Abstract The present research explores how financial status, which influences consumers’ expectations about how companies will treat them, affects consumers’ perceptions and assessments of products that have been given anthropomorphic features by companies. Studies 1 and 2 showed that participants with higher financial status expect more favorable treatment from a humanized entity (e.g., “a self-driving car would prioritize the well-being of the rich over others”). The results of study 3 indicate that participants with higher perceived financial status both afforded greater agency to humanized products and liked these products better than did participants with lower perceived financial status. These effects were mediated by commercial treatment expectations, when we controlled for perceived control and self-efficacy. Further confirming the role of treatment expectations, when participants believed people with low financial status would be treated better than those with high financial status, we observed the reverse pattern (study 4). Lastly, study 5 replicated the effect using a measured, not manipulated, variable of financial status. Findings support the view that effective anthropomorphism requires marketers to take into account consumers’ motivation to interpret a target with humanlike features as having positive agency, which results from treatment expectations. anthropomorphism, financial status, agency perception, product liking Marketers often have the option to give their products and services humanlike features. For example, automobile makers can configure the front end of their cars to resemble a human face, as Google has purportedly done with the prototype model of its self-driving car (see also Landwehr, McGill, and Hermann 2011). The Google car has a pert “nose” for a forward sensor, wide-open “eyes” for headlights, and a smiley “mouth” tilted slightly upward for a bumper line. Other efforts to anthropomorphize offerings go beyond giving products physical features that resemble those of people. For example, Amazon has imbued Echo, a cylindrically shaped interactive speaker, with the human name Alexa, a female voice that employs familiar human intonation, and some quirky “personality” traits. Rapid technological development in recent years has brought more anthropomorphic “things” (e.g., self-driving cars, virtual assistants) into people’s everyday lives. People interact “socially” with these anthropomorphic things; rather than just exerting control over them, people directly interact with the anthropomorphized entities and expect responses from them. As humans interact with anthropomorphic entities more often, it is important to understand the bidirectional dynamic between the two: How and why do people treat anthropomorphic entities in certain ways, and how do people expect to be treated by them? Most prior anthropomorphism studies provide limited insight into the latter question. In particular, past studies have rarely provided opportunities for real interaction between people and entities with humanlike characteristics. A consumer-product relationship with these anthropomorphized objects (i.e., a round-shaped cookie with eyes on it [Hur, Koo, and Hofmann 2015], an orange juice bottle sitting on a beach chair and wearing a hat [Puzakova, Kwak, and Rocereto 2013]) is usually unidirectional: the objects are not expected to react to users. In these contexts, people would likely neither consider how they might appear to the anthropomorphized objects nor have worries that the objects could treat them well or poorly. By contrast, in an interactive relationship, where consumers are more dependent on the actions of the humanlike targets, consumers’ treatment expectations could exert greater influence over their assessment of the anthropomorphized products. The present research is intended to address this question of how consumers’ treatment experiences in the real world affect the way people expect humanized entities to treat them, and consequently influence consumers’ evaluations of these entities. Specifically, we focus on the role of financial status as a practical and prevalent factor in the formation of people’s expectations about how social entities would treat them in a commercial context. Financial standing influences customers’ expectations about their relative attractiveness to businesses and their likely treatment by the agents of these businesses. Because this variable may be readily observed or measured, the present research could provide marketing practitioners with directly actionable insights. In addition, the present work extends theory by revealing not only different evaluations of products that marketers have given human features, but also, moving back a step, differences in consumers’ willingness to see these products as somewhat human depending on their financial standing. To the extent that rich and poor consumers have different expectations about how a company and its representatives will treat them, they may be differently inclined to see products that marketers have given human features as possessing human agency. We acknowledge the possibility that people who feel financially deprived may be more motivated to seek social resources (as a substitute for financial resources), and consequently to afford higher agency to humanlike targets and to exhibit positive attitudes toward them. Financial resources and social resources often seem interchangeable. Recent research provides evidence supporting the compensatory process between the two resources—specifically, that lack of one of the resources heightens the desire for the other type of resource, and once one of the resources is replenished the desire for the other is attenuated. Money motivates people to socialize less (Mogilner 2010) and even mitigates psychological pain caused by social exclusion (Zhou, Vohs, and Baumeister 2009). Additionally, people who experienced social exclusion, and thus temporarily lacked social resources, show higher desire for money (Duclos, Wan, and Jiang 2013). However, a key differentiator of the current research is that we utilize a product, which is provided by a company with a profit motivation, as a potential social target. A profit motivation may lead firms (and the people employed by them) to display preferential treatment toward consumers with the ability to spend more (Mende et al. 2015; Mittal, Sarkees, and Murshed 2008). As a consequence, people with greater financial resources may eagerly perceive products with human features as additional agents in their environment, there to do their bidding and to treat them well, whereas people with limited financial resources may be reluctant to afford these products agency. Therefore, we expect that people who perceive themselves as high in financial resources would be more willing to anthropomorphize products and to like products with humanlike characteristics more. Hence, just as lonely people are more likely to anthropomorphize objects in a social setting because they expect the objects to serve as companions (Epley et al. 2008), affluent people might be more willing to anthropomorphize product offerings in a commercial setting because they expect them to serve as helpful minions. Consumers who perceive themselves as low in financial resources, by contrast, may be wary of marketer-anthropomorphized products, which might not be fully on their side, and thus feel less willing to afford them humanity and like them less. We develop these hypotheses in the following sections and test them in five studies. COMMERCIAL TREATMENT EXPECTATIONS Financial status may especially influence consumers’ expectations regarding the behavior of companies and their representatives. These expectations likely develop as a consequence of the prevalent use of preferential treatment in marketing, which gives selective customers elevated status recognition and/or augmented products and services (Lacey, Suh, and Morgan 2007). From the perspective of a company striving to maximize its profit, all customers should not be treated as equal: without differentiating, a company might fail to maximize its “bang for the marketing buck,” oversatisfying (overserving) less valuable customers and undersatisfying (underserving) more profitable customers. Thus, firms segment their customers based on level of loyalty (which, in the business world, is almost a synonym for spending) and bestow a corresponding class on their customer society. In a well-defined consumption hierarchy, each class of customers is provided with specific benefits that differ from tier to tier. Brands also actively configure and enact the consumer hierarchy in their service encounters by signaling the types of consumers that are welcomed and tailoring service interactions according to their customers in ways that reflect classed treatment expectations (Dion and Borraz 2017). This practice in service interactions further reproduces class dispositions, causing consumers to internalize the inequalities of entitlement. Consumers commonly observe or experience preferential treatment—for example, extra benefits—offered to qualified people (see also Ward and Dahl 2014). Frequent flyers with their “Executive Platinum Memberships”—of course, obtained by spending heavily and contributing to the airline’s bottom line—bypass long check-in lines at the airport and receive a special luggage tag that guarantees priority baggage handling. Through repeated observations and experiences, consumers with different financial standing may develop different expectations regarding the behavior of companies and their representatives, contingent upon their customer value. Consumers who have recognized their high-priority status are more likely to think they deserve commensurate treatment and to believe they are entitled to additional effort and special attention from companies in return for their (potential) contribution (Bolton, Kannan, and Bramlett 2000; Boyd and Helms 2005; Lacey et al. 2007). Further, recent work by Reczek, Haws, and Summers (2014) suggests consumers’ greater feeling of deservingness (because they perceive themselves, as extraordinary customers, to have invested more effort and money in a firm than other ordinary customers) extends to a (mis)belief that they are more likely to receive rewards from the firm even when the rewards are independent of loyalty (i.e., a random draw). By contrast, consumers who lack financial standing may be all too familiar with dismissive, even rude treatment from company representatives who know that those lacking in means offer little financial benefit to companies’ profitability. Based on these findings, the present research examines how such beliefs, which tie financial status to treatment expectations, affect consumers’ motivated perceptions and assessments of products to which companies have given anthropomorphic features. ANTHROPOMORPHISM Anthropomorphism is the attribution of humanlike forms, behavioral characteristics, or mental states to nonhuman objects, such as products, animals, supernatural phenomena, or even abstract concepts. For the most part, anthropomorphism has been thought to enhance favorable attitudes, because it enables people to feel a sense of efficacy with respect to nonhuman entities or to form a social connection with them (Epley, Waytz, and Cacioppo 2007, 2008), satisfying states that enhance positive evaluation of the anthropomorphized entity. For example, people who have anthropomorphic beliefs about objects showed an increased sense of attachment and decreased willingness to replace them (Chandler and Schwarz 2010), and people manifested increased compliance with anthropomorphized social causes (Ahn, Kim, and Aggarwal 2014). Also, socially excluded consumers, who are motivated to establish a social relationship, exhibit greater preferences for anthropomorphized brands than for objectified brands (Chen, Wan, and Levy 2017). Marketers appear aware of these tendencies, strategically imbuing products, brands, and companies with human traits. Indeed, humanlike products, such as differently sized bottles resembling a human family or analog watches set to 10:10 to appear like a smiling face, have scored considerable success in garnering increased attention and liking from consumers (Aggarwal and McGill 2007; Labroo, Dhar, and Schwarz 2008). However, recent research on the effects of anthropomorphism suggests that anthropomorphism does not necessarily lead consumers to positively evaluate anthropomorphized objects. For example, anthropomorphism enhances appeal and liking only when the consumer perceives the object to have positive qualities, such as a helpful functionality, positive outcomes, and congruency between the consumer’s self-concept and brand image (Burgoon et al. 2000; Morewedge 2009; Puzakova, Kwak, and Rocereto 2009). Besides the perception of the anthropomorphized object itself, another important factor makes the anthropomorphism effect more complex: people’s social beliefs. When people encounter objects in anthropomorphic terms, their social worldview appears to extend to the inanimate world. Because people tend to apply their social attitudes or social beliefs to the anthropomorphized entity, researchers have observed different consumer reactions to products depending on the attitudes consumers hold toward other humans. For example, people tend to treat anthropomorphized objects similarly to how they treat other people (e.g., to be less likely to throw away their outdated car when the car is anthropomorphized as an “old and sick friend,” Chandler and Schwarz 2010). People also apply their implicit personality theories about the ability of people to change over time to anthropomorphized products that perform poorly (e.g., to judge anthropomorphized products less favorably after negative publicity compared to nonanthropomorphized products when endorsing an entity vs. an incremental theory of personality, Puzakova, Kwak, and Rocereto 2013). Based on this prior research, we posit that consumers’ social expectations about how companies will treat them will affect their evaluations of products that have been given humanlike features by marketers as well as their motivations to see the nonhuman products as possessing agency, especially an ability to treat the consumers well. (MARKETER-) DESIGNED ANTHROPOMORPHISM VERSUS (CONSUMER-) PERCEIVED AGENCY Our research goes beyond prior research by distinguishing between marketer efforts to present a product in anthropomorphic terms and consumer motivation to afford the product agency to think and act on its own. Marketers may depict products as possessing physical features that appear human, such as facial features and body shapes (Aggarwal and McGill 2007), which most customers would readily discern. However, these efforts by marketers to design humanlike products can be independent from consumers’ attribution of humanlike mental states to those products, depending on consumers’ motivation. Therefore, here we use designed anthropomorphism (DA) to refer to those efforts by marketers to signal humanness with explicit humanlike features in products, whereas perceived agency (PA) refers to consumers’ interpretation of the products with DA as having agency. In the dehumanization literature, which is a conceptual inversion of anthropomorphism, people tend to dehumanize psychologically distant and dissimilar other people with whom they don’t want to be related, such as drug addicts or homeless people, primarily by denigrating their mental capabilities (Waytz, Epley, and Cacioppo 2010), not their human appearance. In a similar way, perceiving a target “being” as a cognitive agent with the ability to think, feel, and act on its own requires an independent, one-step-further progression of anthropomorphism beyond merely noting physical features that humans usually have (e.g., seeing a face on a power outlet). This view accords with that of Guthrie (1993), who proposes three forms of anthropomorphism: partial, literal, and accidental. According to Guthrie, partial anthropomorphism occurs when people see some human traits in objects or events but do not believe the target is fully, literally human. Hence, in a commercial context, marketers can signal some degree of humanity in a product by giving it “cartoonish” features that align with those of people in the hope that customers will see the product in partially human terms. Whether consumers accept or reject the given signals and to what extent they stretch the signals to interpret how partially human the product is, however, is up to them. In particular, when a humanlike entity seems to have a higher cognitive ability to make a set of decisions contingent on situations, people’s treatment expectations—how a social entity would behave toward me—play an important role in their motivation to see the entity as a fully volitional human. Because people desire positive social experiences (Baumeister and Leary 1995), people with a favorable commercial-treatment expectation might be more likely to afford further humanity (agency) to products given human features by marketers, whereas people who expect unfavorable treatment in a commercial context might show decreased motivation to further anthropomorphize such products. In other words, consumers with high financial status, who expect to be treated well by companies and their agents, would be motivated to extend the expectation to humanlike entities, and engage in further anthropomorphism by granting the entities agency to treat them well. In contrast, consumers with low financial status do not expect to receive good treatment from companies and their agents, and so are disinclined to assign agency to the entities with humanlike features. Instead of admitting potential harm into their lives by assigning agency to the humanlike products, these consumers would be motivated not to endow the entities agency in the first place. OVERVIEW OF RESEARCH Integrating these lines of research on financial status, commercial-treatment expectations, and motivated anthropomorphism, we conducted five studies to examine the role of financial status on participants’ evaluations of humanized products. We predict that when consumers see products with humanlike features, consumers with high financial status (vs. those with low financial status) will expect more favorable treatment from these humanlike entities, which leads to higher motivation to give agency, and positively evaluate the entities. More specifically, people with higher financial standing are, in general, treated well by companies and salespeople. Thus, they will likely also expect favorable treatment from the humanized agents, viewing the agents as akin to well-trained employees reflecting the company’s strategy, whereas people with low financial standing will not share these positive beliefs about humanized entities. The asymmetric beliefs about treatment in commercial contexts depending on people’s financial status would influence their motivation to further afford agency to targets with superficial human features. People who perceive themselves as better off financially will be more inclined to confer agency (higher perceived agency [PA]) onto the target with anthropomorphic features (high in designed anthropomorphism [DA]), as the anthropomorphism adds another positive actor to their world. By contrast, people with low financial standing, thus expecting less favorable treatment from social entities, would be motivated to interpret the humanlike features of high-DA products as nothing more than cartoonish design, making the targets powerless with no agency. Further, we predict that this higher motivated perception of agency displayed by consumers with high financial status, resulting from their positive expectation that social entities will treat them well, will lead to a more positive assessment of high-DA products among the rich than the poor. On the other hand, past research predicts no systematic differences regarding products without explicit humanlike features. That is, the rich and the poor would not differ in their evaluations of products that have not been designed by the marketer to seem human. In studies 1 and 2, we first show that people’s financial status (either measured in study 1 or manipulated in study 2) influences their beliefs about how a humanlike product would treat them (e.g., a self-driving car in a moral dilemma situation). Study 3 further explores how different expectations about social entities influence consumers’ motivation to further anthropomorphize and afford agency to products as well as consumers’ evaluation of these products. We investigate the role of treatment expectations as an underlying mechanism, and also rule out possible alternative explanations. Study 4 extends our theorizing about commercial treatment expectations by showing that the interactive effect of financial status and designed anthropomorphism is reversed when people’s expectations about the relationship between financial standing and commercial treatment is disrupted. In particular, we demonstrate that preference for humanized products depending on financial status is reversed when consumers expect commercial agents to treat the poor more favorably. Finally, in study 5, we replicate the interactive effect of financial status and designed anthropomorphism on consumers’ product evaluations using their actual financial status. PILOT SURVEY: FINANCIAL STATUS AND COMMERCIAL TREATMENT EXPECTATIONS Our hypotheses hinge on the assumption that people hold a lay belief that consumers with greater financial resources get better treatment in commercial contexts than do consumers with fewer financial resources. To check this assumption, we assessed people’s beliefs about the typical relationship between financial status and commercial treatment. In a pilot survey (N = 55, 31 male, Mage = 35.2), participants rated the degree to which they agreed with the following statements using a seven-point scale (α = .79, –3 = Strongly Disagree, 3 = Strongly Agree): (1) “I think salespeople tend to treat wealthier customers better”; (2) “I think salespeople tend to show more favorable attitudes toward wealthier customers”; (3) “I think companies tend to train their employees to treat wealthier customers better”; (4) “I think companies are more concerned about keeping wealthier customers happy than those with less money to spend.” In addition, participants reported their need to belong by responding to 10 statements on the same seven-point agree-disagree scale (α = .93; Leary et al. 2013). A sample item for the need-to-belong measure is “I do not like being alone.” They also reported their perceived social power over others in general, using the Sense of Power Scale (α = .90; Anderson, John, and Keltner 2012). A sample item among the eight includes “I can get others to do what I want.” At the end of the survey, participants reported their gender, age, and annual income, the last of which was used as our measure of financial status. As we expected, results demonstrated that across financial statuses, consumers share the belief that companies and sales personnel treat customers in accordance with their level of wealth (one-tail t-test compared to midpoint (0), M = 1.40, SD = .99, t(54) = 10.47, p < .001). Also, individuals did not exhibit any differences depending on their financial status either in their sense of power (b = .03, p = .76) or need to belong (b = –.04, p = .70). These results, along with the prior literature on preferential treatment, suggest that financial status could be a general as well as a practical factor uniquely influencing people’s commercial treatment expectations, without changing their social desires or power perceptions. In the following studies, we experimentally manipulated or measured financial status, and tested whether financial standing influences people’s assessment of humanized products and whether the difference is explained by people’s treatment expectations in a commercial context. STUDY 1 Our primary objective in study 1 was to demonstrate the basic effect of financial status on consumers’ treatment expectations from a product with humanlike features. In this study, we investigate whether consumers’ financial status influences their willingness to disclose their personal financial information to a humanlike product. The logic underlying this study is that if people believe consumers with greater means will get better treatment, wealthier consumers will be more willing to provide their financial information when a service involves a humanlike entity (vs. does not involve a humanlike entity). More specifically, we predict that high-financial-status people, compared to those with low financial status, would be more willing to let a humanlike agent know who they are, expecting more favorable treatment in accordance with their higher level of wealth. However, we predict that for personal information that is less relevant to financial standing (e.g., a shopping wish list), and therefore, less influential to people’s treatment expectations from an entity, the relationship between financial status and willingness to provide the information would be weaker. When people are interacting with a low-DA product, on the other hand, we predict that financial status will not influence people’s motivation to provide their personal information because people do not expect the object to act differently depending on its knowledge about customers. Method We recruited 208 participants in the United States (97 male, Mage = 35.82) from Amazon Mechanical Turk (MTurk) who received $.40 in exchange for completing the study. Participants first recalled how much they currently have in their checking account and reported how easy it was to recall the balance. The first question was used as a measured variable for an individual’s financial status, and the second one was an unrelated question to introduce the concept of a product they were to read about. Participants then read a description of a fictitious product, called Balance, offering financial management services that help track an account balance each month and provide sales and promotions information (see appendix A). After participants viewed general information about Balance, they were informed that they could get a more personalized service if they connected their bank accounts and their shopping wish list with Balance. To manipulate the level of designed anthropomorphism, we created a service agent, Beezy, in the high-DA condition, and “he” explained how he could offer more personalized services when users provide further personal information to him. In the low-DA condition, we introduced Balance Plus, which offered the same benefits when users provide their information. After reading the descriptions, participants rated their willingness to connect their bank accounts and their willingness to connect their wish list with the program. At the end of the survey, to control for confounding demographic features closely related to financial status, we had participants report their education level and employment status in addition to their gender and age. Results and Discussion We conducted regression analyses to examine the relationship between participants’ financial status and their willingness to provide their personal information depending on the level of DA displayed by the entity requesting the information. Because the balance in individuals’ bank accounts, a proxy for financial status, was positively skewed (M = $3,193.79, median = $700, skewness = 4.34, kurtosis = 22.67), we log-transformed the variable. Thus, predictors were participants’ log-transformed checking account balance, the level of DA of the product, their interaction, and a set of controlling variables including education and employment status. Confirming our prediction, we found a significant interaction between financial status and a product’s level of designed anthropomorphism (b = –.72, t(202) = –2.26, p = .02). When the product included a high-DA service agent, wealthier participants were more willing to connect their bank accounts, information that is highly diagnostic of their financial status (b = .53, t(202) = 2.26, p = .02). However, for the low-DA product, participants’ financial status did not influence their willingness to connect their bank accounts (b = –.19, t(202) = –.84, p = .40). The same regression analysis was conducted on participants’ willingness to connect their shopping wish list, which is less diagnostic of their financial status. As expected, participants’ willingness to connect their wish lists did not significantly vary based either on their financial status or on the interactive entity’s level of designed anthropomorphism (all ps >.15). Study 1 provides an initial demonstration that consumers extend their beliefs about commercial treatment from real human beings (exhibited in the pilot study) to treatment expectations from products with anthropomorphic features. Higher willingness among richer consumers to disclose their financial status only to the high-DA product provides evidence supporting the hypothesis that consumers expect to be treated differently, especially by social agents, depending on their financial standing. STUDY 2 In study 2, we manipulated people’s perceived financial status, instead of using a measured variable of financial status, and explored whether people’s perceived financial status impacts their expectations about how a product with anthropomorphic features would treat them. We asked participants how they would expect an automated car in which they were riding to respond in a situation in which the car had to choose between protecting them and protecting others. We predicted that participants who perceived themselves to have high financial status would expect the high-DA car to be more likely to keep them safe, whereas participants of low financial status would expect the high-DA car to protect others while allowing them to be hurt. Method Participants We recruited 215 participants in the United States (109 male, Mage = 36.76) from MTurk. Each participant received $.40 in exchange for their time. We excluded 23 participants due to extremely low (<$10,000/year) or extremely high income (>$200,000/year) that we expected would undermine the financial-status manipulation, or due to an attention-check failure. This exclusion criteria was used consistently in studies where we manipulated participants’ financial status. Design and Procedure The experiment employed a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) between-subjects design. The study consisted of two phases: the wealth-manipulation phase and the decision-making phase. In the wealth-manipulation phase, we used two questions in tandem to manipulate subjective wealth. The first question asked participants to imagine they had just won a lottery, and to describe what they would like to do or buy with the unexpected money. For participants who were being manipulated to feel rich, the winnings were $50, whereas for participants manipulated to feel poor, the winnings were $50,000. The second question asked participants to indicate their monthly income and expected savings in two years. We expected reporting their actual financial status right after imagining what they could do with $50,000 would make people feel financially deprived, but that reporting financial status after imagining themselves to have an additional $50 would not highlight any specific shortcoming in their present level of wealth. To make the contrast more salient, we provided different scales depending on the condition. In the rich condition, participants reported their monthly income and savings on two 11-point scales divided into $50 increments, from 1 ($0–$50) to 11 (more than $500). We expected participants would cluster in the upper ends of the scales, creating a sense of greater financial well-being. In the poor condition, participants reported their monthly income on an 11-point scale from 1 ($0–$10,000) to 11 (more than $30,000) and their savings on a scale from 1 ($0–$10,000) to 11 (more than $450,000). We expected participants presented with scales labeled in this way would cluster at the low ends and so feel relatively poor (Nelson and Morrison 2005; Schwarz 1999). As a manipulation check, we measured participants’ subjective financial status using the sum of four questions: (1) “How satisfied are you with your current personal financial status?” (2) “How satisfied are you with your current material possessions?” (3) “How would you rate your current financial position?” (4) “What would you expect your financial position to be 10 years from now?” using a 100-point scale. After the wealth-manipulation phase was complete, participants entered into the decision-making phase of the study. Participants were asked to imagine a moral-dilemma situation involving an autonomous car. In the high-DA condition, the car was humanized with a speech bubble saying, “Hi, I’m Jasper,” and the phrase “a self-driving car” was used to refer to it (see web appendix A). In the low-DA condition, the speech bubble was replaced by an oval-shaped text box stating, “Here is Jasper,” and the phrase “a driverless car” was used. The moral-dilemma scenario was as follows (see Bonnefon, Shariff and Rahwan 2016): Recently, you got a self-driving car [a driverless car] (a vehicle that is capable of detecting its surroundings and navigating without human input). One day, while you are driving alone, an unfortunate set of events causes the self-driving car [a driverless car] to head toward a crowd of 10 people crossing the road. Since it cannot stop in time, it might swerve into a barrier, which would avoid injuring 10 people. However, if this collision happens, this would injure you seriously. On the other hand, the car might keep its direction, which would keep you safe and intact. However, this would injure the 10 people seriously. After reading the scenario, participants reported on a 10-point scale what they thought the car would do in this situation (1 = Definitely steer into a barrier, 10 = Definitely keep its direction). They also used a 10-point scale to answer how quickly they thought the car would respond to this situation and determine what to do (1 = Almost instantly, 10 = After substantial deliberation/assessment), but this measure showed no significant difference between conditions; therefore, we do not discuss it further. Finally, at the end of the survey, participants answered demographic questions including their annual income, gender, and age. Results Manipulation Check Before the main study, we conducted an independent pretest (N = 44) on MTurk to verify that people are able to discern the difference in the DA in the stimuli regardless of their financial-status perception. After experiencing the same financial-status manipulation as the participants in the main study, participants read generic information about how an autonomous car works. This generic information contained the same anthropomorphism manipulation as in the main study. After reading the description, pretest participants used a seven-point scale (1 = Strongly Disagree, 7 = Strongly Agree) to rate the extent to which they found humanlike features in the product: (1) “I see some humanlike features in Jasper” and (2) “I see the marketers’ intention to design Jasper as resembling a human” (r = .45). A 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) ANOVA revealed only one significant main effect of DA: participants successfully discerned marketers’ anthropomorphic intentions between the high-DA and the low-DA car, regardless of their perceived financial status (Mhi = 5.40, Mlo = 4.59; F(1, 40) = 4.78, p = .03; all other ps >.60). Regarding the financial-status perception (α = .80) measured in the main study, the same 2 × 2 ANOVA revealed only a significant main effect of financial-status manipulation (Mrich = 53.12, Mpoor = 42.46; F(1, 188) = 15.86, p < .01; all other ps >.10), suggesting a successful manipulation. Decision Prediction We predicted that high financial-status perception would increase consumers’ expectations of positive treatment from a humanized product. We conducted a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) ANOVA to test this prediction. The result revealed a significant interaction (F(1, 188) = 7.11, p = .008; see figure 1). We found no other significant effects (all ps >.10). Specifically, participants in the rich condition predicted a high-DA car would be more likely to keep them safe even at the cost of injuring 10 pedestrians than did participants in the poor condition (Mrich hi = 4.36, SDrich hi = 2.99, Mpoor hi = 3.31, SDpoor hi = 2.80; F(1, 188) = 3.46, p = .06), whereas the pattern was reversed when the car was not heavily humanized, although these effects were both marginally significant (Mrich lo = 3.13, SDrich lo = 2.56, Mpoor lo = 4.22, SDpoor lo = 2.77; F(1, 188) = 3.66, p = .06). Also, in the other pair of contrasts, participants in the rich condition predicted the high-DA car would act in a more favorable way toward them than the low-DA car would (F(1, 188) = 4.54, p = .03). By contrast, participants with low financial-status perception thought the low-DA car would be more likely to keep them safe, although this effect was marginally significant (F(1, 188) = 2.66, p = .10). FIGURE 1 View largeDownload slide STUDY 2: THE EFFECT OF PERCEIVED FINANCIAL STATUS AND DESIGNED ANTHROPOMOMRPHISM ON CAR DECISION PREDICTION FIGURE 1 View largeDownload slide STUDY 2: THE EFFECT OF PERCEIVED FINANCIAL STATUS AND DESIGNED ANTHROPOMOMRPHISM ON CAR DECISION PREDICTION Discussion Study 2 provided additional evidence that people have different expectations about how a target with humanlike features would behave toward them depending on their financial-status perception. When a product is highly humanized, participants who felt rich predicted a more favorable treatment from it than those who felt poor did. The perceived rich participants also predicted the highly humanized product would prioritize their benefits over others’ to a greater extent than a product with limited humanlike characteristics would. These results provide an initial validation of our framework that social expectations contingent on subjective financial status moderate the anthropomorphism effect. (We also note with interest that several months after this study was conducted, Mercedes Benz, which targets upscale customers, reported its plans to program its self-driving cars to prioritize the driver and passengers over pedestrians; Morris 2016.) Interestingly, we also found that people with low (vs. high) financial status expected more favorable treatment when a product lacks explicit humanlike features, even though this effect was only marginally significant. We will further look at this unexpected effect in our following studies. STUDY 3 We conducted study 3 with two objectives in mind. First, based on the findings of studies 1 and 2, the current experiment was intended to further examine how different treatment expectations based on financial standing influence (1) the level of agency that consumers actually perceive products with anthropomorphic features to possess and (2) consumers’ assessment of those products. To test the role of treatment expectations as the underlying mechanism, we directly measured consumers’ beliefs about how a target would treat them in the given commercial context. We expected people with high perceived financial standing, who believe the entity would act on their behalf, to be more willing to embrace anthropomorphic features, further giving a high-DA product agency to think and act on its own. Furthermore, we expected such high-financial-status individuals to evaluate the high-DA product more positively than people with low perceived financial standing, who think the entity would not work on their behalf. On the other hand, for a low-DA product, we again explored the opposite possibility on the extended measures: the lack of human features could be interpreted as “no preferential treatment for the rich,” which would make people with high perceived financial status to be less willing to give agency to, and to like the low-DA product less than those with low perceived financial status. In addition, study 3 tested potential alternative explanations for the observed effect by examining other factors that might be correlated with financial status perception: perceived control and self-efficacy. Higher financial status may lead people to feel (illusionary) control over entities and prefer them accordingly. If people believe they can exert power over a target to make it offer a better outcome to them, they could feel more comfortable interacting with the target (and so be willing to give it agency) and be more likely to assess it positively. However, in our suggested framework, it is people’s different expectations about how they will be treated by social entities, independent from the degree of control they perceive themselves to have over interacting entities or over the situation in general, that drives different reactions to humanized targets depending on people’s financial standing. Therefore, the current experiment attempted to minimize participants’ subjective feelings that they could exert control over a target by removing the possibility of direct interaction between the participant and the target (while maintaining the possibility of the target product influencing the quality of the consumer’s experience). We also directly measured perceived control and general self-efficacy to statistically control the two variables. Method Participants We recruited 173 participants in the United States (86 male, Mage = 37.4) from MTurk to complete the study in exchange for $.50, but excluded 11 participants following the established exclusion criteria (see study 2). Design and Procedure The experiment employed a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) between-subjects design. The study was presented as a “Smartphone Application Survey” and consisted of two phases: the wealth-manipulation phase and the product-evaluation phase. We implemented a new financial-status manipulation by slightly modifying the scale of subjective socioeconomic status in Adler et al. (2000). Participants were given a graphical representation of a ladder with eight rungs following the instruction “Think of the ladder below as representing where people stand in the current society.” In the rich condition, where we intended to make participants feel rich, the bottom rung of the ladder was highlighted and labeled “People with the lowest income and least financial resource,” whereas in the poor condition, the top rung was highlighted and labeled “People with the highest income and most financial resource.” Depending on the financial-status-perception condition, participants were instructed to compare themselves with those people at the very bottom (in the rich condition) or top (in the poor condition) rung of the ladder. Additional instruction was provided stating, “These are people who are the worst (best) off, earn the lowest (highest) income, have the least (most) money, and least (most) material possessions. In particular, we’d like you to compare yourself to these people in terms of your own wealth, income, and material possessions.” After the wealth-manipulation phase, participants answered the same manipulation-check questions used in study 2. Participants then read a description of an online food delivery service and were told that the estimated time of arrival is determined either by a logistics expert, Mr. DeLight [high-DA condition] or by a logistics algorithm called the DeLight system [low-DA condition] (see appendix B). Participants were instructed that the delivery time is given to the company and relayed to the consumer, who therefore could not influence the anthropomorphic entity or the system directly, which we hoped would level the differences in perceived control due to the human features depicted. In contrast to the two previous studies, the target product was described from the third-person perspective in all conditions to control for any confounding factors due to the messenger perspective (Touré-Tillery and McGill 2015). After reading the description, participants reported their evaluations of the application. Specifically, they answered three questions: (1) “How much do you like it?” (1 = Not at all, 7 = Extremely); (2) “How much do you want to use it?” (1 = Not at all, 7 = Extremely); (3) “If this service is free, how likely are you to use it?” (1 = Definitely No, 7 = Definitely Yes). As a measure of perceived agency (PA), participants reported the extent to which the DeLight seemed to be a person, have its own intentions, have its own personality, and have free will using a seven-point scale anchored by 1 = Definitely No and 7 = Definitely Yes. For the treatment expectation measure, participants indicated the degree to which they agreed with the following four statements: (1) “Mr.DeLight [high-DA condition]/The DeLight [low-DA condition] system will act in my best interests”; (2) “Mr.DeLight/The DeLight system will act in the company’s best interests, not mine” (reverse-coded); (3) “Mr.DeLight/The DeLight system will act in a way that maximizes my satisfaction”; (4) “Mr.DeLight/The DeLight system will assign the fastest route for my order.” To test potential alternative explanations for the underlying process and control for confounding variables in financial-status perception, we collected measures for perceived control and general self-efficacy. For the control perception, participants responded to the following measures on a seven-point scale: (1) “I believe I can influence DeLight to prioritize my order”; (2) “I believe I can control DeLight to make my order delivered faster”; (3) “I believe I can change DeLight to perform better for me”; (4) “I believe I can exert influence over DeLight.” Participants also completed seven measures of general self-efficacy using a seven-point scale (Chen, Gully, and Eden 2001). Sample items included “I will be able to successfully overcome many challenges” and “In general, I think I can obtain outcomes that are important to me.” Finally, participants indicated their age, gender, and personal annual income for demographic information. Results Manipulation Check We conducted an independent pretest (N = 45) on MTurk to verify that people perceived the intended differences in the level of anthropomorphism of the stimuli, even after the financial-status manipulation. After the same wealth manipulation used in the main study, participants read a description of the same product either in the high- or low-DA manipulation, and rated the extent to which they agreed with the two statements using a seven-point scale (1 = Strongly Disagree, 7 = Strongly Agree): (1) “I see humanlike physical features from DeLight” and (2) “I see marketers’ intention to design DeLight as resembling a human” (r = 82). Analysis of variance on the average score of the two questions revealed only a significant main effect of DA (Mhi = 5.13, Mlo = 3.11; F(1, 41) = 26.41, p < .01; all other ps >.80). We also conducted an independent pretest (N = 92) to confirm that financial status manipulation does not lead to differences in participants’ mood. In the pretest, participants responded to the PANAS upon completion of the same wealth-manipulation task that we used in the main study. Results revealed no difference in participants’ mood, either in their positive (PArich = 2.83, PApoor = 2.73, t(90) = –.54, p = .59) or negative affect (NArich = 1.42, NApoor = 1.53, t(90) = 1.02, p = .31). Regarding the financial-status perception measured in the main study, a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) ANOVA revealed only a significant main effect of financial-status manipulation (Mrich = 59.78, Mpoor = 47.90; F(1, 158) = 15.30, p < .01). No other effects were significant (ps >.10), indicating a successful manipulation. FIGURE 2 View largeDownload slide STUDY 3: THE EFFECT OF PERCEIVED FINANCIAL STATUS AND DESIGNED ANTHROPOMORPHISM ON (1) PRODUCT EVALUATION, (2) PERCEIVED AGENCY (PA), AND (3) TREATMENT EXPECTATION FIGURE 2 View largeDownload slide STUDY 3: THE EFFECT OF PERCEIVED FINANCIAL STATUS AND DESIGNED ANTHROPOMORPHISM ON (1) PRODUCT EVALUATION, (2) PERCEIVED AGENCY (PA), AND (3) TREATMENT EXPECTATION Product Evaluation We conducted a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) ANOVA on the composite score of product evaluations (α = .93). Confirming our predictions, analysis revealed a significant two-way interaction of financial status and designed anthropomorphism (DA) (F(1, 158) = 9.65, p = .002; see figure 2). For the high-DA product, participants in the rich condition evaluated the product more positively than did those in the poor condition (Mrich hi = 5.07, SDrich hi = 1.55, Mpoor hi = 4.23, SDpoor hi = 1.43; F(1, 158) = 6.41, p = .01). However, for the low-DA product, participants in the perceived poor condition reported more positive evaluations than those in the perceived rich condition, although the difference was marginally significant (Mrich lo = 4.11, SDrich lo = 1.62, Mpoor lo = 4.75, SDpoor lo = 1.48; F(1, 158) = 3.50, p = .06). The opposite set of contrasts showed that participants who felt rich evaluated the high-DA product more positively than the low-DA product (F(1, 158) = 8.39, p = .004), but poor participants’ evaluations of the high- and low-DA products did not differ (F(1, 158) = 2.30, p = .13). FIGURE 3 View largeDownload slide STUDY 3: MODERATED MEDIATION RESULTS NOTE.—Financial status: rich = 0, poor = 1; designed anthropomorphism: high = 0, low = 1; β indicates a standardized coefficient; *p < .05; **p < .01; ***p < .001. FIGURE 3 View largeDownload slide STUDY 3: MODERATED MEDIATION RESULTS NOTE.—Financial status: rich = 0, poor = 1; designed anthropomorphism: high = 0, low = 1; β indicates a standardized coefficient; *p < .05; **p < .01; ***p < .001. Perceived Agency (PA) We also explored perceptions of agency for the described application using a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) ANOVA. We averaged the four measures—person, intention, personality, and free will—to create a composite score reflecting perceptions about product agency (α = .95). The ANOVA showed a main effect of the anthropomorphism manipulation (F(1, 158) = 9.45, p = .002). More importantly, and aligned with our theoretical prediction, we observed a significant interaction (F(1, 158) = 3.96, p = .04). Specifically, rich participants were more willing than poor participants to interpret the high-DA target’s anthropomorphic features as signaling agency (Mrich hi = 3.88, SDrich hi = 1.24, Mpoor hi = 2.62, SDpoor hi = 1.37; F(1, 158) = 4.50, p = .03). Regarding the low-DA product, on the other hand, participants did not show significantly different agency perception depending on their financial-status perception (Mrich lo = 3.13, SDrich lo = 1.34, Mpoor lo = 2.88, SDpoor lo = 1.36; F(1, 158) = .50, p = .48). The other pair of contrasts suggests that participants with low financial-status perceptions do not interpret human features embedded in a product by marketers as having agency. The rich participants were more inclined to afford agency to the high-DA product than to the low-DA product (F(1, 158) = 12.88, p < .01), whereas poor participants were disinclined to do so (F(1, 158) = .52, p = .47). Treatment Expectations and Control Measures We then examined whether people have different treatment expectations depending on their perceived financial status and a target’s level of designed anthropomorphism. An ANOVA revealed a significant interaction between financial-status perception and designed anthropomorphism on participants’ treatment expectations (F(1, 158) = 8.46, p = .004; see figure 2). Specifically, treatment expectations from the high-DA product were higher among people with high financial-status perception than those with low financial-status perception (Mrich hi = 5.01, SDrich hi = .95, Mpoor hi = 4.57, SDpoor hi = 1.08; F(1, 158) = 3.47, p = .06). However, the opposite pattern was observed for the low-DA product; people with low financial-status perception expected better treatment than people with high financial-status perception (Mrich lo = 4.28, SDrich lo = 1.19, Mpoor lo = 4.83, SDpoor lo = 1.05; F(1, 158) = 5.04, p = .03). Additionally, the rich participants had more positive treatment expectations regarding the high-DA product than the low-DA product (F(1, 158) = 9.49, p = .002), whereas the poor participants did not have different expectations (F(1, 158) = 1.12, p = .29). ANOVAs on the perceived control and general self-efficacy measures did not show any significant effects (all ps >.10), indicating that these variables do not account for the observed pattern of results. In particular, the perceived control measure was below the midpoint (4 out of seven-point scale) in all the conditions (all ps < .01), suggesting that, as we intended, participants were aware that they had limited direct control over the target. Further, the interactions on the product evaluation, PA, and treatment expectations measures reported above remained significant even after we controlled for both participants’ control perceptions and self-efficacy (F(1, 156) = 9.39, p = .002; F(1, 156) = 4.00, p = .04; F(1, 156) = 8.69, p = .003, respectively). Moderated Mediation Analyses We conducted a moderated mediation analysis with two serial mediators to test whether treatment expectations and agency perception mediate the interactive effect of financial status (rich vs. poor) and designed anthropomorphism (high vs. low) on product evaluations. We ran three multiple regression models. The first mediator model examined the effects of financial status, level of designed anthropomorphism, and their interaction on treatment expectations. This analysis revealed a significant effect of DA (b = .75, p = .002), financial status (b = .57, p = .02), and most importantly, a significant interaction of the two (b = –1.01, p = .003). The second mediator model examined the effects of financial status, DA, their interaction, and treatment expectations on perceived agency. Results showed significant effects of DA (b = .69, p = .04) and, importantly, of treatment expectations (b = .56, p < .001), while all the other effects were not significant (all ps >.50). The third dependent variable model examined the effects of financial status, DA, and their interaction, as well as treatment expectations and agency perception on product evaluations. The analysis revealed a significant effect of treatment expectations (b = .52, p < .001) and, importantly, a significant effect of agency perception on product evaluations (b = .31, p < .001). All the other effects were not significant (all ps >.10). A bootstrap analysis using the PROCESS model (Hayes 2013; Preacher and Hayes 2004) confirmed that the conditional indirect effect of the interaction between financial status and designed anthropomorphism on product evaluations through treatment expectation (mediator 1) and agency perception (mediator 2) was significant (95% CI = .0452 to .3166; 99% CI = .0169 to .3412). Further supporting our theory, a moderated serial mediation path with the reversed order of the two mediators was not significant (95% CI = 0 to 2310; 99% CI = –.0396 to .3657). Specifically, for the high-DA product, the indirect effect of treatment expectations and agency perception on product evaluations was significant and negative (b = –.44; 95% CI = –.8857 to –.0776), whereas it was significant and positive in the low-DA product condition (b = .38; 95% CI = .0066 to .8373). In other words, when the target product was highly humanized by marketers, participants who perceived themselves as being rich (vs. poor) expected better treatment, which led them to give the target product further agency and consequently to evaluate it more positively. On the contrary, when the product did not display explicitly humanlike features, the poor (vs. rich) participants evaluated it more positively, and this effect was serially driven by more favorable treatment expectations from, and greater willingness to afford agency to, the target product. Discussion Study 3 demonstrated that only those participants who perceived themselves to have high financial status evaluated the anthropomorphic features of a product more positively. Participants who felt rich showed more positive attitudes toward the high-DA product (1) than the low-DA product and (2) than did participants who felt poor. The current study also demonstrated that subjective financial standing affects consumers’ willingness to anthropomorphize a target. Participants who felt poor, expecting less favorable treatment from social entities, were reluctant to see the humanlike product as akin to an actual, thinking person. Corroborating our conceptual framework, a series of moderated mediation analyses revealed that it is consumers’ expectation that a company (and its representatives) would treat rich customers more favorably that leads to the heightened agency perception for products given humanlike features, finally resulting in preference for those products. Furthermore, the current experiment ruled out the possibilities that the observed effect is driven either by a heightened control perception over an entity or by higher self-efficacy among people who perceived their financial status as high. These results were replicated in our own follow-up study (N = 216, 119 male, Mage = 36.43; see web appendix B) using a different set of stimuli, an entertainment-recommendation program. Mirroring the current study’s results, we observed significant interactions between financial status and designed anthropomorphism on product evaluations (F(1, 196) = 4.16, p = .02), agency perception (F(1, 196) = 5.97, p = .003), and treatment expectations (F(1, 196) = 5.40, p = .005). People with higher (vs. lower) perceived financial status were more inclined to see the high-DA product as having agency, because they expected it to treat them better (b = .56; 95% CI: .2347 to 1.0246), and this motivated perception of agency increased preference for the humanlike target (b = .30; 95% CI: .0736 to .6920). In addition, unlike in study 3, this study found that the perceived poor participants afforded the low-DA product greater agency than the perceived rich participants did. We address the interesting results for the low-DA product further in the General Discussion. In our next study, we further demonstrate the role of treatment expectation by reversing the conventional consumers’ belief; specifically, we explore a scenario in which affluent consumers would not necessarily expect favorable treatment from an organization. STUDY 4 Study 4 was intended to overturn people’s commercial-treatment expectations depending on their financial standing, so that they believed that the poor would be more likely to be the focus of a company than the rich. In our prior studies, in which people’s expectations about how they would be treated in the marketplace were intact, we observed that people with high perceived financial status were more inclined to give agency and show more positive attitudes toward a product that was given human features by the marketer. However, we predicted the opposite pattern in the current study, where the social expectation linking financial status and commercial treatment was reversed. Method We recruited 270 participants in the United States (125 male, Mage = 36.44) from MTurk who received $.40 in exchange for completing the study. We excluded 15 participants per the established criteria. This study employed a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) between-subjects design, with the same wealth-manipulation method as in study 3. Participants were asked to read a description of a fictitious smartphone application called Tracker that keeps track of movement for a customized health goal. We manipulated the degree of featured humanity of the product using an image of a coach (high-DA) or a stopwatch (low-DA) that seemed to be delivering messages to users according to their performance. In this study, we provided additional information about the developer of the application. Before the product description, participants first read a passage explaining that the product was developed in conjunction with “Fair Economy,” a nonprofit network working against financial inequity and for a better life for people with limited financial resources (see appendix C). This description was intended to communicate organizational preference for people of lower financial status. After reading the description, participants reported the same product-evaluation and agency-perception measures used in the prior experiments. Finally, participants reported the maximum price they would be willing to pay for the product on a sliding bar with a range from $0 to $10, and indicated their demographic information. Results Manipulation Check We conducted an independent pretest (N = 46) on MTurk to ensure that participants perceived the difference in human features in the depiction of the application as we intended. Participants read descriptions of the movement-tracking application in the high- and low-DA conditions and answered for the two DA questions (r = .63) used in the stimuli pretest for the prior studies. Analysis of variance on the composite score revealed a significant difference between the high- and the low-DA conditions (Mhi = 4.36, Mlo = 2.95; F(1, 42) = 8.20, p < .01). No other effects were significant (ps <.80). In the main study, a 2 (financial status: rich or poor) × 2 (designed anthropomorphism: high or low) ANOVA on perceived financial status revealed only a significant main effect of financial-status manipulation (Mrich = 58.39, Mpoor = 50.; F(1, 251) = 11.66, p < .01; all other ps >.10), indicating a successful manipulation. Product Evaluation We predicted a reversal pattern on the product-evaluation index (α = .94). As expected, a 2 × 2 ANOVA revealed a significant interaction (F(1, 251) = 3.82, p = .05; see figure 4). In contrast to the prior results, but as expected, the high-DA product was preferred more by participants in the poor condition than by participants in the rich condition (Mrich hi = 4.12, SDrich hi = 1.73, Mpoor hi = 4.74, SDpoor hi = 1.58; F(1, 251) = 4.93, p = .03). The product evaluation did not differ for the low-DA product (Mrich lo = 4.68, SDrich lo = 1.47, Mpoor lo = 4.52, SDpoor lo = 1.68; F(1, 251) < 1, p > .10). Also, rich participants reported more positive evaluations of the low-DA product than the high-DA product (F(1, 251) = 3.98, p = .04). However, participants with low financial-status perception showed no difference in their product evaluations between high- and low-DA products (F(1, 251) = .32, p = .57). FIGURE 4 View largeDownload slide STUDY 4: THE EFFECT OF PERCEIVED FINANCIAL STATUS AND DESIGNED ANTHROPOMORPHISM ON (1) PRODUCT EVALUATION, (2) WTP ($), AND (3) PERCEIVED AGENCY (PA) FIGURE 4 View largeDownload slide STUDY 4: THE EFFECT OF PERCEIVED FINANCIAL STATUS AND DESIGNED ANTHROPOMORPHISM ON (1) PRODUCT EVALUATION, (2) WTP ($), AND (3) PERCEIVED AGENCY (PA) Perceived Agency (PA) The same 2 × 2 ANOVA was conducted on the composite PA measure (α = .81). In this analysis, only the interaction between financial status and designed anthropomorphism was significant (F(1, 251) = 4.12, p = .04; see figure 4). Also in line with our predictions, poor participants were more willing to give agency to the high-DA product than rich participants were (Mrich hi = 3.06, SDrich hi = 1.23,Mpoor hi = 3.63, SDpoor hi = 1.62; F(1, 251) = 5.17, p = .02), but we found no difference regarding the low-DA product between the rich and the poor participants (Mrich lo = 3.17, SDrich lo = 1.46, Mpoor lo = 3.01, SDpoor lo = 1.40; F(1, 251) = .38, p = .54). Further, the poor participants reported higher agency perceptions in the high- than in the low-DA product condition (F(1, 251) = 5.87, p = .01). By contrast, participants in the rich condition were disinclined to think of the high-DA product as a social being with agency, resulting in no difference in PA depending on the DA level (F(1, 251) = .18, p = .67). Willingness to Pay (WTP) Reflecting the interactive effect on the product-evaluation index, a 2 × 2 ANOVA on WTP also showed an interaction, although this effect was marginally significant (F(1, 251) = 3.18, p = .07; see figure 4). Analysis also revealed a marginally significant main effect of perceived financial status (F(1, 251) = 2.88, p = .09), but interestingly, participants in the poor condition reported higher WTP than those in the rich condition (Mrich = $1.60, Mpoor = $1.94). Planned contrasts revealed that for the high-DA product, participants in the poor condition reported higher WTP than those in the rich condition (Mrich hi = $1.46, SDrich hi = 1.52, Mpoor hi = $2.17, SDpoor hi = 1.77; F(1, 251) = 6.03, p = .01), whereas we found no difference for the low-DA product (Mrich lo = $1.75, SDrich lo = 1.58, Mpoor lo = $1.72, SDpoor lo = 1.55; F(1, 251) = .01, p = .92). Also, the poor participants’ WTP was higher for the high- than the low-DA product, although the difference was only directional (F(1, 251) = 2.18, p = .14). The rich participants showed no difference in WTP depending on the level of designed anthropomorphism (F(1, 251) = 1.08, p = .30). Discussion Findings in study 4 provide additional evidence supporting our theory about the role of treatment expectations as an underlying mechanism of the observed effects. Specifically, we posited that a greater preference for, and perceived agency of, products given humanlike features among people with high financial status would result from those people’s belief that companies want their business and train company “agents” to treat people with high financial status well. Thus, when the spontaneous expectation about money and preferential treatment in a commercial context was reversed in study 4, participants with high perceived financial status evaluated the high-DA product more negatively than did participants with low perceived financial status. The interactive effect of subjective financial standing and the level of designed anthropomorphism extended to a more consequential and practical variable that reflects the perceived value of a product: WTP. Further, the reversal of previous results was observed for perceived agency. Participants who perceived themselves as poor were more inclined to give agency to the humanized product (compared to the rich participants), with the expectation that they then would be treated well, which led to more positive evaluations of that product. Interestingly, we did not observe a difference in product preference between the high- and low-DA product among the poor participants as we had observed for the rich participants in the previous study. Even though we can only speculate for the reason behind this null effect, it might have been due to the unusual “reversed” treatment expectation employed in this study. Under conventional treatment expectations that we used in our prior studies (in which participants in the rich condition expect better treatment from sales agents), participants in the rich condition might have felt that interacting with a low-DA product that lacked humanlike characteristics would involve the loss of their usual special treatment, which in turn may have led to less positive evaluations for the low-DA product than for the high-DA one. However, in this reversal study in which the company appears to care more about customers with low financial status, participants in the poor condition presented with a low-DA product would not necessarily feel deprived of any preferential treatment that they think they deserve and normally get. Therefore, participants in the poor condition might not have shown a decrease in their preference for the low-DA product to the same extent that participants in the rich condition did in study 3. One might argue that the reversal observed in this study did not result from disrupted treatment expectations but rather from some unknown factor related to the Tracker stimuli. To check that possibility, we conducted a study following the same design as study 4 but which omitted the information about the application’s developer that may have favored poor over rich customers. In this study with conventional treatment expectations (N = 180, 84 male, Mage = 36.60; see web appendix C), we observed the same pattern of results as study 4—that is, a significant interaction of financial status and designed anthropomorphism on product evaluations (F(1, 177) = 3.66, p = .057), perceived agency (F(1, 177) = 6.78, p = .01), and WTP (F(1, 177) = 8.40, p = .003), but in the opposite direction. Participants in the rich condition saw greater agency in, preferred, and were willing to pay more for the high-DA product than participants in the poor condition were. We also observed the opposite for the low-DA product, for which the poor customers saw greater agency and higher WTP than the rich customers, an effect we discuss further in the General Discussion. STUDY 5 Results for studies 3 and 4 consistently support our theory that consumers show different attitudes toward a product with humanlike features depending on their financial status, which shapes their treatment expectations in a commercial context. Finally, in study 5, we aimed to demonstrate the robustness of our effect by employing a measured variable for participants’ financial status, instead of manipulating it. More importantly, we attempted to replicate our findings in a more naturalistic consumption setting: when consumers are not forced to think about their financial status. A limitation of our previous studies is that we manipulated or measured participants’ financial status before they evaluated the product. This process might have made consumers’ financial status more salient than usual and affected their assessment on a product. Therefore, in the current study, we minimize the salience of financial status by measuring participants’ financial standing at the end of the survey. Method We recruited 286 participants in the United States (135 male, Mage = 34.99) from MTurk who received $.30 in exchange for completing the study. We used the same Tracker stimuli as a target product. Participants were instructed to think about their daily experiences as a consumer and to evaluate the product, which was displayed either in the high- or low-DA condition. After reading the description, participants reported the same product-evaluation measure used in our prior studies. Because the extent to which people are health-conscious, and hence interested in the target product, could differ depending on their financial standing, we also asked participants how important 1) keeping in shape, 2) exercising regularly, 3) staying physically active was to them (α = .94) to control for any direct influences of financial status on product evaluation. At the end of the survey, to control for demographic features that could covary with financial status, we asked participants to report their education level and employment status, as well as their gender and age. Finally, participants indicated their personal and household annual income, both on an 11-point scale (1 = Less than $10,000, 11 = More than $100,000). We used the average of the two income-related answers as a measured variable of participants’ actual financial status. Results and Discussion We conducted a regression analysis on the product evaluation with three independent variables—participants’ financial status (M = 5.23, SD = .50), the product’s level of designed anthropomorphism, and the interaction of the two—and controlling variables including education level, employment status, and health consciousness. Results showed a significant main effect of health consciousness (b = .29, t(264) = 4.78, p < .001). However, more importantly, we observed a significant interaction between financial status and designed anthropomorphism (b = .29, t(264) = 3.70, p < .001). Simple slope tests revealed that participants with higher financial status evaluated the high-DA product more positively (b = .18, t(264) = 3.00, p = .003), whereas participants evaluated the low-DA product less positively as they got wealthier (b = –.10, t(264) = –1.98, p = .05). To further explore this interaction, we conducted a floodlight analysis to identify regions in the range of the moderator variable (financial status) in which the effect of the independent variable (designed anthropomorphism) on the dependent variable (product evaluation) is significant (Johnson and Neyman 1936; Spiller et al. 2013). The Johnson–Neyman point for p < .05 for the financial-status moderator occurred at the values of 4.32 and 8.39. This result indicates that the high-DA product was evaluated more positively than the low-DA product for all values of financial status above 8.39. In addition, the preference was reversed for participants with financial status below 4.32 (figure 5). FIGURE 5 View largeDownload slide STUDY 5: INTERACTION OF FINANCIAL STATUS AND DESIGNED ANTHROPOMORPHISM FIGURE 5 View largeDownload slide STUDY 5: INTERACTION OF FINANCIAL STATUS AND DESIGNED ANTHROPOMORPHISM Combined with the results in our prior studies, study 5 affirms that financial status and a product’s level of designed anthropomorphism interact to influence consumers’ reactions toward products. Whether we manipulated participants’ financial-status perception in two different ways or measured their actual financial standing by two different proxies, our key results were robust: consumers with high (vs. low) financial status preferred and more willing to give agency to the products that are given anthropomorphic features by marketers, and this effect was driven by different treatment expectations. GENERAL DISCUSSION Our findings suggest that anthropomorphizing may create favorable attitudes toward products only when consumers believe they are valued customers to the company. Across five studies, we found that products with humanlike features were perceived and evaluated differently depending on consumers’ financial status. For the most part, people with higher perceived financial status were more willing to perceive agency in products that were given humanlike features by marketers, and they liked these products better than did people with lower perceived financial status. Studies 1 and 2 as well as a pilot survey showed the foundational effect of financial status on people’s treatment expectations about other social entities’ (including products with humanlike features) behavior toward them. Extending their treatment expectations from salespeople to humanlike products, people showed differences in their willingness to disclose their personal information to those humanized products (study 1). In study 2, participants who perceived themselves as having high financial status tended to believe a self-driving car with human features would choose their well-being over the well-being of others, whereas participants with low perceived financial status tended to expect the car with human features would sacrifice their well-being for others’. Building on this finding regarding people’s treatment expectations depending on financial standing, studies 3–5 looked further into consumers’ evaluations of humanized products. Participants expecting better treatment from a commercial entity interpreted anthropomorphic features that marketers had designed into the product as signaling greater agency, and they liked the products with human features more than did those in the poor condition. Product evaluations appear to be serially driven by differences in treatment expectations and agency perception. Mediational evidence suggested that high financial-status perceptions increased favorable commercial treatment expectations from social entities (high-DA products), which in turn led to higher agency perceptions and more positive product evaluations. Study 3 provided converging evidence for the interactive effect between financial status and designed anthropomorphism through treatment expectations, while it ruled out alternative accounts based on social power or efficacy over the entity. Further supporting the role of treatment expectations, study 4 demonstrated the opposite pattern of results when the conventional relationship between financial status and commercial treatment expectations was reversed. Specifically, when people with low financial status expected to be treated better than people with high financial status, they perceived greater agency in a product that marketers had given humanlike features and liked the product better. Findings in the Low-DA Condition Throughout the studies, we observed a pattern of results that was highly interesting and not anticipated in our hypothesis derivation regarding the low-DA products, which were not intentionally designed by the marketer to resemble a person. In the studies that involved conventional treatment expectations (i.e., not the study that reversed expectations) and in which we directly measured treatment expectations and product evaluations, participants with low financial standing expected more positive treatment from, and showed more positive evaluations of, the low-DA products than those with high financial standing did (study 2: [car’s decision prediction] 3.13 (rich) < 4.22 (poor), p = .06; study 3: [product evaluation] 4.11 < 4.75, p = .06, [treatment expectation] 4.28 < 4.83, p = .03; study 5 [product evaluation; comparing ±1SD from mean]: 4.41 < 5.21, p = .07). These results suggest that when consumers are presented with products like those used in our studies—products that offer a modest level of artificial intelligence (AI)—and those products are not designed to seem like a person, consumers might differ in their attitudes depending on how they are accustomed to being treated by the actual people that these products might conceivably replace. When anticipating interaction with low-DA entities, consumers who are financially well-off might expect to experience deprivation of their usual special treatment, whereas those with low financial resources would not feel a loss of status. In other words, in a world of objects, consumers would expect the mechanistic entities to function in an egalitarian way. This “equal ground” for both the rich and the poor could be perceived differently depending on how consumers have been treated in the real commercial world. The very lack of special treatment, to which consumers with high financial status think they are entitled, could be particularly discomforting to the rich, while particularly comforting to the poor. Therefore, when consumers encounter products low in designed anthropomorphism, those consumers with low financial status might expect more favorable treatment than do those with high financial status. That effect could flow through to more positive evaluations on the low-DA products among those people with low (vs. high) financial status. However, given the marginal results around this effect in our studies, we believe future research is required to systematically test this hypothesis. This interesting pattern was also observed in the perceived agency measure. Follow-up studies to studies 3 and 5 revealed that participants who perceived themselves as poor saw more agency in the low-DA products than did those who perceived themselves as rich. Consistent with the discussion above, we speculate that one possible explanation for this effect is that people with low financial status might have felt more comfortable with a somewhat human agent that didn’t seem to have been trained by marketers, and so were inclined to give it agency. One important thing to note is that in our studies, even though not explicitly depicting humanlike physical features, the products in the low-DA conditions could nevertheless be seen as having some degree of agency—by offering advice and keeping track of tastes or goals, for example. Consumers could believe that a product given humanlike features by marketers, a high-DA product, explicitly represents a company’s business strategy, whereas a less-humanized product, a low-DA product, would be more detached from the company’s value maximizing position—much the same as salespeople in the front of the store would be expected to execute marketing strategy, but the stock worker in the back would not. Thus, people with low financial standing might have felt a sense of solidarity with products with at least some human characteristics and so were inclined to grant these products agency. This argument finds some conceptual support in research by Duclos, Wan, and Jiang (2013), which showed that consumers sometimes trade off social and financial resources. Hence, those participants who perceived themselves to have low financial status in our studies might have been looking for “human” support to compensate for their relative poverty, more so than those who perceived themselves to have high financial status—so long as the targets were not obviously manipulated by the marketer to seem human. This argument is also consistent with the results of the two studies in which we did not observe an effect for the low-DA product: specifically, study 3, in which customers and the low-DA product could not interact directly (and so could not establish any sense of bonding or solidarity), and study 4, in which the wealthy participants would likely not feel a strong sense of connection with the low-DA “worker.” Future research might therefore find it fruitful to investigate moderating factors for the effect of financial resources on perceived agency, which can be traced to beliefs about how the new social relationship would pan out. Theoretical Contributions This research makes novel contributions to the anthropomorphism literature. Our findings indicate that consumers’ treatment expectations could determine the level of agency they afford products given human features. Beyond the three determinants of anthropomorphism (sociality, effectance motivation, and elicited agent knowledge; Epley, Waytz, and Cacciopo 2007) that prior research has suggested, the current results offer a different lens to explore how consumers’ different cognitive or social motivations influence their tendency to anthropomorphize a nonhuman target. The methods, context, and design of most prior consumer anthropomorphism studies lacked an “interactive” relationship between consumers and the anthropomorphized products, which might have led to negligible difference in consumers’ motivation to perceive agency from the humanlike targets. By contrast, in our article, target products are more engaged in consumer interactions—for example, determining and then notifying customers of delivery time in study 3, providing exercise feedback depending on users’ performance in studies 4 and 5, and even determining consumers’ safety in study 2. An exception to this argument is presented by studies in work by Kim and McGill (2011) in which participants had the opportunity to consider the dangers of a disease that they might encounter, which was described in human terms, or to gamble with a slot machine, which had been given human features. However, these authors did not find differences in perceived anthropomorphism or object agency, depending on the social power of the participant, although they observed differences in perceived risk in interactions with these entities depending on participants’ perceived power. This null effect might be traced to the participants’ expectations regarding the possibility of preferential treatment by these entities. In Kim and McGill’s studies, it is possible that social power could have affected people’s motivation to see the entity as being able to treat them well or poorly at least somewhat—the powerful might have imagined good treatment and the weak might have feared bad treatment—and that could have affected willingness to afford the entity agency. However, this tendency would likely be weaker than in the present studies because participants in the Kim and McGill studies would not have had reason to expect that the disease or the machine was inclined or trained, respectively, to treat some people better than others. The disease scenario lacked a marketer with any particular goal or profit motive and the gambling scenario offered no reason to prioritize low- versus high-power customers. As a result, the low-power participants would not expect that a “sentient” disease or machine would single them out for poor treatment. Hence, they may have been less inclined to deny the humanized entity agency. Practical Implications Our findings suggest that marketers should be mindful when presenting their products in anthropomorphic terms. Those consumers who feel financially well-off might willingly accept these products as being like real people, because they believe commercial agents usually treat them well. Those who do not feel well-off, however, might resist marketers’ efforts to anthropomorphize products, because these products could then become yet more people who are able to make their lives difficult. Less affluent consumers would be less likely to perceive the possibility of a positive interaction with these highly humanized products. Such a response could damage marketing attempts targeting lower-income consumers. In particular, as suggested in studies 1 and 2, the consequences of anthropomorphism go beyond the mere positive or negative evaluations of a product. People have different expectations about how an anthropomorphized social entity will serve them even in situations involving their lives and safety. Such differences would have particularly meaningful implications for the healthcare, transportation, and security industries, which are major domains where AI is being developed and adopted with astonishing speed. Marketers and public policy makers in these domains should use caution in utilizing anthropomorphism strategies, especially when describing how their services work, depending on their target populations. In particular, the wise usage of anthropomorphism strategies could contribute to closing the technology gap between the rich and poor by increasing trust and reducing unnecessary anxiety around new technology adoption among lower-income individuals. Future Research Directions Our work suggests several interesting avenues for future research. Although the current study did not specify the role of the anthropomorphized entity (i.e., product-as-partner vs. product-as-servant), future work examining how the intended characteristics of the consumer-product relationship influence the interactive effect of perceived financial status and anthropomorphism would constitute an important contribution to the literature. Recent research has shown materialists respond more favorably to an anthropomorphized brand, especially when the brand role is a servant instead of a partner (Kim and Kramer 2015). Further, Kim and Kramer’s findings suggest that differences in consumers’ desire to dominate the target brand drove the distinctive responses to the partner and servant positioning. Considering these findings, we might expect the influence of (perceived) financial status to be magnified when a consumer interacts with a product or service positioned as a servant, whereas that influence could be attenuated or even reversed when the target product is positioned as a partner or an expert (with a seemingly superior status) who teaches/gives professional advice to users. Another opportunity for future research is to explore other factors that impact consumers’ treatment expectations, such as a consumer’s physical attractiveness or perceived fit with a company’s target market. These attempts could further enlarge our findings regarding the effect of financial status toward a more general role of consumers’ expectations from the humanlike entities, which better captures the essence of our theorizing. For example, our own follow-up study, which employed age, supported our theorizing regarding the process driving our observed effects. A pretest (N = 44) showed that as people get older, they are more likely to think their age can positively impact their treatment in typical commercial contexts β = .04, t(42) = 2.73. Based on this preliminary finding, we conducted a study using the same target product as study 4, using age instead of financial status as an independent variable influencing treatment expectations. (We omitted information about the developer of the application, which we had employed in study 4 to disrupt expectations regarding who is likely to be treated well in the marketplace.) In the follow-up study, instead of the wealth-manipulation phase, we had participants (N = 119, 63 male, Mage = 36.9, SDage = 10.3) read a paragraph suggesting that health in every life stage is important. Then, they read the description of the movement tracking application and responded to the product evaluation, PA, WTP, and treatment expectation measures. A series of 2 (age: low vs. high; split into Mage ± .5* SDage) × 2 (designed anthropomorphism: high vs. low) ANOVAs revealed significant interactions on product evaluation (F(1, 115) = 21.49, p < .01), PA (F(1, 115) = 4.53, p < .05), WTP (F(1, 115) = 6.20, p < .05), and treatment expectation measures (F(1, 115) = 16.80, p < .01), mirroring the effect found in our prior studies. These effects remained significant even after we controlled for personal income or when using age as a continuous variable (see web appendix D and E for additional details). Such results provide additional process evidence for the role of treatment expectations by showing that the moderating effect on anthropomorphism is not confined to financial status or due to other factors that covary with financial status, but rests upon consumers’ belief about how other entities would treat them, which includes social norms. In sum, the current findings show that consumers evaluate products conveying human features differently depending on their financial status. People with high financial status like high-anthropomorphized products more than people with low financial status do, and more than they like low-anthropomorphized products, as long as they hold the belief that commercial entities treat the rich better. Our findings offer a novel contribution to the anthropomorphism literature by distinguishing consumers’ willingness to afford human agency from marketers’ intentions to add human features to a product. Even with the same level of marketer effort to anthropomorphize a product, people with high perceived financial status, having more positive commercial-treatment expectations, are more likely to accept and extend the anthropomorphism to give the product greater agency. This different assessment of high-/low-anthropomorphized products is based on different commercial treatment expectations people infer from a target in the given context. Our results indicate that there is more to effective anthropomorphism than simply adding superficial humanlike features to the product: it also needs to take into account consumers’ motivation to interpret a target with humanlike features as having positive agency. DATA COLLECTION INFORMATION All data were collected on Amazon Mechanical Turk. The data collection periods are as follows: pilot survey in summer 2017, study 1 in fall 2017, study 2 in spring 2016, study 3 in spring 2017, study 4 in fall 2016, and study 5 in fall 2017. All the data were collected and analyzed by the first author under the guidance of the second author. Financial support from the Kilts Center for Marketing, University of Chicago, is gratefully acknowledged. The authors would like to thank Pankaj Aggarwal and Sara Kim for their insightful and valuable comments on an earlier version of this article. The authors also thank the editor, associate editor, and reviewers for their constructive feedback. This article is based on part of the first author’s doctoral dissertation. Supplementary materials including follow-up studies and experimental materials are included in the web appendix accompanying the online version of this article. Appendix A PRODUCT ANTHROPOMORPHISM (STUDY 1) Appendix B PRODUCT ANTHROPOMORPHISM (STUDY 3) Appendix C COMMERCIAL-TREATMENT-EXPECTATION MANIPULATION (STUDY 4, HIGH-DA CONDITION) References Adler Nancy E. , Epel Elissa S. , Castellazzo Grace , Ickovics Jeannette R. ( 2000 ), “Relationship of Subjective and Objective Social Status with Psychological and Physiological Functioning: Preliminary Data in Healthy, White Women,” Health Psychology , 19 6 , 586 – 92 . Google Scholar CrossRef Search ADS PubMed Aggarwal Pankaj , McGill Ann L. ( 2007 ), “Is That Car Smiling at Me? Schema Congruity as a Basis for Evaluating Anthropomorphized Products,” Journal of Consumer Research , 34 4 , 468 – 79 . Google Scholar CrossRef Search ADS Ahn Hee-Kyung , Kim Hae Joo , Aggarwal Pankaj ( 2014 ), “Helping Fellow Beings: Anthropomorphized Social Causes and the Role of Anticipatory Guilt,” Psychological Science , 25 1 , 224 – 9 . Google Scholar CrossRef Search ADS PubMed Anderson Cameron , John Oliver P. , Keltner Dacher ( 2012 ), “The Personal Sense of Power,” Journal of Personality , 80 2 , 313 – 44 . Google Scholar CrossRef Search ADS PubMed Baumeister Roy F. , Leary Mark R. ( 1995 ), “The Need to Belong: Desire for Interpersonal Attachments as a Fundamental Human Motivation,” Psychological Bulletin , 117 3 , 497 – 529 . Google Scholar CrossRef Search ADS PubMed Bolton Ruth N. , Kannan P. K. , Bramlett Matthew D. ( 2000 ), “Implications of Loyalty Program Membership and Service Experiences for Customer Retention and Value,” Journal of the Academy of Marketing Science , 28 1 , 95 – 108 . Google Scholar CrossRef Search ADS Bonnefon Jean-François , Shariff Azim , Rahwan Iyad ( 2016 ), “The Social Dilemma of Autonomous Vehicles,” Science , 352 6293 , 1573 – 76 . Google Scholar CrossRef Search ADS PubMed Boyd Henry C. , Helms Janet E. ( 2005 ), “Consumer Entitlement Theory and Measurement,” Psychology & Marketing , 22 3 , 271 – 86 . Google Scholar CrossRef Search ADS Burgoon Judee K. , Bonito Joseph A. , Bengtsson Bjorn , Cederberg Carl , Lundeberg Magnus , Allspach L. ( 2000 ), “Interactivity in Human–Computer Interaction: A Study of Credibility, Understanding, and Influence,” Computers in Human Behavior , 16 6 , 553 – 74 . Google Scholar CrossRef Search ADS Chandler Jesse , Schwarz Norbert ( 2010 ), “Use Does Not Wear Ragged the Fabric of Friendship: Thinking of Objects as Alive Makes People Less Willing to Replace Them,” Journal of Consumer Psychology , 20 2 , 138 – 45 . Google Scholar CrossRef Search ADS Chen Gilad , Gully Stanley M. , Eden Dov ( 2001 ), “Validation of a New General Self-Efficacy Scale,” Organizational Research Methods , 4 1 , 62 – 83 . Google Scholar CrossRef Search ADS Chen Rocky Peng , Wan Echo Wen , Levy Eric ( 2017 ), “The Effect of Social Exclusion on Consumer Preference for Anthropomorphized Brands,” Journal of Consumer Psychology , 27 1 , 23 – 34 . Google Scholar CrossRef Search ADS Dion Delphine , Borraz Stéphane ( 2017 ), “Managing Status: How Luxury Brands Shape Class Subjectivities in the Service Encounter,” Journal of Marketing , 81 5 , 67 – 85 . Google Scholar CrossRef Search ADS Duclos Rod , Wan Echo Wen , Jiang Yuwei ( 2013 ), “Show Me the Honey! Effects of Social Exclusion on Financial Risk-Taking,” Journal of Consumer Research , 40 1 , 122 – 35 . Google Scholar CrossRef Search ADS Epley Nicholas , Akalis Scott , Waytz Adam , Cacioppo John T. ( 2008 ), “Creating Social Connection through Inferential Reproduction Loneliness and Perceived Agency in Gadgets, Gods, and Greyhounds,” Psychological Science , 19 2 , 114 – 20 . Google Scholar CrossRef Search ADS PubMed Epley Nicholas , Waytz Adam , Cacioppo John T. ( 2007 ), “On Seeing Human: A Three Factor Theory of Anthropomorphism,” Psychological Review , 114 4 , 864 – 86 . Google Scholar CrossRef Search ADS PubMed Guthrie Stewart E. ( 1993 ), Faces in the Clouds: A New Theory of Religion , New York : Oxford University Press . Hayes Andrew F. ( 2013 ), Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach , New York : Guilford Press . Hur Julia D. , Koo Minjung , Hofmann Wilhelm ( 2015 ), “When Temptations Come Alive: How Anthropomorphism Undermines Self-Control,” Journal of Consumer Research , 42 2 , 340 – 58 . Johnson Palmer O. , Neyman J. ( 1936 ), “Tests of Certain Linear Hypotheses and Their Application to Some Educational Problems,” Statistical Research Memoirs , 1 , 57 – 93 . Kim Hyeongmin Christian , Kramer Thomas ( 2015 ), “Do Materialists Prefer the ‘Brand-as Servant’? The Interactive Effect of Anthropomorphized Brand Roles and Materialism on Consumer Responses,” Journal of Consumer Research , 42 2 , 284 – 99 . Google Scholar CrossRef Search ADS Kim Sara , McGill Ann L. ( 2011 ), “Gaming with Mr. Slot or Gaming the Slot Machine? Power, Anthropomorphism, and Risk Perception,” Journal of Consumer Research , 38 1 , 94 – 107 . Google Scholar CrossRef Search ADS Labroo Aparna A. , Dhar Ravi , Schwarz Norbert ( 2008 ), “Of Frowning Watches and Frog Wines: Semantic Priming, Perceptual Fluency, and Brand Evaluation,” Journal of Consumer Research , 34 6 , 819 – 31 . Google Scholar CrossRef Search ADS Lacey Russell , Suh Jaebeom , Morgan Robert M. ( 2007 ), “Differential Effects of Preferential Treatment Levels on Relational Outcomes,” Journal of Service Research , 9 3 , 241 – 56 . Google Scholar CrossRef Search ADS Landwehr Jan R. , McGill Ann L. , Herrmann Andreas ( 2011 ), “It’s Got the Look: The Effect of Friendly and Aggressive ‘Facial’ Expressions on Product Liking and Sales,” Journal of Marketing , 75 3 , 132 – 46 . Google Scholar CrossRef Search ADS Leary Mark R. , Kelly Kristine M. , Cottrell Catherine A. , Schreindorfe Lisa S. ( 2013 ), “Construct Validity of the Need to Belong Scale: Mapping the Nomological Network,” Journal of Personality Assessment , 95 6 , 610 – 24 . Google Scholar CrossRef Search ADS PubMed Mende Martin , Scott Maura L. , Lemon Katherine N. , Thompson Scott A. ( 2015 ), “This Brand Is Just Not That into You,” in Strong Brands, Strong Relationships , ed. Fournier Susan , Breazeale Michael J. , Avery Jill , New York : Routledge . Mittal Vikas , Sarkees Matthew , Murshed Feisal ( 2008 ), “The Right Way to Manage Unprofitable Customers,” Harvard Business Review , 86 4 , 94 – 103 . Google Scholar PubMed Mogilner Casey ( 2010 ), “The Pursuit of Happiness: Time, Money, and Social Connection,” Psychological Science , 21 9 , 1348 – 54 . Google Scholar CrossRef Search ADS PubMed Morewedge Carey K. ( 2009 ), “Negativity Bias in Attribution of External Agency,” Journal of Experimental Psychology: General , 138 4 , 535 – 45 . Google Scholar CrossRef Search ADS PubMed Morris Davis Z. ( 2016 ), “Mercedes-Benz’s Self-Driving Cars Would Choose Passenger Lives over Bystanders,” Fortune, http://fortune.com/2016/10/15/mercedes-selfdriving-car-ethics. Nelson Leif D. , Morrison Evan L. ( 2005 ), “The Symptoms of Resource Scarcity Judgments of Food and Finances Influence Preferences for Potential Partners,” Psychological Science , 16 2 , 167 – 73 . Google Scholar CrossRef Search ADS PubMed Preacher Kristopher J. , Hayes Andrew F. ( 2004 ), “SPSS and SAS Procedures for Estimating Indirect Effects in Simple Mediation Models,” Behavior Research Methods, Instruments, and Computers , 36 4 , 717 – 31 . Google Scholar CrossRef Search ADS Puzakova Marina , Kwak Hyokjin , Rocereto Joseph ( 2009 ), “Pushing the Envelope of Brand and Personality: Antecedents and Moderators of Anthropomorphized Brands,” Advances in Consumer Research , 36 1 , 413 – 9 . Puzakova Marina , Kwak Hyokjin , Rocereto Joseph ( 2013 ), “When Humanizing Brands Goes Wrong: The Detrimental Effect of Brand Anthropomorphization amid Product Wrongdoings,” Journal of Marketing , 77 3 , 81 – 100 . Google Scholar CrossRef Search ADS Reczek Rebecca Walker , Haws Kelly L. , Summers Christopher A. ( 2014 ), “Lucky Loyalty: The Effect of Consumer Effort on Predictions of Randomly Determined Marketing Outcomes,” Journal of Consumer Research , 41 4 , 1065 – 77 . Google Scholar CrossRef Search ADS Schwarz Norbert ( 1999 ), “Self Reports: How the Questions Shape the Answers,” American Psychologist , 54 2 , 93 – 105 . Google Scholar CrossRef Search ADS Spiller Stephen A. , Fitzsimons Gavan J. , Lynch John G. Jr. , McClelland Gary H. ( 2013 ), “Spotlights, Floodlights, and the Magic Number Zero: Simple Effects Tests in Moderated Regression,” Journal of Marketing Research , 50 2 , 277 – 88 . Google Scholar CrossRef Search ADS Touré-Tillery Maferima , McGill Ann L. ( 2015 ), “Who or What to Believe: Trust and the Differential Persuasiveness of Human and Anthropomorphized Messengers,” Journal of Marketing , 79 4 , 94 – 110 . Google Scholar CrossRef Search ADS Ward Morgan K. , Dahl Darren W. ( 2014 ), “Should the Devil Sell Prada? Retail Rejection Increases Aspiring Consumers’ Desire for the Brand,” Journal of Consumer Research , 41 3 , 590 – 609 . Google Scholar CrossRef Search ADS Waytz Adam , Epley Nicholas , Cacioppo John T. ( 2010 ), “Social Cognition Unbound Insights into Anthropomorphism and Dehumanization,” Current Directions in Psychological Science , 19 1 , 58 – 62 . Google Scholar CrossRef Search ADS PubMed Zhou Xinyue , Vohs Kathleen D. , Baumeister Roy F. ( 2009 ), “The Symbolic Power of Money: Reminders of Money Alter Social Distress and Physical Pain,” Psychological Science , 20 6 , 700 – 6 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc. All rights reserved. For permissions, please e-mail: email@example.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Journal of Consumer Research – Oxford University Press
Published: Jan 29, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera