Master thesis MSc Marketing, specialization Marketing Management University of Groningen Faculty Economics and Business

(1)

THE UNEXPECTED SURPRISE IN CONSUMPTION: HOW SURPRISE LABELS

AFFECT CONSUMPTION BEHAVIOURS

by

Jildou Bouma

Master thesis

MSc Marketing, specialization Marketing Management

University of Groningen

Faculty Economics and Business

(2)

THE UNEXPECTED SURPRISE IN CONSUMPTION: HOW SURPRISE LABELS

AFFECT CONSUMPTION BEHAVIOURS

ABSTRACT

In the marketplace, brands are often using elements of surprise in their marketing communication. The unexpected-framing can help marketers to make their product more attractive. Although this is a common practice, it remains unclear whether this strategy is also positive for the consumption experience. Should brands indeed explicitly label surprises in their marketing communication or not? This research examined the extent to which surprise labels influence consumption behaviours. In addition, the moderating role of self-affirmation was examined. Although scholars have predominantly focused on the positive side of unexpected-framing in pre-consumption stages, I propose that surprise labels may entail negative implications in the consumption stage. In particular, the prospective meaning violation and uncertainty in face of a surprise label in an ongoing consumption experience elevates arousal-induced anxiety, resulting in avoidance and compensatory behaviours. Hence, I propose that the threat response will lead to lower levels of consumption and that consumers will compensate for the prospective meaning violation by reinforcing highly accessible schemata in unrelated consumption domains. An online experiment did not confirm that a surprise label lowers levels of consumption and encourages compensatory consumption, such that it did not heighten the preference for dominant brands. Furthermore, self-affirmation did not alleviate the threat response, such that if consumers positively affirmed their self before being exposed to a surprise label, their consumption and compensation levels were not alleviated. Taken together, surprise labels do not seem to harm consumption behaviours, and marketers can for now keep employing surprise labels.

Keywords: Surprise, Meaning violation, Consumption, Compensatory consumption, Fluid

compensation, Self-affirmation

First supervisor: A. Schumacher

(3)

TABLE OF CONTENTS

1. INTRODUCTION ... 4

2. THEORETICAL FRAMEWORK ... 6

2.1 Surprise and the nonconscious threat response ... 6

2.2 Ongoing consumption experience ... 8

2.3 Fluid Compensation ... 10

2.4 Self-affirmation – Alleviating the nonconscious threat response ... 12

(4)

1. INTRODUCTION

In the marketplace, we can observe the ubiquitous notion of brands communicating their value by underlining the element of surprise. In other words, in their marketing communication they frame their brand with unexpected benefits. For example, the tagline of the Korean automobile manufacturer Kia reads “The power to surprise”. Recent ad campaigns of Volvo (2018) and Reebok (2019) promise “See the Unexpected” and “Sport the Unexpected”. The Canadian TD Bank (2019) even claims to be “Unexpectedly Human” in a repositioning campaign and a Dutch surprise travel agency SRPRS.ME (2019) advertises with the slogan “Make a Date with the Unexpected”. Such brand promises might have a positive impact on consumers. Prior research suggests that relevant unexpected-framing elicits more positive attitudes towards the product and increases product choice (Wadhwa, Kim, Chattopadhyay, & Wang, 2019; Lee & Mason, 1999). This is because unexpected benefits cue potential consumers to believe that other unexpected rewards from the product can be anticipated (Wadhwa et al., 2019). Moreover, surprises can heighten emotional reactions in that unexpected outcomes can cause more pleasantness than expected outcomes (Valenzuela, Mellers & Strebel, 2010; Mellers, Schwartz, Ho, & Ritov, 1997).

(5)

influence on consumption behavior (Bechara & Damasio, 2005; Winkielman, Berridge, & Wilbarger, 2005). Moreover, in situations such as an ongoing consumption experience, consumers show an approach aversion to stimuli that they perceive to be approaching (Hsee, Tu, Lu, & Ruan, 2014). Similarly, I expect that surprise labels cue prospective meaning violations and therefore may trigger a nonconscious threat response among consumers. Especially, I expect that the ongoing consumption experience attenuates the aversiveness of the prospective meaning violation and accompanied uncertainty compared to pre-consumption stages. This nonconscious threat response in turn manifests in consumption disengagement and thus in earlier termination of consumption.

Previous research has shown that positive self-affirmation prior to exposure to a threat may reduce the nonconscious threat response (Finley, Crowell, & Schmeichel, 2018; McQueen & Klein, 2006; Sherman & Cohen, 2006; Sherman & Cohen, 2002; Steele, 1988). Thus, the defensive response in one domain can be attenuated by positively affirming the self in another domain. Furthermore, consumers can also alleviate this aversive tension to surprise labels by compensating in other consumption domains, such as the preference for dominant brands (Taylor & Noseworthy, 2020; Proulx & Inzlicht, 2012; Proulx & Heine, 2008). Thus, if prospective surprises indeed threaten consumers’ need for certainty and meaning, surprise labels may have unintended consequences for consumption levels and stimulate compensation behaviour. However, these negative effects of a surprise label may be avoided by positive self-affirmation beforehand.

(6)

labels is questionable. Even though surprise labels seem to be helpful for marketers in making the product more attractive, the lower levels of consumption could reduce revenues and even customer loyalty. Specifically, consumers may not be able to consciously name the aversive feelings that are caused by the surprise labels, but they may still experience the consumption as less enjoyable. Furthermore, other non-dominant brands in the product portfolio may lose sales because consumers may compensate by favouring category dominant brands when facing a surprise label. Consequently, understanding what drives consumption in the presence of surprise labels can help marketers to avoid decreasing consumption levels. In particular, it would help marketers in deciding the optimal surprise framing by showing that earlier termination of consumption can be avoided by affirming the consumer’s positive self-image beforehand. For example, in an advertisement marketers could first trigger the consumer to think about a universal positive value that makes them feel good about themselves before showing the surprise label.

In the following sections the theoretical framework, the research design and the results are discussed. In the last section the discussion of the findings and the theoretical and practical contributions of this study are presented. Furthermore, the limitations of the study and directions for future research are discussed.

2. THEORETICAL FRAMEWORK

In this section, I draw upon different streams of research to explore how surprise labels could impact consumption behaviours and how this effect could be alleviated.

2.1 Surprise and the nonconscious threat response

(7)

mental representations of expected relations to incoming information (Proulx & Inzlicht, 2012; Heine et al., 2006). In other words, people maintain schemata that organize their experiences and allow them to make sense of their experiences. Moreover, these schematic expectations help people to predict and control their world (Proulx & Inzlicht, 2012; Heine et al., 2006). Hence, they derive a sense of certainty from these expected relations. Seen in this light, meaning refers to the ability to confirm expected relationships. However, schematic expectations are sometimes violated by experiences that do not cohere with these expectations and therefore perceptions of uncertainty arise. For instance, a surprise cannot be easily integrated into activated schemata and thus we do not know what to expect. This threatens one’s sense of meaning, even if the surprise is prospective. Hence, a surprise shares resemblance with a meaning violation, that is any experience that is inconsistent with expectations that follow from our understandings (Proulx & Inzlicht, 2012; Heine et al., 2006). According to MMM, these schema inconsistencies nonconsciously motivate us to reaffirm other available and intact schemata to restore a sense of meaning, rather than reinterpreting or revising our existing schemata (Taylor & Noseworthy, 2020; Proulx & Inzlicht, 2012; Heine et al., 2006). This coping strategy to reduce the aversive tension is driven by a physiological state of aversive arousal accompanied by the psychological experience of anxiety or threat (Townsend, Eliezer, & Major, 2013; Proulx & Inzlicht, 2012). Thus, consumption experiences with a prospective surprise can announce a threat to one’s sense of meaning, ultimately leading to an anxiety response.

(8)

other hand, prior research also suggests that products might be evaluated more positively when they are moderately incongruent, because consumers can resolve the inconsistency. Hence, the resolution is without a fundamental change in the consumer’s cognitive structure. Subsequently, this leads to more positive affect than when the product attribute is congruent or extremely incongruent (Meyers-Levy & Tybout, 1989; Mandler, 1982). However, the anticipation of a surprise is similar to extreme incongruency that similarly implies uncertainty and prospective violations of meaning. Moreover, the anticipation of a surprise can also yet not be revolved as with extreme incongruency.

The schema congruity theory is based upon the view that people have a sense-making impulse and that inconsistencies can cause tension (Taylor & Noseworthy, 2020; Piaget, 1960). Hence, both the schema congruity literature and the meaning maintenance literature involve around the inability to make sense of the presented input and thus the violation of expectations, followed by an aversive tension. In both theories, the inconsistencies elicit an autonomic physiological response (i.e., arousal) that manifests as anxiety (Noseworthy et al., 2014; Proulx & Inzlicht, 2012). Hence, both theories can be unified under a common emotional mechanism: arousal-induced anxiety (Taylor & Noseworthy, 2020). Specifically, the arousal-induced anxiety engenders a nonconscious need to regain meaning by affirming pre-existing schemata in unrelated domains and causes extreme incongruent products to be evaluated negatively (Noseworthy et al., 2014; Proulx & Inzlicht, 2012).

In this research, to study the effect of surprise labels on consumption we should focus on the ongoing consumption experience. In a similar vein to the theory of schema congruity and the meaning maintenance model, I expect that the anticipation of a surprise, induced by a mere surprise label, signals a prospective threat to one’s sense of meaning. This threat to our meaning and certainty triggers a nonconscious threat response, which manifests in arousal-induced anxiety and behavioural avoidance. As a result, the consumption is terminated earlier than a consumption experience without a surprise label. In the next part, I will further explain this underlying mechanism and why it is unique to the consumption stage.

2.2 Ongoing consumption experience

(9)

perceived as positive, but if the stimulus is becoming closer in time or in probability, the stimulus is received as less positive (Hsee et al., 2014). Hence, the difference in the surprise effect between the stages can be explained by the perceived dynamic movement of the surprise. The dynamic movement of the stimuli can be divided in both the temporal movement and in the probabilistic movement (Hsee et al., 2014). The former explains that the moment of the surprise approaches the here and now. The latter explains that the probability of the occurrence of a possible surprise increases. The perceived dynamic movement increases the vivid representation of the event and leads to heightened levels of perceived risk as the perceived likelihood of occurring increases, and ultimately affects behavioural response (Cian, Krishna, & Elder, 2015). Thus, a surprise label in an ongoing consumption experience causes more perceived risk (i.e., threat), due to the increase in dynamic movement compared to pre-consumption stages. Therefore, surprise labels may increase product preference in pre-pre-consumption stages but exert a negative influence during the consumption stage (Wadhwa et al., 2019). In other words, the aversiveness of the prospective meaning violation is accentuated in the ongoing consumption experience.

(10)

affective reaction which people are simply not aware of but does alter consumption behavior (Winkielman et al., 2005). Thus, emotions may occur in the absence of conscious awareness and therefore cannot be identified through introspection. However, the nonconscious affective states can still drive consumers’ judgments and even consumption behaviours (Winkielman et al., 2005; Winkielman & Berridge, 2004). In our case, I believe that a surprise label in an ongoing consumption experience triggers a nonconscious threat response that is manifested in heightened anxiety among consumers, which generates avoidance behaviours. Prior research has indeed found that automatic negative emotions, such as anxiety, often drive behaviour rather than cognitive evaluations when facing risky choices (Loewenstein et al., 2001). Moreover, anxiety is a signal that the environment is uncertain and uncontrollable. Therefore, anxiety drives consumer’s decisions and behaviours that reduce risk and uncertainty (Raghunathan et al., 2006). Hence, during the ongoing consumption experience the automatic affective responses may overrule cognitive evaluations and guide consumer behaviour. In particular, the anticipation of a surprise will violate one’s sense of meaning and thus disrupts consumers’ ability to predict and control their world, resulting in uncertainty. The subsequent nonconscious anxiety thus manifests itself in increased risk avoidance behaviours, such as less consumption. Prior research has already found that primed stimuli can unconsciously evoke anxiety as a threat response among consumers, and that the anxiety in turn induces an increase in risk avoidance (Lowe et al., 2019). Furthermore, expectancy violations can indeed cause situational uncertainty and increase threat responses, such as avoidance from the threatening stimulus and even motoric freezing (Mendes et al., 2007). Therefore, I believe that the consumer may not be aware of the triggered threat response but still affects their consumption behavior through avoidance.

In sum, I expect that the aversive tension evoked by the prospective meaning violation which is accentuated by the perceived approaching surprise causes a nonconscious threat response in the ongoing consumption experience. In particular, the accentuated uncertainty of the prospective meaning violation and the subsequent experienced nonconscious anxiety can turn into avoidance behaviour of the perceived threat. Therefore, surprise labels decrease the amount consumed. In the next section, I will discuss an alternative manner in which consumers can alleviate this aversive tension.

2.3 Fluid Compensation

(11)

reconstruct meaningful associations. In other words, people have the natural tendency to maintain a sense of meaning and thus are motivated to resolve inconsistencies (Taylor & Noseworthy, 2020; Heine et al., 2006). Meaning is defined as “the expected relationships that allow us to make sense of our experiences” (Proulx & Inzlicht, 2012). Hence, incongruency prompts that schematic expectations are violated because the experience is inconsistent with prior established relational understandings. Incongruency can be resolved through employing assimilation, implicating within the activated product schema. If this is impossible incongruency can be resolved through employing accommodation, implicating modifying the product schema or developing a new product schema (Piaget, 1960). Fluid compensation is the third way a person can resolve expectancy violations and restore meaning, in the sense that reaffirming a readily available established schema can make up for the inability to resolve another. Hence, when people’s sense of meaning is sufficiently threatened through unexpected experiences, people will often nonconsciously attempt to compensate by reinforcing pre-existing beliefs in unrelated domains, such as brand dominance, rather than resolving the violations itself (Taylor & Noseworthy, 2020; Proulx & Inzlicht, 2012; Heine et al., 2006). A threat to a consumer’s meaning is any experience that is inconsistent with the expected relationships. The source of the threat could therefore be that the consumer feels uncertain and is unable to predict future events (Proulx & Inzlicht, 2012; Heine et al., 2006). For example, this would be the case when consumers are encountering a prospective surprise that cannot be resolved yet. Thus, surprise labels would create the need to restore meaning through the process of fluid compensation.

Fluid compensation is fuelled by the arousal-induced anxiety when our sense of meaning is threatened and aims to alleviate the aversive tension by regaining meaning in other domains than that of the expectancy violation (Taylor & Noseworthy, 2020; Proulx & Inzlicht, 2012; Proulx & Heine, 2008). Moreover, fluid compensation is a nonconscious coping strategy driven by nonconscious anxiety (Randles, Proulx, & Heine, 2011; Proulx & Heine, 2008). This mechanism is in accordance with product schema incongruity, which suggests that extreme incongruity automatically heightens physiological arousal, that in turn evokes anxiety resulting in aversive behaviour (Noseworthy et al., 2014). Prior research has indeed found that if extreme product incongruity sufficiently threatens expectations so that it induces anxiety, it in turn creates a nonconscious need to regain meaning by reaffirming pre-existing schemata in unrelated consumption domains. At that instant, extreme product incongruity enhances compensatory behaviour by increasing the preference for products that can affirm unrelated beliefs, such as dominant brands (Taylor & Noseworthy, 2020). Similarly, I propose that surprise labels trigger an arousal-induced anxiety below our awareness and that we cope with this aversive tension by compensatory consumption in unrelated domains, such as brand dominance.

(12)

labels enhance fluid compensation, such that they increase the preference for dominant brands to compensate for the prospective loss of meaning and to elevate the aversive tension. In the next section, I discuss the moderating role of self-affirmation on the relationship between surprise labels and consumption behaviours.

2.4 Self-affirmation – Alleviating the nonconscious threat response

In a similar vein to the theory of fluid compensation, positive self-affirmation in an unrelated domain reduces the nonconscious threat response that is evoked by the surprise label in another domain. A threat can thus be reduced through the process of self-affirmation. People are motivated to maintain an overall worth and integrity of the self. When people’s self-concept is threatened, people seek a way to restore their global positive self-image. The indirect affirmation of alternative aspects of the self to protect their self-worth in face of a threat, enables people to cope with a threatening event and lowers defensive behaviours (Sherman & Cohen, 2006; McQueen & Klein, 2006; Sherman & Cohen, 2002; Steele, 1988).

(13)

2.5 Hypotheses

From the expectations the following hypotheses are proposed:

H1. The presence of a surprise label (vs. a regular label) decreases product consumption.

H2. The presence of a surprise label (vs. a regular label) increases compensatory consumption.

H3A. Self-affirmation moderates the relationship between a surprise label and product

consumption, such that self-affirmation before encountering a surprise label weakens the negative effect of a surprise label on product consumption.

H3B. Self-affirmation moderates the relationship between a surprise label and

compensatory consumption, such that self-affirmation before encountering a surprise label weakens the positive effect of a surprise label on compensatory consumption.

The conceptual model of the expected relationships between the independent and dependent variables and the moderator are depicted in figure 2.1.

Figure 2.1 Conceptual model

3. RESEARCH DESIGN

(14)

is increased. However, the need of compensatory consumption could also be alleviated when the consumer can affirm their positive self-image beforehand.

3.1 Method

Participants and design. To test the hypotheses, an online study in Dutch and English using

the survey software Qualtrics was conducted. The Dutch participants were recruited via snowball sampling by sharing the experiment online via social media in April 2020. I recruited 255 participants online in return for participation in a lottery of a gift card of €30 from an online retailer. Participants were randomly assigned to one of the four conditions in a 2 (surprise label vs. regular label) x 2 (self-affirmation vs. no self-(self-affirmation) between-subjects factorial design. The data was analysed using SPSS, such that two-way ANOVA analyses were conducted to find the interaction effects and group differences on the continuous dependent variables. Prior to the analyses, 32 participants who failed the attention check (see procedure section) were excluded. Moreover, four participants failed to follow the instructions from the self-affirmation manipulation. Two of them reflected on a negative characteristic or value and the other two did not want to elaborate on their positive characteristic or value. Lastly, after checking for outliers on consumption another participant was excluded (see figure A.1 in appendix A). These 37 cases were excluded leaving a final N of 218 (68.8% females; Mage = 25.34,

SD = 7.68). Due to unfinished sessions of the study, the conditions sizes were not equal. In the surprise label condition with self-affirmation N was 45, in the surprise label condition without self-affirmation N was 64, in the regular label condition with self-affirmation N was 46 and in regular label condition without self-affirmation N was 63. This affected the assumption of equal variance between samples and thus affected the statistical power. Hence, a larger sample size would have given more reliable results with greater precision and power.

3.2 Stimuli

Surprise and consumption. In order to simulate an ongoing consumption experience with a

(15)

Self-affirmation. Participants were either positively self-affirmed before the photo album

browsing experience or not. Prior research found that reflecting upon one’s most important characteristics or values is an effective manner of activating self-affirmation (McQueen & Klein, 2006; Steele, 1988). In a similar vein to study one of Cohen, Aronson and Steele (2000), participants in the self-affirmation condition completed a writing task about a value they chose to be personally important. However, my study asked participants to come up with a personal characteristic or value on their own, such as sense of humour or creativity, since a review of experimental manipulations of self-affirmation showed that it is also common and effective to let participants identify their own important aspects (McQueen & Klein, 2006). This reflective writing task served as the self-affirmation manipulation. In particular, participants were asked to describe one to two personal experience(s) where the selected value had been important to them and made them feel good about themselves. The value they wrote about should have been in all cases be unrelated to a surprise. The participants in the control condition wrote about a time when they did groceries, the chores or another ordinary task around the house as a filler task to avoid any opportunity to self-affirm (Cohen et al., 2000). I propose that self-affirmation before encountering a surprise label weakens the negative effect of a surprise label on the time spent browsing and weakens the positive effect on the preference for dominant brands.

Compensation. For this study, I based the stimulus of compensatory consumption on prior

research that used brand dominance (Taylor & Noseworthy, 2020). I define brand dominance as the strength of the association between a brand and the related product category. It has been shown that when consumers’ expectations are violated as a result of extreme incongruency, that the subsequent anxiety state elicits compensatory behaviours, such as reaffirming the highly accessible schemata of brand dominance (Taylor & Noseworthy, 2020). In other words, they unconsciously compensate by favouring category dominant brands in unrelated domains. Thus, I propose that if a surprise label triggers an aversive tension, then these participants are more likely to compensate by preferring dominant brands than participants exposed to a regular label. Similar to the compensation task of Taylor and Noseworthy (2020) I selected five product categories: soup, smartphone, battery, vacuum cleaner and coffee. Each product category consisted of 10 brands that were adapted to the Dutch market and the option “Other” if the preferred brand did not appear in the list. Within the 10 brands there was one brand that had the strongest category dominance. Hence, the compensation level was measured by the preference for this dominant brand in each category. A pre-test (N = 26) was conducted to determine the dominants brands for the selected product categories in the Dutch market. The pre-test used convenience sampling to recruit the participants (34.6% females; Mage =

(16)

categories with one clear identifiable dominant brand: Unox (soup), Duracell (battery), Philips (vacuum cleaner) and Douwe Egberts (coffee). For the smartphone category Apple and Samsung were named equally. Therefore, for this product category there were two dominant brands. Next to online brand rankings, frequent named brands in the pre-test were included in the choice set for each category.

3.3 Procedure

(17)

participants were then confronted with the following attention check: “Please show that you pay attention to this study and answer (1) not all to this question.” (1 = not at all, 7 = very much). Participants then performed a filler task where they had to find as many words as possible in a word search puzzle within three minutes. After the filler task, participants were asked to respond to the following question for five product categories in the context of their shopping behaviour in general: “Consider if you were going to buy a product from [Insert product category name] today. Which brand would be your top choice (please select ‘Other’ if your preferred brand does not appear in the list)?”. In reality, this served as the compensation exercise. The five product categories were soup, battery, vacuum cleaner, smartphone, and coffee. For each category, participants were given a choice set consisting of 10 brand names plus an “Other” option. The order of the product categories and the category brands were randomized. For each category, participants chose their brand of preference. The study concluded with basic demographic questions and questions about the study experience itself. Before the participants were thanked for their participation, they were asked if they had participated in a similar study before.

4. RESULTS

First the results on product consumption are reported, followed by the findings on compensatory consumption.

4.1 Results consumption

(18)

label, F(1,214) = .01, p = .91, and self-affirmation, F(1,214) = .01, p = .94, had no significant effects on consumption. A surprise label did not significantly lead to a lower consumption (M = 44.82, SD = 31.84) than a regular label (M = 44.69, SD = 35.35). Moreover, self-affirmation did not significantly lead to a higher consumption (M = 44.96, SD = 32.41) than no self-affirmation (M = 44.61, SD = 34.50). An additional two-way ANOVA analysis was carried out without excluding participants and another additional analysis with also excluding participants who participated in a similar study before (N = 99) (see appendices C and D for the analyses summaries). Neither analysis did reveal a significant change.

Lastly, one-way ANOVA analyses were conducted to rule out that surprise labels change the level of expected and experienced enjoyment (see appendix E for the analyses summaries). First, a two-way ANOVA was conducted to establish that there was no significant interaction between surprise label and self-affirmation on expected enjoyment, F(1,214) = 0.07, p =.80. Next, the Levene’s test of the one-way ANOVA on expected enjoyment confirmed that the assumption of homogeneity of variance was met, F(1,216) = 2.96, p =.09. There was no significant difference between the product labels on expected enjoyment, F(1,216) = 2.66, p = .10. Another one-way ANOVA was conducted to see if surprise labels change the level of experienced enjoyment after consumption. First, a two-way ANOVA was conducted to establish that there was no significant interaction between surprise label and self-affirmation on experienced enjoyment, F(1,214) = 1.71, p =.19. The Levene’s test of the one-way ANOVA on experienced enjoyment confirmed that the assumption of homogeneity of variance was met, F(1,216) = .15, p =.70. There was also no significant difference between the product labels on experienced enjoyment, F(1,216) = .03, p = .86.

(19)

4.2 Results compensation

The dependent variable of compensation was first coded as binary for each category (1 = dominant brand, 0 = non-dominant brand). Then, the sum of the total categories was taken to count the number of chosen dominant brands across the five product categories. Compensation was considerately normally distributed, with a skewness of -.04 (SE = 0.17). A two-way ANOVA was conducted to analyse the influence of a surprise label and self-affirmation on compensation in terms of participants’ dominant brand preferences. Given that there were multiple brands the participants could choose from, their choices were not independent. Hence, a multilevel binary logistic regression was deemed appropriate as it permits the use of repeatedly measured categorical-dependent variables (Taylor & Noseworthy, 2020). However, this analysis was beyond the scope of this research. The next optimal choice was a Poisson regression. However, the distribution of the count of the chosen dominant brands across the five product categories did not follow a Poisson distribution (see table F.1 in appendix F for the distribution analysis). As a consequence, the mean and variance of the model were not identical. Hence, the Poisson model violated the assumption of equidispersion through underdispersion. A Poisson regression was run to predict the number of dominant brands based on the type of product label and self-affirmation (see table F.2 in appendix F for the analysis summary). Indeed, the Pearson Chi-Square/df (= .46) was smaller than one, indicating that the Poisson regression model might not have been well suited for this data, apparently because of the underdispersion in the data count. Moreover, the addition of the independent variables did not generate a significant overall model (p = .44).

(20)

smartphone category with the two dominant brands was taken out altogether (see appendix H for the analysis summary). There was no significant change. Hence, including the smartphone category with two dominant brands did not cause any issues. As previously, an additional two-way ANOVA analysis was carried out without excluding participants and another additional analysis with also excluding participants who participated in a similar study before (N = 99) (see appendices I and J for the analyses summaries). Without excluding participants, the main effect of self-affirmation became not significant, F(1,251) = 2.48, p = .12. Also, with excluding participants who indicated that they participated in a similar study before, the main effect of self-affirmation became not significant, F(1,152) = 2.82, p = .10.

Figure 4.2 Impact surprise labels on compensationin the average number of chosen dominant brands across the five product categories by self-affirmation (note. no significant interaction).

(21)

5. DISCUSSION

Companies often engage in communicating unexpectedness in their brands’ marketing communication, which could involve the use of surprise labels. This unexpected-framing can create more positive attitudes towards the product and thereby increase product preference (Wadhwa et al., 2019; Lee & Mason, 1999). One question that arises is, if brands explicitly label surprises in their marketing communication, would this also be positive for the ongoing consumption experience? The present study tried to answer this question.

(22)

The results of the study did not show any evidence in support of the prediction that consumers will engage in more consumption when they face a surprise label and have the opportunity to positively affirm their image prior than when they cannot (H3A). One could argue that the chosen self-affirmation manipulation was not successful in alleviating the threat response to a surprise label. Therefore, it showed no significant difference in consumption levels if the person had the opportunity to self-affirm or not before being exposed to a surprise label. However, there was also no main effect of surprise labels on consumption levels found. This could mean that self-affirmation was unnecessary, since the participants seemed not to engage in defensive behaviours as an answer to exposure to a surprise label. Self-affirmation has namely been found to reduce defensive behaviours, such as avoidance, to threatening events (Finley et al., 2018; Sherman & Cohen, 2006). Hence, it is difficult to determine if the current manipulation of self-affirmation was effective or not.

The results of the study did not support the prediction that consumers will compensate for the prospective loss of meaning and elevate the aversive tension by favouring more dominant brands when facing surprise labels (H2). Specifically, the experience of a surprise label did not lead to an increase in the choice of dominant brands in unrelated product categories than the experience of a regular label. An explanation could be that people’s sense of meaning was not sufficiently threatened for them to nonconsciously feel the need to compensate by reinforcing pre-existing beliefs in unrelated domains, since the need to restore meaning through the process of fluid compensation is present when people experience a threatening meaning violation (Taylor & Noseworthy, 2020; Heine et al., 2006; Proulx & Inzlicht, 2012). Another possibility is that surprise labels do not cause the same extent of threat to our sense of meaning as extreme product incongruity does. Consequently, the insufficient amount of arousal-induced anxiety does not turn into the need to seek meaning in unrelated consumption domains. Therefore, extreme product incongruity does enhance compensatory behaviour by increasing the preference for dominant brands (Taylor & Noseworthy, 2020).

(23)

of a threatening surprise label or a regular label. Future research is needed to explain the reason behind this finding and if it can be replicated at all.

Although the findings did not correspond with my predictions, this research still makes several meaningful contributions. From a theoretical perspective, this research contributes to the surprise literature by demonstrating that consumers do not consume less due to the possible threat from a surprise label during the consumption stage. This adds to previous literature by broadening the unexpected-framing effects beyond the focus on consumer product preference and attitude (Wadhwa et al., 2019; Lee & Mason, 1999) to the actual consumption experience. In particular, this study found that the surprise effect does not exist negatively nor positively during consumption. Moreover, the current research contributes to the literature on affirmation and surprise by showing that self-affirmation does not seem to moderate the effect of a surprise label on consumption levels nor compensatory consumption. Hence, the process of self-affirmation to cope with a threatening event that elicits defensive responses did not seem to work in this domain. Therefore, it could be that self-affirmation is only effective when the event is directly threatening to the self-concept. The findings of this research also contribute to the self-affirmation literature by providing some initial evidence of the role of self-affirmation on the preference of dominant brands. Furthermore, the current work contributes to the fluid compensation literature by demonstrating that not all meaning violations, such as a prospective surprise, necessarily translate in compensating behaviours in unrelated consumption domains. Especially, consumers did not affirm other meaningful aspects of consumption, such as preference for dominant brands, after experiencing a possible tension from a surprise label. Hence, consumers did not employ fluid compensation as a means of alleviating this aversive tension.

(24)

when using surprise labels. Lastly, positively affirming the consumer’s self-image by triggering the consumer to think about a universal positive value that makes them feel good about themselves, is not needed in the marketing communication when marketers are considering surprise framing. However, self-affirmation did increase the choice for dominant brands. This could be beneficial for marketers who are looking to revitalize sales of dominant brands.

It is important to pay attention to some of the potential limitations of this study. Firstly, the study was based on a small convenience sample. Thus, the sample is not representative and does not allow generalisations to the general population. Hence, it could be that there was an insufficient sample size in order to conclude a valid research result. Secondly, a small part of the sample had participated in a similar study before. Although the purposes of the similar studies were not shared, the repetition might have affected the findings. Thirdly, the design in which the data was collected might have limited the ability to find the predictions. In this research I conducted the self-affirmation manipulation by asking participants to self-identify a positive value or characteristic important to them and describe an accompanied experience. It could be argued that the employed manipulation was not successful in generating a positive self-image, resulting in no significant findings. A manipulation check could have been included. Moreover, compensatory behaviour was measured by the semantic association of category dominant brands. It could be possible that this measure failed to measure compensatory behaviour and contained product categories with ambiguous dominant brands. However, the measure was adopted from prior research and thus it seemed valid to measure compensatory behaviour with the pre-identified dominant brands. Another possible explanation could be that the time between the exposure to the threatening surprise label and the compensation task with a distractor task in-between was too long to have an effect. The participants could have been reminded that they consumed a product with a surprise label before performing the compensation task. Furthermore, several remarks from the participants regarding the study described that the photo album was not representational. For example, the photos were not from a real vacation but were instead stock photos and the photo album itself did not mirror the feel of a real photo album. This could have had the effect that the threat of the surprise label was not realistic enough. The stimulus could be improved into a more representative photo album. Lastly, the data was collected under extremely special circumstances of a nationwide lockdown due to the global pandemic of COVID-19. This might have impacted the findings in an unforeseen way. For example, the generally higher levels of anxiety could have made it difficult to manipulate anxiety levels any further. When interpreting the results, these limitations must be borne in mind and the results must be interpreted with caution.

(25)

that measured consumption as the total browsing time of the photo pages. For example, future research may explore the consumption of food, such as chocolates, with a surprise label on the packaging that measures the amount of chocolates eaten. Examining this possibility would increase the generalizability of this study to other consumption domains. Other future research could investigate other compensatory acts of fluid compensation when facing a surprise label. For example, consumers may compensate for a meaning violation by favouring culturally relevant products to affirm their cultural identity (Taylor & Noseworthy, 2020). This would allow us to see if the findings are replicable. Moreover, future research might also investigate consumer responses to a product with an actual surprise (e.g., unexpected benefits) during the consumption stage. In this research I focused on a surprise that was merely communicated and never happened. Hence, any sense-making of the surprise could not take place. This would be different with an actual surprise. After sense-making of an actual surprise, the valence of the surprising outcome itself is more likely to be attributed to the response (Noordewier & van Dijk, 2019; Noordewier et al., 2016; Noordewier & Breugelmans, 2013). Then, people may experience a positive surprise as pleasant or a negative surprise as painful. The induced arousal during a surprise may carry over and amplify the subsequent affective responses (Valenzuela et al., 2010; Noordewier & Breugelmans, 2013; Mellers et al., 1997). Therefore, the experienced value of the stimulus might be enhanced compared to receiving the expected experience. Prior research suggests that an unexpected outcome could activate an approach or an avoidance motivation, depending on the valence of the outcome. When expectancy violations outcomes are positive, consumers anticipate that there are other potential rewards resulting in approach behaviours. However, when the expectancy violations outcomes are negative, it cues other potential losses and thus enhances avoidance behaviours (Wadhwa et al., 2019). Thus, future research is needed to explore if consumption levels and enjoyment indeed change after sense-making of an actual surprise. Moreover, future research could measure the underlying processes that drive consumers’ consumption when faced with surprise labels and the effect of self-affirmation. The physiological response of arousal could be measured by looking at cardiovascular measures, such as blood pressure and heart rate. Self-affirmation is namely found to reduce elicited physiological responses (Taylor et al., 2003). Lastly, it would be interesting to further examine the effect of self-affirmation on the preference for dominant brands, as in this study a significant effect was found for this possible relationship. This finding raises questions for future research to explore if this indeed is true and why. Overall, one can imagine that more research is needed in the area of surprises.

(26)

(27)

REFERENCES

Bechara, A., & Damasio, A. R. (2005). The somatic marker hypothesis: a neural theory of economic decision. Games and Economic Behavior, 52(2), 336-372.

Cian, L., Krishna, A., & Elder, R. (2015). A sign of things to come: Behavioral change through dynamic iconography. Journal of Consumer Research, 41(6), 1426-1446.

Cohen, G., Aronson, J., & Steele, C. (2000). When beliefs yield to evidence: Reducing biased evaluation by affirming the self. Personality and Social Psychology Bulletin, 26(9), 1151-1164. Finley, A., Crowell, A., & Schmeichel, B. (2018). Self-affirmation enhances processing of negative stimuli among threat-prone individuals. Social Cognitive and Affective Neuroscience, 13(6), 569-577. Gawronski, B., & Strack, F. (2012). Cognitive consistency: A fundamental principle in social cognition. New York etc.: Guilford Press.

Heckler, S., & Childers, T. (1992). The role of expectancy and relevancy in memory for verbal and visual information: What is incongruency? Journal of Consumer Research, 18(4), 475-492.

Heine, S., Proulx, T., & Vohs, K. (2006). The meaning maintenance model: On the coherence of social motivations. Personality and Social Psychology Review: An Official Journal of the Society for

Personality and Social Psychology, Inc, 10(2), 88-110.

Hsee, C., Tu, Y., Lu, Z., & Ruan, B. (2014). Approach aversion: Negative hedonic reactions toward approaching stimuli. Journal of Personality and Social Psychology, 106(5), 699-712.

Kahneman, D., & Miller, D. (1986). Norm theory: Comparing reality to its alternatives. Psychological Review, 93(2), 136-153.

Kunst-Wilson, W. R., & Zajonc, R. B. (1980). Affective discrimination of stimuli that cannot be recognized. Science, 207(4430), 557-558.

Lee, Y., & Mason, C. (1999). Responses to Information Incongruency in Advertising: The Role of Expectancy, Relevancy, and Humor. Journal of Consumer Research, 26(2), 156-169.

Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as feelings. Psychological bulletin, 127(2), 267.

Lowe, M., Loveland, K., & Krishna, A. (2019). A quiet disquiet: Anxiety and risk avoidance due to nonconscious auditory priming. Journal of Consumer Research, 46(1), 159-179.

Mandler, G. (1982). The structure of value: Accounting for taste. In M. S. Clark, & S. T. Fiske (Eds.), Affect and cognition: The seventeenth annual carnegie symposium on cognition (p. 3–36). Hillsdale, NJ: Lawrence Erlbaum.

McQueen, A., & Klein, W. (2006). Experimental manipulations of self-affirmation: A systematic review. Self and Identity, 5(4), 289-354.

Mellers, B., Schwartz, A., Ho, K., & Ritov, I. (1997). Decision affect theory: Emotional reactions to the outcomes of risky options. Psychological Science, 8(6), 423-429.

(28)

Meyer, W., Reisenzein, R., & Schützwohl, A. (1997). Toward a process analysis of emotions: The case of surprise. Motivation and Emotion, 21(3), 251-274.

Meyers-Levy, J., & Tybout, A. (1989). Schema congruity as a basis for product evaluation. Journal of Consumer Research, 16(1), 39-54.

Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: Affective priming with optimal and suboptimal stimulus exposures. Journal of Personality and Social Psychology, 64(5), 723-739. Noordewier, M. K., & Breugelmans, S. M. (2013). On the valence of surprise. Cognition &

Emotion, 27(7), 1326-1334.

Noordewier, M. K., & van Dijk, E. (2019). Surprise: Unfolding of facial expressions. Cognition & Emotion, 33(5), 915-930.

Noordewier, M., Topolinski, S., & Van Dijk, E. (2016). The temporal dynamics of surprise. Social and Personality Psychology Compass, 10(3), 136-149.

Noseworthy, T. J., Muro, F. D., & Murray, K. B. (2014). The role of arousal in congruity-based product evaluation. Journal of Consumer Research, 41(4), 1108-1126.

Piaget, J. (1960). General problems of the psychological development of the child. In J. M. Tanner & B. Inhelder (Eds.), Discussion on child development, proceedings of the world health organization study group on psychological development of the child (Vol. 4, p. 3-27). New York, NY: International Universities Press.

Proulx, T., & Heine, S. (2008). The Case of the Transmogrifying Experimenter: Affirmation of a Moral Schema Following Implicit Change Detection. Psychological Science, 19(12), 1294-1300.

Proulx, T., & Inzlicht, M. (2012). The five "a"s of meaning maintenance: Finding meaning in the theories of sense-making. Psychological Inquiry, 23(4), 317-335.

Raghunathan, R., Pham, M., & Corfman, K. (2006). Informational properties of anxiety and sadness, and displaced coping. Journal of Consumer Research, 32(4), 596-601

Randles, D., Proulx, T., & Heine, S. (2011). Turn-frogs and careful-sweaters: Non-conscious perception of incongruous word pairings provokes fluid compensation. Journal of Experimental Social

Psychology, 47(1), 246-249.

Rumelhart, D.E. (1980) Schemata: the building blocks of cognition. In R.J. Spiro et al. (Eds.), Theoretical Issues in Reading Comprehension, Hillsdale, NJ: Lawrence Erlbaum.

Sherman, D., & Cohen, G. (2002). Accepting threatening information: Self-affirmation and the reduction of defensive biases. Current Directions in Psychological Science, 11(4), 119-123.

Sherman, D., & Cohen, G. (2006). The psychology of self-defense: Self-affirmation theory.Advances in Experimental Social Psychology, 38, 183-242.

Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the self. In L. Berkowitz (Ed.), Advances in experimental social psychology, Vol. 21. Social psychological studies of the self: Perspectives and programs (p. 261-302). Academic Press.

Taylor, N., & Noseworthy, T. (2020). Compensating for innovation: Extreme product incongruity encourages consumers to affirm unrelated consumption schemas. Journal of Consumer

(29)

Taylor, S., Lerner, J., Sherman, D., Sage, R., & McDowell, N. (2003). Are self-enhancing cognitions associated with healthy or unhealthy biological profiles? Journal of Personality and Social Psychology, 85(4), 605-15.

Townsend, S. S. M., Eliezer, D., & Major, B. (2013). The embodiment of meaning violations. In K. D. Markman, T. Proulx, & M. J. Lindberg (Eds.), The psychology of meaning (p. 381-400). American Psychological Association.

Valenzuela, A., Mellers, B., & Strebel, J. (2010). Pleasurable surprises: A cross-cultural study of consumer responses to unexpected incentives. Journal of Consumer Research, 36(5), 792-805. Wadhwa, M., Kim, J. C., Chattopadhyay, A., & Wang, W. (2019). Unexpected-Framing Effect: Impact of Framing a Product Benefit as Unexpected on Product Desire. Journal of Consumer Research, 46(2), 223-245.

Winkielman, P., & Berridge, K. (2004). Unconscious emotion. Current Directions in Psychological Science, 13(3), 120-123.

Winkielman, P., Berridge, K., & Wilbarger, J. (2005). Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality & Social Psychology Bulletin, 31(1), 121-35.

Winkielman, P., Zajonc, R. B., & Schwarz, N. (1997). Subliminal affective priming resists attributional interventions. Cognition and Emotion, 11(4), 433-465.

(30)

APPENDICES

(31)

Appendix A – Identification of outliers

(32)

Appendix B – Two-way ANOVA summary for consumption Table B.1 Descriptive statistics of consumption in seconds

Self-affirmation Product label M SD N

No self-affirmation Regular label 42.92 32.27 63

Surprise label 46.27 36.73 64

Total 44.61 34.50 127

Self-affirmation Regular label 47.13 39.43 46

Total 44.96 32.41 91

Total Regular label 44.69 35.35 109

Total 44.76 33.57 218

Table B.2 Two-way ANOVA summary for consumption

Source SS df MS F p

Self-affirmation 6.36 1 6.36 0.01 0.94

Surprise label 13.79 1 13.79 0.01 0.91

Self-affirmation * Surprise label 791.93 1 791.93 0.70 0.41

Error 243690.88 214 1138.74

(33)

Appendix C – Two-way ANOVA summary without excluding participants Table C.1 Descriptive statistics of consumption in seconds without excluding participants

Total 46.93 38.78 142

Total 44.63 35.20 113

Total 45.91 37.18 255

Table C.2 Two-way ANOVA summary for consumption without excluding participants

Source SS df MS F p

Surprise label 159.24 1 159.24 0.11 0.73

Error 348012.86 251 1386.51

Note. R2_{= .009, Adj. R}2 _{= -.003.}

(34)

Appendix D – Two-way ANOVA summary with excluding participants similar study Table D.1 Descriptive statistics of consumption in seconds with excluding participants similar study

Total 44.68 32.78 90

Total 48.02 32.72 66

Total 46.09 32.69 156

Table D.2 Two-way ANOVA summary for consumption with excluding participants similar study

Source SS df MS F p

Surprise label 1360.06 1 1360.06 1.26 0.26

Error 163788.84 152 1077.56

Note. R2_{= .011, Adj. R}2 _{= -.008.}

(35)

Appendix E – One-way ANOVA summaries for expected and experienced enjoyment Table E.1 Descriptive statistics of expected enjoyment on a scale of 1 to 100

Product label M SD N

Regular label 59.62 21.33 109

Total 61.83 20.09 218

Table E.2 One-way ANOVA summary for expected enjoyment

Source SS df MS F p

Between groups 1065.71 1 1065.71 2.66 0.10

Within groups 86530.35 216 400.60

Total 87596.06 217

Table E.3 Descriptive statistics of experienced enjoyment on a scale of 1 to 100

Product label M SD N

Regular label 66.96 17.43 109

Total 67.17 17.68 218

Table E.2 One-way ANOVA summary for experienced enjoyment

Source SS df MS F p

Between groups 9.29 1 9.29 0.03 0.86

Within groups 67789.43 216 313.84

(36)

Appendix F – Poisson regression summary for compensation

Table F.1 One-sample Kolmogorov-Smirnov test for Poisson distribution of compensation Compensation

N 218

Poisson parameter Mean 2.48 Most Extreme Differences Absolute 0.11

Positive 0.10

Negative -0.11

Kolmogorov-Smirnov Z 1.67

Asymp. Sig. (2-tailed) 0.01

Table F.2 Poisson regression summary for compensation

Goodness of Fit Value df Value/df

Deviance 115.48 214 .54

Pearson Chi-Square 98.16 214 .46

Omnibus test Likelihood Ratio Chi-Square df p

2.72 3 .44

Source Wald df p

Intercept 447.43 1 .00

Surprise label .37 1 .55

Self-affirmation 2.32 1 .13

(37)

Appendix G – Two-way ANOVA summary for compensation

Table G.1 Descriptive statistics of compensation in total number of chosen dominant brands

Total 2.34 1.08 127

Total 2.67 1.03 91

Total 2.48 1.07 218

Table G.2 Two-way ANOVA summary for compensation

Source SS df MS F p

Surprise label 0.95 1 0.95 0.84 0.36

Error 241.56 214 1.13

(38)

Appendix H – Two-way ANOVA summary without smartphone category

Table H.1 Descriptive statistics of compensation in total number of chosen dominant brands without smartphone category

Total 1.54 0.94 127

Total 1.84 0.98 91

Total 1.66 0.97 218

Table H.2 Two-way ANOVA summary for compensation without smartphone category

Source SS df MS F p

Surprise label 0.69 1 0.69 0.75 0.39

Error 197.39 214 0.92

Note. R2_{= .027, Adj. R}2 _{= .013.}

(39)

Appendix I – Two-way ANOVA summary without excluding participants

Table I.1 Descriptive statistics of compensation in total number of chosen dominant brands without excluding participants

Total 2.32 1.07 142

Total 2.54 1.14 113

Total 2.42 1.11 255

Table I.2 Two-way ANOVA summary for compensation without excluding participants

Source SS df MS F p

Surprise label 1.06 1 1.06 0.87 0.35

Error 307.21 251 1.22

Note. R2_{= .015, Adj. R}2 _{= .003.}

(40)

Appendix J – Two-way ANOVA summary with excluding participants similar study

Table J.1 Descriptive statistics of compensation in total number of chosen dominant brands with excluding participants similar study

Total 2.36 1.12 90

Total 2.65 1.05 66

Total 2.48 1.10 156

Table J.2 Two-way ANOVA summary for compensationwith excluding participants similar study

Source SS df MS F p

Surprise label 0.15 1 0.15 0.12 0.73

Error 183.40 152 1.21

Note. R2_{= .019, Adj. R}2 _{= .000.}

(41)

The Unexpected Surprise in Consumption: How

Surprise Labels Affect Consumption Behaviors

Jildou Bouma

(42)

(43)

(44)

•

Meaning maintenance model (Proulx & Inzlicht, 2012; Heine et al.,

2006)

•

Schema congruity theory (Taylor & Noseworthy, 2020;

Noseworthy et al., 2014)

•

Arousal-induced anxiety and behavioural avoidance

03 Fluid compensation

01 Meaning Violation

04 Self-affirmation

•

Fluid compensation (Proulx & Inzlicht, 2012; Heine et al.,

2006)

•

Brand dominance (Taylor & Noseworthy, 2020)

•

Restores meaning and reduces the aversive tension

•

Positive self-affirmation (Finley et al., 2018; Sherman & Cohen,

2006)

•

Reduces defensive behaviors

02 Ongoing consumption experience

•

Approach aversion (Hsee et al., 2014)

•

Somatic marker hypothesis (Bechara & Damasio, 2005)

•

Affective primacy hypothesis (Winkielman et al., 2005; Zajonc,

1980)

(45)

(46)

(47)

Impact surprise label on compensation in the average number of chosen dominant brands across the five product categories by self-affirmation (note. no significant interaction)

(48)

(49)

• No decrease in consumption levels

• No increase in compensatory

consumption

surprise labels are

not threatening

Self-affirmation was

not effective

• No altering of consequences

of surprise labels

• Increases preference for

dominant brands

The Practical

implications

• Not harmful for the

consumption experience

• Increases product desire

• No loss of sales non-dominant

brands

• The employment of

(50)

CREDITS: This presentation template was created by Slidesgo, including icons by Flaticon, and

infographics & images by Freepik and illustrations by Stories

THANKS

(51)

• Bechara, A., & Damasio, A. R. (2005). The somatic marker

hypothesis: a neural theory of economic decision.Games and

Economic Behavior, 52(2), 336-372.

• Finley, A., Crowell, A., & Schmeichel, B. (2018).

Self-affirmation enhances processing of negative stimuli among

threat-prone individuals. Social Cognitive and Affective

Neuroscience, 13(6), 569-577.

• Heine, S., Proulx, T., & Vohs, K. (2006). The meaning

maintenance model: On the coherence of social motivations.

Personality and Social Psychology Review: An Official Journal of the Society for Personality and Social Psychology, Inc, 10(2), 88-110.

• Hsee, C., Tu, Y., Lu, Z., & Ruan, B. (2014). Approach aversion: Negative hedonic reactions toward approaching stimuli.

Journal of Personality and Social Psychology, 106(5), 699-712.

• Noseworthy, T. J., Muro, F. D., & Murray, K. B. (2014). The role

of arousal in congruity-based product evaluation. Journal of

Consumer Research, 41(4), 1108-1126.

• Proulx, T., & Inzlicht, M. (2012). The five "a"s of meaning maintenance: Finding meaning in the theories of

sense-making.Psychological Inquiry, 23(4), 317-335.

• Sherman, D., & Cohen, G. (2006). The psychology of

self-defense: Self-affirmation theory. Advances in Experimental

Social Psychology, 38, 183-242.

• Taylor, N., & Noseworthy, T. (2020). Compensating for

innovation: Extreme product incongruity encourages

consumers to affirm unrelated consumption schemas.Journal

of Consumer Psychology, 30(1), 77-95.

• Winkielman, P., Berridge, K., & Wilbarger, J. (2005).

Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality & Social Psychology Bulletin, 31(1), 121-35.

• Zajonc, R. B. (1980). Feeling and thinking: Preferences need