independent Independent Liberal Conservative Independent Political Affiliation Scenario Liberal Hotel Conservative Hotel Conservative Hotel Liberal Hotel Conservative Hotel Liberal Hotel No affiliation information No affiliation information No affiliation information Alignment Condition based on Matching Aligned Aligned Mis-aligned Mis -aligned Control Control Control Control Control Notes. Manipulation check: Overall, we find strong support for our manipulation. Towards the end of the survey, respondents were asked “Based on the information you read about “selected hotel” in this survey, what is its position on gun control? (1=Supports gun control, 2=Supports gun freedom; 3=I don’t know/no such information provided in the survey). 91.64% of the respondents answered this question correctly. 112 Table O4: Correlations and Descriptive Statistics - Study 5 Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Political Affiliation (5) Customer Satisfaction (6) Age (7) Gender (8) Race (9) Education (10) Income Mean S. D. (1) 1 0.931 0.165 0.226 0.778 0.136 0.028 0.068 0.15 -0.021 4.579 2.055 (2) (3) (4) (5) (6) (7) (8) (9) (10) 1 0.22 0.22 0.756 0.13 -0.03 0.053 0.166 -0.019 4.512 2.052 1 0.017 0.066 0.114 -0.124 0.076 0.165 0.021 1.508 1.066 1 0.145 0.124 0.028 -0.053 -0.042 0.067 3.883 1.845 1 0.165 0.012 0.1 0.117 0 4.773 1.797 1 -0.155 0.116 0.05 -0.003 41.151 12.329 1 -0.023 0.033 0.003 0.505 0.501 1 0.056 0.042 0.836 0.371 1 0.333 4.448 1.298 1 6.759 3.047 Notes: All correlations above 0.102 are statistically significant at p <0.05 or less. N = 299. 113 Table O5: Results for Main Effect (H1) and Moderation (H5) - Study 5 Predictor Variables Political Intensity (H1) Political Intensity *Customer-Brand Political Misalignment (H5) Political Intensity * Customer-Brand Political Alignment Customer-Brand Political Misalignment Customer-Brand Political Alignment Political Affiliation Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Main Effect Model (1) 0.291*** (0.080) -- -- Customer Loyalty Interaction Model (2) 0.384*** (0.101) -0.375** (0.190) -0.204 (0.208) -1.082*** (0.241) -0.006 (0.201) 0.132*** (0.040) 0.747*** (0.056) -0.003 (0.006) 0.095 (0.148) -0.035 (0.223) 1.388*** (0.533) Incl. Incl. Incl. 299 0.690 -1.133*** (0.244) -0.026 (0.182) 0.131*** (0.040) 0.754*** (0.055) -0.004 (0.006) 0.098 (0.148) 0.004 (0.215) 1.280** (0.536) Incl. Incl. Incl. 299 0.686 Main Effect Model (3) 0.436*** (0.080) -- -- Brand Advocacy Interaction Model (4) 0.525*** (0.096) -0.454** (0.185) -0.108 (0.216) -0.982*** (0.250) -0.281 (0.216) 0.131*** (0.041) 0.731*** (0.059) -0.005 (0.006) -0.145 (0.155) -0.110 (0.219) 1.906*** (0.535) Incl. Incl. Incl. 299 0.673 -1.057*** (0.250) -0.254 (0.195) 0.131*** (0.041) 0.742*** (0.058) -0.005 (0.006) -0.144 (0.155) -0.073 (0.212) 1.763*** (0.538) Incl. Incl. Incl. 299 0.668 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Intensity was mean-centered in the interaction models (2 and 4) before creating the interactions. 114 APPENDIX P: DETAILS OF COMPLEMENTARY STUDY 5W In this study, we validate the main effects of political intensity on customers’ loyalty and advocacy (H1) and the moderating role of the internal threat of customer-brand political misalignment (H5) from study 5 utilizing customers actual experiences with real brands in the context of services (fast food chains). Design and Procedure. We recruited 448 respondents from Cloud Research through two eligibility screening questions, 1) indicating that they had eaten at or ordered take-out/delivery from one of six listed fast-food chains over the prior 12 months and 2) had no prior knowledge about the chosen fast-food chain’s political affiliation22. The latter is an important consideration for subsequent manipulation of firms’ political affiliations to examine customer-firm political misalignment. Participants were randomly allocated to one of the three scenarios, where they read a brief description of a new loyalty rewards program being introduced by their selected fast-food chain. The three scenarios include 1) a Democratic scenario, where the fast-food chain was presented as having a strong affiliation with the Democratic Party; 2) a Republican scenario, where the fast-food chain was presented as having a strong affiliation with the Republican Party; and 3) a no affiliation-scenario, where no information regarding fast-food chain’s political party affiliation was offered. See Panel A of Table P1 for complete information about these three 22Participants were asked to select one fast food restaurant (from a list of six) which they had visited most frequently in the past 12 months (including a “none of the above” option which terminated the survey). Subsequent scenarios and questions in the survey concerned the restaurant selected. The order of presentation of restaurants in the list was randomized to avoid potential ordering bias. The restaurants included in the list were: Taco Bell, Burger King, Dunkin’ Donuts, Panera Bread, Pizza Hut, and KFC. These restaurants were chosen for the study because they do not affiliate strongly with either the Democratic or the Republican Party (i.e., contributing to both in roughly equal quantities or not contributing at all to either party), based on the composition of their political donations during the last 2020 presidential election cycle. 115 scenarios and the manipulation check. Subsequently, participants responded to a sequence of questions concerning the constructs of interest to this study. Measures. We utilize the same measures used in Study 5 for all of the common variables except political intensity, which is based on information about customers’ political party affiliation (versus political ideology in Study 5) used in Study 2A. Additionally, given that this is a recall study, we measure and account for their respondents’ overall attitude towards the brand (same measure as used in studies 2AW and 3), satisfaction with the fast food, and perceptions about the brands’ political affiliation. Table P2 provides correlations among the variables included in this study along with their descriptive statistics. Similar to study 5, while our objective in this study is to test the moderating role of customer-firm political misalignment (H5) our study setup allows for testing moderation by alignment as well. To test the moderation effect of customer-firm political misalignment, participants were divided into three groups – misaligned, aligned, and control – based on the extent of misalignment or alignment between their own revealed party affiliation and the scenario presented to them during the survey. Detailed information regarding this matching process can be found in Panel B of Table P1. Table P2 provides details of the correlations among the variables and their descriptive statistics. Table Q2 provides correlations among all the variables included in this study along with their descriptive statistics. Estimation and Results. We use the same model specification as study 5 to test the main effect of political intensity (H1) and the moderating role of customer-firm political misalignment (H5). The results from this analysis are presented in Table P3, wherein the main effect of political intensity is substantiated for loyalty (.183, p < .01) and advocacy (.207, p < .01). Further, in support of H5, such effects are observed to be weaker in the case of misalignment on 116 both loyalty (-.416, p < .05) and advocacy (-.334, p < .10). Lastly, similar to study 5, alignment does not seem to significantly alter the strength of the effects of customers’ political intensity. 117 Table P1: Scenarios Used in Manipulation and Control Conditions – Study 5W A: Scenarios Politically Republican Fast-Food Chain Scenarios Used to Manipulate Hotel’s Political Affiliation Description “Selected Fast Food Chain,” a national restaurant chain that predominately donates to and supports Republican causes, is planning to launch a new free-of-cost loyalty rewards program for its customers. As part of this new program, members could collect points and redeem them for free items, enjoy additional customization options for both foods and beverages and receive other exclusive perks. Members will also have the option of donating the points earned through purchases to the Republican National Committee. Politically Democratic Fast-Food Chain “Selected Fast Food Chain,” a national restaurant chain that predominately donates to and supports Democratic causes, is planning to launch a new free-of-cost loyalty rewards program for its customers. As part of this new program, members could collect points and redeem them for free items, enjoy additional customization options for both foods and beverages and receive other exclusive perks. Members will also have the option of donating the points earned through purchases to the Democratic National Committee. No Political Affiliation Information About Fast- Food Chain “Selected Fast Food Chain,” a national restaurant chain, is planning to launch a new free-of-cost loyalty rewards program for its customers. As part of this new program, members could collect points and redeem them for free items, enjoy additional customization options for both foods and beverages and receive other exclusive perks. Members will also have the option of donating the points earned through purchases to a charitable organization of their choice. B: Political Affiliation of Respondent Democratic Republican Democratic Republican Independent Independent Democratic Republican Independent Construction of the Alignment Conditions Political Affiliation Scenario Democratic Fast-Food Chain Republican Fast-Food Chain Republican Fast-Food Chain Democratic Fast-Food Chain Republican Fast-Food Chain Democratic Fast-Food Chain No affiliation information No affiliation information No affiliation information Alignment Condition based on Matching Aligned Aligned Mis-aligned Mis -aligned Control Control Control Control Control Notes. Manipulation check: Overall, we find strong support for our manipulation. Towards the end of the survey, respondents were asked “Based on the information you read about “selected hotel” in this survey, what is its position on gun control? (1=Supports gun control, 2=Supports gun freedom; 3=I don’t know/no such information provided in the survey). 93.15% of the respondents answered this question correctly. 118 Table P2: Correlations and Descriptive Statistics – Study 5W Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Political Affiliation (5) Customer Satisfaction (6) Age (7) Gender (8) Race (9) Education (10) Income Mean S. D. (1) 1 0.791 0.126 0.075 0.409 0.01 -0.081 0.006 0.084 -0.004 5.301 1.585 (2) (3) (4) (5) (6) (7) (8) (9) (10) 1 0.144 0.089 0.419 0.027 -0.116 0.005 0.101 0.026 5.141 1.700 1 -0.112 0.058 0.133 -0.104 -0.026 0.102 0.068 1.583 1.167 1 -0.006 -0.036 0.063 -0.114 -0.037 0.075 3.739 1.950 1 0.054 -0.124 0.109 0.065 0.05 5.482 1.513 1 -0.041 0.102 -0.07 0.011 40.962 11.178 1 -0.041 0.013 0.016 0.302 0.460 1 -0.119 0.054 0.436 0.496 1 0.283 4.498 1.224 1 6.676 2.942 Notes: All correlations above 0.09 are statistically significant at p <0.05 or less. N = 448. 119 Table P3: Results for Main Effect (H1) and Moderation (H5) - Study 5W Predictor Variables Political Intensity (H1) Political Intensity *Political Misalignment (H5) Political Intensity *Political Alignment Political Misalignment Political Alignment Political Affiliation Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Main Effect Model (1) 0.183*** (0.061) -- -- Customer Loyalty Interaction Model (2) 0.283*** (0.072) -0.416** (0.190) -0.129 (0.143) -0.925*** (0.201) 0.299* (0.154) 0.051 (0.032) 0.363*** (0.054) 0.001 (0.007) -0.122 (0.131) -0.120 (0.178) 3.970*** (0.802) Incl. Incl. Incl. 448 0.323 -1.002*** (0.191) 0.340** (0.145) 0.055* (0.032) 0.367*** (0.055) 0.001 (0.007) -0.133 (0.133) -0.116 (0.182) 3.642*** (0.813) Incl. Incl. Incl. 448 0.313 Main Effect Model (3) 0.207*** (0.069) -- -- -0.924*** (0.208) 0.335** (0.163) 0.073** (0.035) 0.415*** (0.060) 0.002 (0.007) -0.233 (0.145) -0.144 (0.191) 3.921*** (0.834) Incl. Incl. Incl. 448 0.312 Brand Advocacy Interaction Model (4) 0.321*** (0.089) -0.334* (0.183) -0.265 (0.180) -0.901*** (0.197) 0.358* (0.200) 0.072* (0.037) 0.410*** (0.048) 0.003 (0.007) -0.219 (0.145) -0.144 (0.189) 4.324*** (0.841) Incl. Incl. Incl. 448 0.319 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Intensity was mean-centered in the interaction models (2 and 4) before creating the interactions. 120 APPENDIX Q: CURVILINEAR ESTIMATION (REMAINING STUDIES) Additional Analysis – Studies 2a, 2b, 2bw, 3, 4, 5, and 5w. Table Q1: Results for Main Effect of Political Intensity (H1) - Study 2A Predictor Variables Political Affiliation Political Affiliation2 (H1) Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Brand's Perceived Political Affiliation Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.067*** (0.025) 0.038*** (0.013) 0.089*** (0.034) 0.001 (0.004) -0.007 (0.105) -0.279 (0.280) 6.567*** (0.457) Incl. Incl. Incl. Incl. 303 0.243 Brand Advocacy (2) 0.026 (0.025) 0.044*** (0.011) 0.101*** (0.033) 0.003 (0.003) -0.059 (0.087) 0.143 (0.241) 5.612*** (0.421) Incl. Incl. Incl. Incl. 303 0.359 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 121 Table Q2: Results for Main Effect of Political Intensity (H1) - Study 2B Predictor Variables Political Affiliation Political Affiliation2 (H1) Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) -0.066 (0.044) 0.079*** (0.020) 0.230*** (0.065) 0.000 (0.007) 0.102 (0.146) -0.110 (0.221) 3.527*** (0.679) Incl. Incl. Incl. 195 0.266 Brand Advocacy (2) -0.062 (0.043) 0.053** (0.021) 0.193*** (0.060) -0.006 (0.006) -0.221 (0.138) -0.362** (0.180) 3.828*** (0.812) Incl. Incl. Incl. 195 0.247 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. Table Q3: Results for Main Effect of Political Intensity (H1) - Study 2BW Predictor Variables Political Affiliation Political Affiliation2 (H1) Attitude Towards the Brand Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand‘s Political Affiliation Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.241*** (0.070) 0.099*** (0.030) 0.034 (0.066) 0.204*** (0.058) 0.001 (0.007) -0.473*** (0.140) -0.220 (0.143) 4.477*** (0.577) Incl. Incl. Incl. Incl. 294 0.361 Brand Advocacy (2) 0.270*** (0.051) 0.090*** (0.024) 0.005 (0.061) 0.174*** (0.054) 0.001 (0.005) -0.015 (0.105) -0.115 (0.166) 4.627*** (0.491) Incl. Incl. Incl. Incl. 294 0.360 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 122 Table Q4: Results for Main Effect (H1) – Study 3 Predictor Variables Political Affiliation Political Affiliation2 (H1) External Threat (Compettor’s Price Promotion) Customer Satisfaction Attitude Towards the Company Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand’s Perceived Political Affiliation Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.257*** (0.047) 0.092*** (0.017) -0.026 (0.093) 0.045* (0.027) 0.176*** (0.041) -0.001 (0.006) 0.058 (0.103) -0.010 (0.140) -0.412 (0.953) Incl. Incl. Incl. Incl. 279 0.424 Brand Advocacy (2) 0.246*** (0.062) 0.086*** (0.021) 0.063 (0.118) 0.071** (0.032) 0.240*** (0.051) -0.008 (0.010) 0.140 (0.131) 0.183 (0.203) -1.400 (1.180) Incl. Incl. Incl. Incl. 279 0.321 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 123 Table Q5: Results for Main Effect (H1) - Study 4 Predictor Variables Political Affiliation Political Affiliation2 (H1) Service Failure Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Restaurant Brand Fixed Effects Restaurant Brand Perceived Political Affiliation Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.241*** (0.040) 0.118*** (0.026) -0.349*** (0.091) 0.508*** (0.042) -0.011* (0.006) 0.115 (0.105) 0.058 (0.185) 2.150*** (0.562) Incl. Incl. Incl. Incl. 603 0.540 Brand Advocacy (2) 0.284*** (0.044) 0.104*** (0.028) -0.234** (0.097) 0.439*** (0.047) -0.005 (0.006) 0.175 (0.108) 0.045 (0.210) 2.690*** (1.003) Incl. Incl. Incl. Incl. 603 0.494 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 124 Table Q6: Results for Main Effect (H1) - Study 5 (Hotel Controlled Context) Predictor Variables Political Affiliation Political Affiliation2 (H1) Political Misalignment Political Alignment Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.111*** (0.040) 0.074*** (0.024) 0.075 (0.179) -1.024*** (0.241) 0.762*** (0.055) -0.003 (0.006) 0.078 (0.149) 0.033 (0.216) 1.231** (0.613) Incl. Incl. Incl. 299 0.681 Brand Advocacy (2) 0.100** (0.041) 0.113*** (0.024) -0.106 (0.193) -0.898*** (0.247) 0.753*** (0.059) -0.005 (0.006) -0.171 (0.158) -0.029 (0.214) 1.395** (0.623) Incl. Incl. Incl. 299 0.657 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 125 Table Q7: Results for Main Effect (H1) - Study 5W (Fast Food Recall) Predictor Variables Political Affiliation Political Affiliation2 (H1) Political Misalignment Political Alignment Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.029 (0.034) 0.045** (0.018) 0.410*** (0.142) -0.927*** (0.189) 0.372*** (0.055) 0.001 (0.007) -0.142 (0.133) -0.121 (0.183) 3.576*** (0.799) Incl. Incl. Incl. 448 0.310 Advocacy (2) 0.044 (0.036) 0.050** (0.020) 0.419*** (0.159) -0.836*** (0.204) 0.421*** (0.060) 0.003 (0.007) -0.244* (0.146) -0.150 (0.191) 3.889*** (0.823) Incl. Incl. Incl. 448 0.307 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 126 APPENDIX R: POLITICAL INTENSITY AND POLITICAL ORIENTATION Additional analysis – studies 1a 2a, 2b, 2bw, 3, 4, 5, and 5w. Thus far, we have provided evidence regarding the effects of customers’ political intensity, while also taking into account the direction of their political affiliation or ideology. In this analysis, we estimate regression models (separately for Studies 1A, 2A, 2B, 2BW, 3, 4, 5, and 5W) of customer loyalty and brand advocacy as a function of political intensity, political affiliation, and the interaction between the two while continuing to include all other covariates and fixed effects and using similar estimations approaches from respective studies. To construct the political affiliation variable, we classify customers as either left-wing (- 1) or right-wing (+1), while assigning independents a value of 0 (Zwicker, Prooijen, and Krouwel 2020). Detailed results from these analyses are presented in Tables S1-S7. While several of the coefficients of the interaction terms are positive, that is, indicating potentially stronger effects of political intensity among conservatives (Republicans), most are statistically non-significant (p>.05). 127 Table R1: Interaction of Political Intensity and Political Affiliation - Study 1A Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation Internal Threat (Product/Service Failure) External Threat (Product Market Fluidity) Customer Satisfaction Price Perception Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Industry Fixed Effects Brand Fixed Effects Year Fixed Effects Month Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.106*** (0.013) 0.040** (0.019) 0.001 (0.016) -1.008*** (0.045) -0.007 (0.011) 0.276*** (0.013) 0.261*** (0.017) -0.001 (0.001) 0.019 (0.029) -0.242*** (0.036) 7.436*** (0.500) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 0.247 Brand Advocacy (2) 0.113*** (0.013) 0.031 (0.020) -0.015 (0.017) -1.132*** (0.047) -0.003 (0.011) 0.276*** (0.013) 0.288*** (0.018) -0.003*** (0.001) -0.000 (0.030) -0.245*** (0.037) 7.873*** (0.466) Incl. Incl. Incl. Incl. Incl. Incl. 22,444 0.274 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Intensity and external threat (product market fluidity) variables were mean-centered in the interaction models before creating the interactions (2 and 4). Most Firm, Industry, and Year-Month fixed effects are statistically significant at least at p < .05 across all models. 128 Table R2: Interaction of Political Intensity and Political Affiliation - Study 2A Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects Brand’s Perceived Political Affiliation Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.103** (0.043) 0.114* (0.063) 0.064 (0.077) 0.091*** (0.035) 0.002 (0.004) -0.003 (0.109) -0.288 (0.279) 6.788*** (0.458) Incl. Incl. Incl. Incl. 303 0.238 Brand Advocacy (2) 0.136*** (0.035) 0.061 (0.063) -0.009 (0.068) 0.102*** (0.033) 0.003 (0.003) -0.062 (0.088) 0.143 (0.243) 5.835*** (0.407) Incl. Incl. Incl. Incl. 303 0.357 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 129 Table R3: Interaction of Political Intensity and Political Affiliation - Study 2B Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.271*** (0.075) 0.124 (0.095) 0.011 (0.085) 0.224*** (0.065) 0.001 (0.007) 0.116 (0.146) -0.119 (0.217) 3.889*** (0.675) Incl. Incl. Incl. 195 0.266 Brand Advocacy (2) 0.164** (0.076) 0.052 (0.102) -0.020 (0.091) 0.192*** (0.061) -0.005 (0.006) -0.212 (0.138) -0.367** (0.181) 4.044*** (0.832) Incl. Incl. Incl. 195 0.240 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 130 Table R4: Interaction of Political Intensity and Political Affiliation - Study 2BW Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation Attitude Towards the Brand Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand’s Political Affiliation Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.310*** (0.106) 0.194 (0.125) -0.057 (0.096) 0.197*** (0.057) 0.035 (0.063) 0.001 (0.007) -0.441*** (0.140) -0.200 (0.143) 4.821*** (0.613) Incl. Incl. Incl. Incl. 294 0.368 Brand Advocacy (2) 0.274*** (0.088) 0.210** (0.098) 0.059 (0.083) 0.170*** (0.053) 0.008 (0.059) 0.001 (0.006) 0.010 (0.105) -0.094 (0.166) 4.854*** (0.509) Incl. Incl. Incl. Incl. 294 0.364 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 131 Table R5: Interaction of Political Intensity and Political Affiliation – Study 3 Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation External Threat (Competitor’s Price Promotion) Customer Satisfaction Attitude Towards the Company Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Perceived Political Affiliation Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.430*** (0.078) -0.099 (0.091) 0.065 (0.065) -0.007 (0.091) 0.040 (0.026) 0.178*** (0.040) -0.003 (0.006) 0.075 (0.100) -0.021 (0.138) -0.006 (0.936) Incl. Incl. Incl. Incl. 279 0.461 Brand Advocacy (2) 0.359*** (0.084) 0.040 (0.092) 0.027 (0.075) 0.067 (0.118) 0.067** (0.031) 0.238*** (0.050) -0.009 (0.010) 0.141 (0.131) 0.169 (0.200) -0.990 (1.182) Incl. Incl. Incl. Incl. 279 0.336 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 132 Table R6: Interaction of Political Intensity and Political Affiliation – Study 5 Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation Political Misalignment Political Alignment Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.291*** (0.080) 0.147 (0.099) 0.186** (0.092) -0.028 (0.182) -1.134*** (0.245) 0.755*** (0.055) -0.004 (0.006) 0.097 (0.149) 0.007 (0.214) 1.293** (0.537) Incl. Incl. Incl. 299 0.686 Brand Advocacy (2) 0.436*** (0.080) 0.098 (0.101) 0.222** (0.100) -0.250 (0.195) -1.055*** (0.251) 0.742*** (0.058) -0.005 (0.006) -0.141 (0.155) -0.080 (0.211) 1.782*** (0.542) Incl. Incl. Incl. 299 0.668 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 133 Table R7: Interaction of Political Intensity and Political Affiliation - Study 5W Predictor Variables Political Intensity Political Intensity*Political Affiliation Political Affiliation Political Misalignment Political Alignment Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty (1) 0.187*** (0.062) 0.153* (0.086) 0.003 (0.087) -0.996*** (0.190) 0.339** (0.145) 0.370*** (0.055) 0.001 (0.007) -0.135 (0.132) -0.106 (0.183) 3.797*** (0.806) Incl. Incl. Incl. 448 0.316 Brand Advocacy (2) 0.212*** (0.069) 0.193** (0.087) 0.014 (0.091) -0.918*** (0.207) 0.334** (0.164) 0.418*** (0.060) 0.003 (0.007) -0.236 (0.144) -0.132 (0.193) 4.137*** (0.829) Incl. Incl. Incl. 448 0.315 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Affiliation was mean-centered before creating its quadratic term. 134 APPENDIX S: HIGH COMPLEXITY CONTEXTS Across the studies (1A, 1B, 2A, 2B, 2BW, 3, 4, 5, and 5W), we test the effects of political intensity on customer loyalty and brand advocacy across a variety of low complexity (e.g., soft drinks, Study 3) and limited complexity (e.g., cellular phones, Study 5W) product contexts. In this study, using a controlled context of hypothetical brands, we experimentally manipulate context complexity: low (cookies) and high (hybrid SUVs) complexity to assess the sensitivity of the effects to such contexts. Design, Procedure, and Manipulation of Context Complexity. We recruited a total of 198 U.S.-based respondents from Amazon Mechanical Turk who were provided with appropriate compensation. Participants were randomly assigned to one of the two conditions: 1) a low complexity context, wherein they read descriptions of three new cookie brands and selected one to purchase (see Table S1 for details); 2) a high complexity context, wherein they read descriptions of three hybrid SUVs and selected one to purchase (see Table S2 for details). Participants then answered a sequence of questions about other constructs of interest to this study, including their likelihood to repurchase the same brand of cookies or hybrid SUV in the future. Measures. We adopt the same political party affiliation-based measure of political intensity and control variables (satisfaction, gender, education, income, race, and age) as used in Study 2A and the same measures of customer loyalty and brand advocacy as used in Study 4. Table S3 provides correlations among all the variables included in Study 6W along with descriptive statistics. Estimation and Results. For the manipulation check, participants were asked whether they agreed (or disagreed) that deciding which cookies (hybrid car) to buy was an expensive and 135 risky decision (1= “strongly disagree”, 7 = “strongly agree”). Prior research shows that high- complexity contexts typically concern more expensive products/services and involve higher risk (cf. Tanner and Raymond 2010). As intended, in the low complexity (cookies) condition (M = 4.71), participants evaluated their decisions as less expensive and risky than those in the high complexity condition (hybrid SUV, M = 5.34, t-test of difference = -0.63, p < .021). Adopting the same estimation procedure as in other studies, in support of H1, the results (Table S4) indicate that political intensity has a strong positive effect on customer loyalty (.272, p < .01) and on brand advocacy (.338, p < .01). However, the interaction models show that the effects of political intensity do not differ significantly in high-complexity (versus low- complexity) context for both customer loyalty (-.029, p >.10) and brand advocacy (.020, p>.10). 136 Table S1: Cookie (Low Complexity Product Category Condition) Brand Options and Their Descriptions – Study 6W Instructions Option 1 Option 2 Option 3 Imagine that you are shopping at a grocery store and during the shopping process, you feel like eating cookies. You find three options on the shelf. NutriBite Cookies are designed for those who prioritize health but still crave a delicious snack. These vegan-friendly cookies are made with organic ingredients like quinoa flour, almond butter, and dark chocolate chips. They are a good source of plant-based protein and dietary fiber, making them not just tasty but also nourishing. DoubleDelight is the ultimate cookie for chocolate aficionados. Each cookie is loaded with both dark and milk chocolate chunks, offering a deep, rich chocolate experience in every bite. The soft, moist center combined with a slightly crunchy exterior makes these cookies a delectable treat for any time of day. Caramel Crave cookies are a dream come true for those with a sweet tooth. These butter cookies feature a rich, buttery base filled with gooey caramel and a pinch of sea salt for a perfect balance of flavors. Their indulgent, chewy texture makes them an irresistible treat for any caramel lover. Post-Choice Scenario Please read the description of your experience with the car carefully: Imagine that it has been a while since you purchased “chosen cookie brand.” The cookies tasted as expected. 137 Table S2: Hybrid Cars (High Complexity Product Category Condition) Brand Options and Their Descriptions – Study 6W Instructions Option 1 Option 2 Option 3 Imagine that you are given the opportunity to purchase a brand-new hybrid SUV. Read the descriptions below about the three options to help you select one to purchase. Fulton Motors Ridge Runner: A sleek and powerful SUV, it combines luxury with performance. Equipped with a high-performance engine and state-of-the-art suspension, the Ridge Runner delivers a smooth ride even on the toughest roads. Its elegant design and high-tech features make it a top choice for those who seek comfort and style in their adventures. Zenith Autoworks Horizon Hiker: The Horizon Hiker is an eco-friendly SUV, offering a perfect blend of efficiency and power. With its hybrid engine, it ensures lower emissions without compromising on performance. The spacious, tech-loaded interior is designed for long journeys, providing a comfortable and connected experience for all passengers. Emagine Automotive Wanderer: This SUV stands out with its unique design and advanced safety features. Aimed at providing a secure and reliable ride, it comes with innovative driver-assist technologies and a reinforced structure for enhanced protection. The Wanderer is ideal for families, offering ample space, and a smooth, quiet ride. Post-Choice Scenario Please read the description of your experience with the car carefully: Imagine that it has been a few years since you purchased your hybrid SUV. The car worked as expected and you are ready for a new SUV. 138 Table S3: Correlations and Descriptive Statistics - Study 6W Variables (1) Customer Loyalty (2) Brand Advocacy (3) Political Intensity (4) Political Affiliation (5) Context Complexity (6) Customer Satisfaction (7) Gender (8) Education (9) Income (10) Age (11) Race Mean S. D. (1) 1 0.573 0.356 0.005 0.152 0.066 0.103 -0.146 -0.033 0.061 0.148 6.056 0.856 (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 1 0.366 0.036 0.113 0.067 0.071 -0.132 -0.059 0.019 0.152 5.869 0.925 1 0.048 -0.091 0.177 0.191 0.009 -0.219 0.086 0.023 2.015 1.147 1 0.023 -0.069 0.002 0.003 -0.08 -0.114 -0.015 0.510 0.501 1 -0.11 -0.115 0.181 -0.26 -0.065 -0.015 4.136 2.319 1 0.154 -0.161 -0.059 0.102 0.084 5.162 1.812 1 -0.134 0.215 -0.025 -0.012 33.939 6.155 1 -0.037 0.028 -0.01 0.682 0.467 1 0.139 -0.068 0.939 0.239 1 0.133 4.955 0.833 1 6.283 2.359 Notes: All correlations above 0.12 are statistically significant at p <0.05 or less. N = 198. 139 Table S4: High versus Low Complexity Contexts – Study 6W Predictor Variables Political Intensity (H1) Political Intensity *Context Complexity (High=1, Low-0) Context Complexity (High=1, Low-0) Political Affiliation Customer Satisfaction Age Gender Race Intercept Education Level Fixed Effects Income Level Fixed Effects Brand Perceived Political Affiliation Fixed Effects Brand Fixed Effects N (Customers) R-squared Customer Loyalty Brand Advocacy Main Effect Model (1) 0.272*** (0.051) -- -0.044 (0.114) 0.082*** (0.027) -0.003 (0.030) -0.002 (0.009) -0.310** (0.147) 0.512* (0.263) 5.685*** (0.625) Incl. Incl. Incl. Incl. 198 0.296 Interaction Model (2) 0.286*** -0.029 (0.095) (0.141) -0.046 (0.114) 0.081*** (0.027) -0.003 (0.030) -0.002 (0.009) -0.311** (0.148) 0.505* (0.265) 5.678*** (0.630) Incl. Incl. Incl. Incl. 198 0.297 Main Effect Model (3) 0.338*** (0.064) -- 0.020 (0.127) 0.039 (0.031) 0.000 (0.037) -0.008 (0.012) -0.428*** (0.156) 0.380 (0.257) 6.199*** (0.658) Incl. Incl. Incl. Incl. 198 0.312 Interaction Model (4) 0.328*** 0.020 (0.112) (0.189) 0.021 (0.127) 0.039 (0.031) -0.000 (0.037) -0.008 (0.012) -0.427*** (0.156) 0.385 (0.259) 6.205*** (0.664) Incl. Incl. Incl. Incl. 198 0.312 Notes: Indicates statistical significance at ***p < .01; **p < .05; *p < .10. Heteroskedasticity robust standard errors (SE) in parentheses. Political Intensity was mean-centered in the interaction models (2 and 4) before creating the interactions. 140