The impact of emotions in service encounters
A meta-analysis on the emotions’ influence on consumers
evaluations
MSc Business Administration: Marketing track
University of Amsterdam, Amsterdam Business School
June, 2017
Student Edoardo De Iuliis
Id number 11386142
Msc Business Administration: Main track in Marketing
email: edo.uliis@hotmail.it
Statement of Originality
This document is written by Edoardo De Iuliis who declares to take full
responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and
that no sources other than those mentioned in the text and its references have
been used in creating it.
The Faculty of Economics and Business is responsible solely for the
supervision of completion of the work, not for the contents.
ABSTRACT 1. INTRODUCTION 1 2. THEORETICAL FRAMEWORK 2.1. Antecedents 3 2.1.1. Expectations 2.1.2. Affect 2.2. Consequences 8 2.2.1. Customer satisfaction 9 2.2.2. Customer loyalty 10 2.2.3. Complaint behavior 11 2.2.4. Negative WOM 13 2.2.5. Industry moderators 14
2.2.6. Moderators based on sample size 15
2.2.6.1. Students 2.2.6.2. Gender 2.2.6.3. Industry moderators 5. DATA GATHERING 19 6. METHODOLOGY 24 7. RESULTS 26 8. CONCLUSIONS 34 8.1. Discusion 34 8.2. Findings 35 8.2. Theoretical implications 35
8.3. Managerial implications and future research 39
10. REFERENCES 43
Abstract
The growing interest towards the influence of emotions on service encounters is producing a good amount of publications in such regard. However, the mixed reported findings, complicate the understanding of the interaction between emotions and service encounters. These findings and the need for managers of having a satisfied clientele leave space for a meta-analytical review of the empirical research. This thesis documents the diverse effects that emotions, and thus affect, have on post-purchase behaviors. In doing so several moderators are taken into account. In conclusion, managerial implications on these effects are presented, offering directions for future research.
Keywords:
Introduction
Economics and business studies classically focus on human rationality and the consumer’s capability to make rational choices. Although these models are extremely useful in understanding and summarizing complex situations and scenarios, their capability to explain reality is limited. In the case of economics, for example, many authors have argued that consumer choices are not based on rationality, but more on behavior and heuristics. Likewise, business studies focused attention on the different elements that define consumer choices and responses in different situations. Edvardsoon (2004) described service encounters as a tripartite process in which both cognitive, affect and behavior influences the nature of consumer reaction to service encounters. These elements are nowadays considered fundamental, even though in the past most academic attention was given to the understanding of cognitive reaction, both rational and irrational (e.g. heuristics).
Academically, there has been a substantial increase in the scientific recognition of the influence of feelings and emotions on the rules of thinking and behavior (Pham, 2004). Emotions and their behavioral effects have acquired importance and recognition in the last decades. Nowadays, emotions, and their effects on consumer behavior, represent an important field of study and are the object of interest in many scientific subjects and studies. From business to politics, their effect on people’s behavior, consumer behavior and decision-making are studied and analyzed. Accordingly, Lerner and Keltner (2000) concluded that emotions have a specific influence on judgement and choice, making them interesting elements for Business and Marketing Studies. Following their theory, it is clear that disposition emotions play a large role for consumers. They have the power to shape consumer judgement and, subsequently, to influence consumer choices in many different forms.
Emotions therefor play and will play a key role in defining commercial and business strategies for business-to-consumer industries. It appears clear that, due to the ongoing transformation in the retailing world, retailers and managers, from any kind of industry, from fashion firms to financial institutions, will have to borrow many ideas, theories and organizational structure from “emotional” providers (such as hospitality sector, or the museum industry). This will be done in order to make the consumer journey a playful and educative process about the offered good and services of the physical store.
The first meta-analysis studying customer satisfaction, “Customer Satisfaction: A Meta-Analysis of the Empirical Evidence” was published in 2001. By the author’s own admission, times were not mature for research to systematically study, through the use of a meta-analysis, the effect of emotions. Due to this limitation, the emotional influence was included as a moderator in a wider research regarding customer satisfaction. Nowadays, despite the relatively small amount of research regarding these elements, the time is right to undertake the first meta-analysis regarding the influence of emotional status on consumer outputs.
The goal of this study is to advance the understanding of affect and their effects on consumer’s behavior. By conducting a meta-analysis on the role of emotions in service encounters, a quantitative synthesis of what is known about the dynamics of consumers outputs will be produced. The aim is to understand to which extent affect is able to influence post-purchase behaviors. In particular, which are the industries where affect plays a major role and which are the post-purchase constructs influenced by it. At the same time, the hope is to create a fruitful and useful framework for other researchers, in order to better understand and synthesize this particular topic.
To fulfill this purpose, a summary of the rationale beyond the antecedents, the outcomes, and the potential moderators of behavioral outputs in service encounters will be presented. Afterwards, the methodology behind this process will be explained. Subsequently, the findings from a statistical analysis of 24 empirical studies reporting 103 correlations involving consumer output in service encounters are presented. This meta-analysis will be concluded by discussing the implications of the findings and the limitations of this research. Finally, new directions for future research will be explained.
Literature Review
Antecedents of Post-Purchase Behaviors
Current literature identifies a series of elements that influence consumer behavior. Among them, expectations (and subsequently disconfirmation) play an important role in defining and explaining the cognitive elements behind consumer behavior (Szymanski and Henard 2001). Besides that, another important element defining behavioral outcomes is affect. As a matter of fact, service encounters are described as a tripartite process, in which cognitive, emotional and behavioral elements result in the creation of a mental mark or memory (Edvardsson et al. 2005, p. 151).
Expectations.
According to current literature, it is clear that expectations have a clear role in triggering post-purchase behaviors. There are two main roles that expectations can fulfill: anticipative and comparative (Szymanski et al 2001). When expectations have a role in anticipation, the general idea is that they are able to influence post purchase levels (Szymanski et al 2001). The first one happens when consumers have no previous experience with a product or service and are thus not able to compare with pre-existing evaluations and assessments (Oliver 1988). This leads consumers to anticipate experiences and create expectations around a product or service (Szymanski et al 2001). These expectations are then used as a point of reference for future satisfaction assessments (Oliver 1993). Consequently, dissonance plays an important role as well. Indeed, consumers tend to avoid it by modifying their satisfaction levels in order to stay consistent with the expectations they had of a product or service. The majority of empirical research supports the existence of a positive relationship between expectations and consumer satisfaction (Oliver and Linda 1981; Swan and Trawick 1981). Indeed, the more positive the expectations are, the more positive the satisfaction will be.
Second, expectations play a comparative role when consumers already have a point of reference in their purchase history. The disconfirmation paradigm plays an important role in this configuration. Disconfirmation can first be positive if the assessment of the service exceeds expectations (positive disconfirmation), negative if the assessment of the service is inferior to the expectation (negative disconfirmation), and neutral if the actual consumer experience matches its expectations (Oliver, 1981).
Affect
Disconfirmation theory has been largely used as a theoretical framework to explain different service constructs (Oliver, 1980). Disconfirmation theory explains consumer behavior through the analysis of cognitive processes. The theory explains customer satisfaction as an element that depends on three factors: expectations, perceived performance and confirmation/disconfirmation of precedents beliefs. Thereafter, the theoretical framework behind disconfirmation theory, is centered on the cognitive processes of the post-purchase behavior. Unfortunately, little attention has been given to the influence of emotions in the process and formation of post-purchase behaviors (Westbrook, 1987). Nevertheless, it is possible to highlight how a focus on cognitive perspectives can overshadow other perspectives and neglect important aspects of post-purchase behaviors. For instance, there is a growing body of research investigating the effects of affect on consumer’s behavior, and thus on post-purchase decisions (Szymanski et al 2001). The term affect encompasses a broad group of specific mental processes (Szymanski et al 2001). These processes can be divided into three different categories: emotions, moods and attitudes. Bagozzi defines emotion as a “mental state that arises from cognitive appraisals of events or thoughts” (1999). Emotions are linked as well to physiological processes as they can generate physical expressions by affecting posture, gestures and facial expressions. The nuance between emotions and moods can be difficult to define and understand. Indeed, moods usually last for a relatively long period of time compared to emotions but with lower intensity, while emotions are usually short-lived and intense. Finally, attitudes are considered instances of affect. Some authors define them as being evaluative judgments in terms of good or bad reactions rather than truly emotional states (Bagozzi et al. 1999). The idea of emotions, the definition of them and the correct way to study them, have widely been debated among management and marketing scholars (Plutchik 1994). Clore (1987) suggested that marketing science should analyze and understand emotions, in order to better understand the processes behind consumer’s judgment and consumer’s reaction to events. Afterwards, some authors define the role of emotions as a fundamental element in order to analyze and understand the consumption experience (Oliver 1997).
Several studies (e.g., Mano and Oliver 1993; Westbrook and Oliver 1991), analyzed the role that affect plays in post-purchase judgments. These studies suggest that emotions aroused during service encounters are able to leave affective traces. Specifically, these traces are
fundamental for post-purchase judgments, since they are integrated into post-purchase assessments (Westbrook and Oliver 1991).
Besides the Pleasure, Arousal and Dominance model (PAD Model) (Mehrabian & Russell 1974), the first comprehensive studies regarding the influence of emotions on consumer behavior were made in the 70’s and 80’s. Gerdner (1985) identified service encounters as one the key areas for fruitful mood research, and Morris (1989) defined the power of mood in an element able to alter our mental processes. This power resides in the capability that affect has in twisting cognitive processes. In doing so, these constructs are altered, with the subsequent generation of different and more powerful behavioral outcomes. This results in identifying affect as a key element to clearly understand customer’s behaviors (Morris 1989).
Normally customer emotions can be divided into two main groups, one characterized by positive emotions and the other side by negative (Tronvoll, 2011, Bagozzi et al. , 1999; Brainerd et al., 2008; Chaudhuri, 1998; Liljander and Strandvik, 1997; Machleit and Mantel, 2001)
A first school of thought, stated that positive emotions, and thus positive moods, lead consumers to more positive evaluations (Mano and Oliver 1993). Oliver (1997) suggests that consumers’ emotional responses during consumption experiences or service encounters, should be analyzed regarding the consequences of specific events. This theoretical interpretation of affect links emotions with the final behavioral outcome of the encounter, giving to emotions an important role as antecedent of the behavioral outcome.
A second school of thought stated that negative emotions are much more complex than positive ones. They account for the majority of the variance reported in several studies (Berenbaum et al., 1995; Diener et al., 1995; Watson and Clark, 1991). Due to the complexity of these second category of emotions, they are a subject of interest for this meta-analysis.
Positive Negative Trust Anger Joy Frustration Pleased Sadness Hopeful Fear Happy Shame Joyful Disappointed Pride Discontented Gratitud Depressed
Positive surprise Guilty
Excited Humiliated
Fulflled Desire of revenge
Contended Regret
Envious
Several articles and academic publications reported the effects that affect has on different behavioral outcomes. In the following table a summary of the literature regarding affect in consumer behavior can be found:
CONSTRUCT AUTHOR AND TITLE FOCUS FINDINGS Customer
satisfaction I Have Paid Less Than You! The Emotional and Behavioral Consequences of Advantaged Price Inequality
Katja Gelbrich, 2011
This article analyzes emotions and post -purchase consumer behaviors. Theory is based on appraisal theories of emotion and includes social comparison theory. An experiment (n = 272) and a field study (n = 261) are part of this research.
Positive emotions include happiness, gratitude, pride, and malicious joy; while negative emotions include pity, outrage, and guilt. Findings suggest that these emotions mediate the occurrence of post-purchase behaviors (i.e., customer satisfaction, loyalty, WOM referral, and WOM activity).
Satisfaction:
A Meta-Analysis of the Empirical Evidence David M. Szymanski and David H. Henard, 2001
A high number of academic studies on customer satisfaction were made in the last years. Due to mixed findings a meta-analysis is required in order to synthesizing the empirical evidence on customer satisfaction and to assess current knowledge.
Equity and disconfirmation are strongly related to customer satisfaction. It is also find that measurement and method factors characterizing the research have the power to moderate the relationship strength between satisfaction and its antecedents and outcomes.
Loyalty The impact of experiential consumption cognitions and emotions on behavioral intentions J. Enrique Bigné, Anna S. Mattila, Luisa Andreu, 2008
This study examines cognitive and affective antecedents and consequences of satisfaction in the context of hedonic services.
Findings analyze the importance of emotions in defining consumer responses to hedonic services. More specifically, is highlighted that pleasure is positively linked to satisfaction and loyalty.
Asymmetric effects of customer emotions on satisfaction and loyalty in a utilitarian service context Aude Rychalski, Sarah Hudson Sarah Hudson 2016
This study investigates the effects positive and negative emotions on customer satisfaction and customer loyalty in a utilitarian service setting.
Results shows that positive emotions influence satisfaction more strongly than negative emotions
Complaint
Behavior Negative emotions and their effect on customer complaint behavior Bård Tronvoll 2007
This paper's aim is to understand pre-complaint situations. In doing so, three different purposes are achieved: the identification a set of negative emotions, the examination of the patterns of these negative emotions and the links that exist between negative emotions and complaint behaviour.
Results show that the negative emotion of frustration is the best predictor for complaint behaviour.
Negative Word
of Mouth “Never Eat In ThatRestaurant, I Did!”: Exploring Why People Engage In Negative Word-Of-Mouth Communication Inge M. Wetzer, Marcel Zeelenberg, and Rik Pieters, 2007
Two studies explore this question and reveal that consumers pursue specific goals when engaging in Negative WOM
The results reveal that consumers who experience anger engage in N-WOM to vent feelings or to take revenge.
This reveals the functionality of specific emotions to WOM, and how goals for N-WOM are associated with these emotions.
Repurchase intention
Gratitude, Delight, or Guilt: The Role of Consumers’ Emotions in Predicting Post consumption Behaviors, Isabella Soscia 2007
This study analyzes the relationships between consumption emotions (gratitude, happiness, guilt, anger, pride, and sadness), and post-consumption behaviors (positive and negative word of mouth, repurchase intention, and complaint behavior).
Findings demonstrate that emotions are able to predict different post-consumption behaviors.
Post-purchase behaviors Customer Satisfaction
The term customer satisfaction, defines the fulfillment of the customer, regarding a firm or a product (Vesel and Zabkar, 2009, Wu, 2011) or the degree in which a product or a service has met a customer’s expectation (Akhtar et al., 2011). In today’s consumer behavior and marketing science, customer satisfaction is among the most important post purchase construct. It is in fact, a key element in order to understand other constructs that are considered direct consequences of customer satisfaction such as: complaint behavior, word of mouth, customer loyalty (Szymanski et al 2001). In a more practical perspective, customer satisfaction is among the most important factor on firms success and growth rate, it is worthy of study and scientific attention (Kanning and Bergmann, 2009; Hoq and Amin, 2010).
Among the different definitions that were given for satisfaction, Kotler and Keller defined customer Satisfaction as “a person's feelings of pleasure or disappointment that results from comparing a product's perceived performance or outcome with his/her expectations” (Kotler and Keller, 2009, p. 789). Moreover, Johnson and Fornell (1991) defined customer satisfaction as, “A customer’s overall evaluation of the performance of an offering to date”. These definitions are clearly based on cognitive elements (Although Kotler and Keller define it as a “feeling,” giving to it an emotional status) and disconfirmation theory, linking the construct with expectations and performance.
Although cognitive elements play an important role in the definition of customer satisfaction, they incorporate a strong and important affective element generated through repeated usage of the product or the service (Oliver 1999). This affective component, present in customer satisfaction, is worthy of scientific and academic attention. Nowadays, the emotional/cognitive nature of customer satisfaction is still debated among authors. Scholars wonder whether customer satisfaction can be defined and studied as an emotional component (Westbrook, 1980) or a cognitive construct, that is in some way influenced by affect (Bagozzi et al., 1999; Crooker and Near, 1998, Brooks 1995). Nevertheless, research has shown on several occasions that emotional status and moods have an influence on it, and that, by a large extent, are important antecedents of customer satisfaction (Dubé and Menon, 2000; Westbrook and Oliver, 1991). Oliver (1993) tried to pacify the two schools of thought. He suggested that emotions, although being subordinated to cognitive structures (disconfirmation theory), are the
elements that differentiate customer satisfaction from other constructs like service quality or product quality. In accordance to that, emotional statuses play a role in consumer behavior and in triggering consumers outputs. In particular, it is proposed that customer satisfaction, although being a cognitive construct, is positively influenced by emotional status and affect in general.
H1: Affect, specifically emotions, has a positive influence on consumer satisfaction .
Consequences of customer satisfaction
Customer Loyalty
Customer loyalty is defined by Oliver as “a deeply held commitment to rebuy or repatronizes a preferred product or service consistently in the future, despite situational influences and marketing efforts having the potential to cause switching behavior” (Oliver, 1997: p. 392). Both marketing and management literature show a strong link between customer loyalty and consumer satisfaction (Chen, 2012; Kumar et al., 2013; Suh and Yi, 2006). As a matter of fact, customer satisfaction is considered to be amongst the most prominent factors on customer loyalty (Oliver, 1993, Kanning and Bergmann, 2009; Hoq and Amin, 2010). Similarly to what was stated in the previous chapter, both the constructs of customer satisfaction and customer loyalty present affective (emotional) and cognitive components (Oliver, 1993; Liljander and Strandvik, 1997). As previously said, authors suggest that particular emotions, have an influence on human behavior, and thus, customer satisfaction (Oliver, 1993; Dubé and Menon, 2000; Yi-Ting Yu and Dean, 2001). In particular, following the previous pattern, Yi-Ting Yu and Dean, (2001) linked positive emotions to customer loyalty. In their theory is suggested that personal positive emotions create links with a firm or a service provider, and thus, lead consumers to develop a strong customer loyalty. Meanwhile, the effects that negative emotions have on consumers are the opposite, linking customer to negative post-purchase behaviors. Emotions are in fact able to create the necessary links that transform a repeat purchase behavior in a loyal attitude towards the brand or the firm. More specifically, during the process in which the consumer becomes loyal to the firm, two types of loyalty can be underlined. On one side cognitive loyalty, that is a belief-based form of loyalty (cost-benefit, performance) and in the aftermath, affective loyalty, that is an attitude based on cumulative satisfying experiences (pleasurable fulfillment) (Oliver 1993). These two form of loyalty are part of
larger process, in which cognitive loyalty is antecedent to affect loyalty. It is then possible to argue that loyalty is a construct activated by cognitive elements and then enhanced and influenced by satisfactory experiences and positive attitudes towards the preferred brand/firm (Beerli et al. 2004). Hence loyalty is a construct based on previous judgments, furthermore, is a tool that consumers have in order to avoid negative experiences (disconfirmation theory). Although its cognitive nature, it can be enhanced by affect and positive emotions in general. Following this theoretical framework, it is evident that positive emotions have a clear and defining role in the affirmation of customer loyalty. It is thanks them if repeat purchasing based on inertia becomes customer loyalty.
H2: Customer loyalty is positively influenced by positive emotions, and thus, affect.
COMPLAINT BEHAVIOR
Another important post-purchase behavior is complaint behavior. Complaint behavior is deeply linked to consumer satisfaction and is an important element and indicator for managers and firms (Tronvoll, 2011). The number of complaints is a clear indicator of the level of dissatisfaction towards the firm, the service or the product (Oliver 1997). For managers, the rate of complaint behaviors, or the willingness to complain regarding a given service, must be considered as a clear and evident outcome of dissatisfaction towards the firm.
Similarly to customer satisfaction, complaint behavior has a clear cognitive nature. The theoretical framework behind the concept is linked to expectations and, thus, to the disconfirmation theory. Although its cognitive nature, strong links with negative emotions can be detected (Oliver, 1987).
From a theoretical point of view, complaint behaviors has been widely discussed in literature. Among the others, Oliver (1987) gave to the construct some theoretical background. He defines it as one mechanism available to consumers for relieving cognitive dissonance when the consumption experience is not fully satisfying. He also argues that complaint behavior is clearly linked with dissatisfaction, being it, in given circumstances, the easiest tool that consumers have in order to express their dissatisfaction. In his words, dissatisfaction consists in “A customer experience that is lower to what was previously expected” Oliver (1987). Defining it as a tool, or
mechanism, that customers have in order to relieve cognitive dissonance in case their purchase, or consumption experience is not alined with the previous expectation. This definition, and interpretation of the phenomena, links the construct to the disconfirmation theory and thus, gives it a cognitive nature. However, other authors, gave to the construct a different connotation, in which cognitive aspects are mixed to affect aspects of the construct. Nyer (1999) for example defines it as “A mechanism for venting anger and frustration and a mechanism for initiating or seeking redress for failed consumption experiences”. Evidencing the importance that negative emotions such as anger and frustration play in triggering this behavioral outcome. Following this path, Bagozzi (1999) argues that emotions plays a fundamental role in the complaint process. Following his analysis, the antecedents of this post-purchase behavior are both cognitive and emotional. It is then possible to affirm that complaints do not only originate from dissatisfaction and therefore, cognitive mechanism are clearly not enough to understand the cause for customers to complain (Day, 1984; Singh and Pandya, 1991).
A difference is underlined between negative and positive emotions. In fact, the influx that positive emotions have on consumers is clearly effecting positive consumer behaviors, while negative emotions lead consumers to negative behaviors, one of them, probably one of the most important, is complaint behaviors (Bagozzi et al., 1999, Liljander and Strandvik, 1997).
Tronvoll (2011), similarly to Nyer, links this particular post-consumption behavior to negative emotions. In his analysis of negative emotions in post-purchase behaviors, he defines customer complaining behavior as a process that arise when the service experience that a customers has, is outside the “acceptance zone”. In this context, being outside their comfort zone, consumers are leaded to an emotional, rather than cognitive response. Emotions are in fact able to influence the storage, the organization and the retrieval of cognitive information (Ger, 1989; Nasby and Yando, 1982). In conclusion, following this theoretical background it is possible to affirm that, although complaint behavior arises when the expectations are not fulfilled (dissatisfaction theory), the nature of this construct is both cognitive, both emotional. Accordingly to what was said, it is worth of academic interest, to better investigate the influence that emotions have in this construct. The hypothesis for this construct is then the following:
The aim is to understand in which degree emotions have an influence in the arousal of complaint behavior.
2. Negative Word Of Mouth
The word of mouth phenomena has an unquestioned power on consumers behaviors and decisions. It as a tool that is significantly more effective than advertising in business’s communication (Day, 1971). However, the literature regarding WOM is narrow and limited to particular applications. This behavioral construct has a significant effect on consumer behavior. The first authors analyzing the phenomena were Bass (1969) and Moore (1995), who defined it as one of the most influential factors in sales growth and products diffusion process. WOM is defined as “an oral person to person communication between a receiver and a communicator whom the receiver perceives as non-commercial, regarding a brand, a product or a service” (Arndt, 1967). Jillian C. Sweeney et al. (1997) made a distinction between the two different forms of WOM: Positive and Negative.
When positive, word of mouth is able to have a strong power for the company’s performance in several aspects; it is able, when properly managed, to inform potential customers in an efficient and fast way. In doing so, it saves the company money and the consumer time, and finally, being offered by informal sources, it is independent and reliable. Thanks to the uncommercial source of this construct, WOM communication is perceived by most of the consumers with less skepticism than typical promotional efforts (Herr et al., 1991). Most of the existing research argues that positive word of mouth presents a cognitive structure (Herr et al., 1991).
On the other hand, when negative, word of mouth behavior can be considered a different form of complaint behavior that is positively correlated with dissatisfying experiences. This behavioral outcome is triggered under some conditions such as high severity in the service failure and high levels of social activity in the disappointed customer (Richins 1983). It offers customers a mechanism to overcome the tension and disappointment of a negative encounter and to demonstrate superior consumption standards to the others (Nyer 1999). Empirical evidence suggests that negative information, and subsequently negative WOM, plays a much greater role in consumers’
evaluation than other positive informations. Some authors suggested that, in comparison to positive word of mouth, negative WOM one has a stronger emotional influence (Tronvoll, 2011).
Finally, following the results from Wetzer et al (2007) we can define negative WOM as a construct that is deeply influenced by negative emotions, giving the possibility to affirm that negative WOM is among the constructs that are mostly influenced by affect, and emotions in general. Finally, in accordance to thus hypothesis, Wetzer (2007) argues that negative emotions are predictors of this construct. Their influence in scientifically proven, and, most importantly, the general WOM construct is extremely emotional when the outcome is negative (Wetzer (2007).
Due to these conditions, our second hypothesis will be the following:
H6: Negative WOM is positively influenced by negative emotions, and thus, affect.
Moderators of post-purchase behaviors
One of the main purposes of a meta-analysis is to recognize the effect of the different moderators in post-purchase process. These moderators include the industry and the subject population
Moderators based on the Industry
One of the most interesting moderators to study is the industry in which emotions are triggered. The goal is to understand and define which are the industries that the influence of affect, and thus emotions, has the most significant and influential role in the definition of customer outputs. This meta-analysis takes in account five different industries, however due to limitations in the sample, only three different industries will be taken in account. These industries are hospitality, retail, and education.
Accordingly to previous literature and research, the industry in which affect plays the major role is the hospitality industry (Jie Zhang et al 2010). By hospitality, a broad category of services is usually defined. These services include lodging, event management, amusement parks, transportation, restoration, recreation and further areas within the tourism industry (LuJun Su & Maxwell K. Hsu, 2013). The importance of affect is likely due to the importance of the role that
excitement and arousal play in this particular kind of industry (Quix, 2012). Consumers are receiving experiences, not a good or a proper service (Quix, 2012). In this scenario, it is undeniable that affect has a powerful impact on consumer’s outcomes. Therefore, through this analysis the following hypothesis is proposed:
H3: Among three industries (hospitality, Retail and Education), hospitality is the strongest moderator on the relationship between positive emotions and consumer’s outputs.
This being said, another industry that might be partially influenced by affect, and thus emotions is retail (Jie Zhang et al 2010, Amatulli & Guido, 2012). Retail is defined by the process of market goods or services to consumers via various channels of delivery in order to gain a profit (Amatulli & Guido, 2012). Due to the affirmation of the internet channel, several authors argued that the industry is undergoing several transformation, making it, one of the most dynamic and interesting industries to be studied. In the last decade, many scholars claimed that physical stores would be a remembrance of the past. Nowadays, the internet appears to serve as a facilitating technology in many domains, serving traditional business as a complementary tool, rather than a disruptive technology able to erase it (Jie Zhang et al 2010). This faith is motivated by the fact that the arousal of emotions in physical settings (e.g. stores) can achieve strong improvements in consumers’ outcomes and post-purchase behaviors such as satisfaction (Quix, 2012, Amatulli & Guido, 2012). Although, internet channels are bel to enhance operational efficiency and provided benefits to customers, they are not a substitute for traditional retailing. Traditional settings are able to provide certain unique emotional benefits that new channels are not able to provide (Ravi Kalakota, 2012). In particular, these benefits include: the potential use of the senses in product evaluation, the option of cash payment, and more importantly, research, the personal service, the entertainment and social experiences. These specific benefits, have a power in triggering post-purchase responses. In order to enhance these benefits, physical retailers will deal with different forms of emotional management. As argued by Jie Zhang et al (2009), it is clear that among the most valuable characteristics of physical stores there is the presence of personal service, that leads to human interactions. In fact, emotions are defined as "responses to outcomes of social interaction" (Kemper, 1978, p. 80) and “Social rather than isolated, individual processes” (Domagalski, 1999, p. 836). This enables us to say, the presence of others humans is the most valuable characteristic of physical retailing, enabling firms to provide customers with emotional benefits that in other channels are infrequently and less strong. In fact, the literature
suggests that the way in which individuals display emotions has a strong impact on the perceived quality of transaction and service encounters (Jie Zhang et al 2009).
Emotional management creates new directions, and management paths for retailing management. These new opportunities consist of the possibility, thanks to a wise use of the emotions, to offer great experiences to their costumers (Amatulli & Guido, 2012, p. 193). Managers should focus on retail stimuli, the environment, and all those characteristics that could lead the customer to a strong and positive emotional involvement. For example, as many industries are already doing: adding educational labs, that enhance curiosity (Quix, 2012), Design furniture, that causes arousal (Quix, 2012), perfumes (Jie Zhang et al. 2010), or pop iconography (Quix, 2012). The first examples of these trends were introduced in the 80’s.
In conclusion, the influence of affect, most specifically of emotions, in the retail industry is going to be tested. The general prevision is that, emotions, although having a smaller effect in the retail industry rather than in the hospitality one, have a positive influence on post-purchase behaviors.
H4: As a moderator of the relationship between positive emotions and consumer’s outputs, Retail presents a positive effect, lower, but similar to the one expected in the hospitality
Industry.
Moderators based on study characteristics
Moderators based on sample size
Concerns were raised among different researchers about the reliability of student-based findings among consumer’s population (Burnett and Dunne 1986; Park and Lessig 1977). This is due to a series of characteristics that define this type of sample. As a matter of fact, students have different characteristics from the rest of the population, among others: modest income, young age, small experiences, different needs. Although being a very atypical group, they are present in large proportions in scientific studies (Park and Lessig 1977). These characteristics of this segment group of the population suggest that they could account for some variance in our analysis.
H7: The relationship between emotional states and post-purchase behaviors is moderated by characteristics of the researched sample: Non-students, as compared to students are more
effected by emotions.
Gender
An interesting element to analyze is the difference between the two different genders. Thanks to the collected data, it is possible to analyze if the two genders presents different correlations.
H8: The relationship between emotional states and post-purchase behaviors is moderated by characteristics of the researched sample: males, as compared to females are more effected by
Overview of the studies included in this meta analysis:
Author Journal Title
Post-Purchase Behaviors
Industry Country Emotions
1 LuJun Su &
Maxwell K. Hsu (2013)
Journal of Travel & Tourism Marketing, 30:8, 786-805
Service Fairness, Consumption Emotions, Satisfaction, and Behavioral Intentions: The Experience of Chinese Heritage Tourists SAT REVISIT WOM ALTERNAT HOSP CHINA PE NE 2 Davoud Nikbin et
al. (2015) Asia Pacific Journal of Tourism Research,
2015
The Determinants of Customers' Behavioral Intentions after Service Failure: The Role of Emotions
NWOM AIRTRAVE MALASYA NE
3 LuJun Su et al
(2015) Journal of Travel & Tourism Marketing,
31:8, 1018-1038
Understanding the Relationship of Service Fairness, Emotions, Trust, and Tourist Behavioral Intentions at a City Destination in China
REVISIT,
WOM HOSP CHINA PE, NE, TRUST
4 Katja Gelbrich
(2011) Journal of the Academy of
Marketing Science (2010) 38:567–585
I Have Paid Less Than You! The Emotional and Behavioral Consequences of Advantaged Price Inequality
SAT, LOYALTY, WOM
RETAIL GERMANY HAPPY, GRATITUD, PROUD, SELFPITY, OUTRAGE, GUILTY, MAL_JOY
5 Heesup Han PhD &
Ki-Joon Back PhD (2007)
Journal of Hospitality & Leisure Marketing, 15:3, 5-30
Investigating the Effects of Consumption Emotions on Customer Satisfaction and Repeat Visit Intentions in the Lodging Industry
SAT HOSP USA PE, NE
6 Vera Pedragosa et
al (2015) Motriz, Rio Claro, v.21 n.2, p. 116-124,
Apr./Jun. 2015
The role of emotions on consumers’
satisfaction within the fitness context SAT PLEASURE PORTUGAL PE
7 Aude Rychalski Sarah Hudson (2016) Journal of Business Research 71 (2017) 84–91
Asymmetric effects of customer emotions on satisfaction and loyalty in a utilitarian service context
SAT, LOYALTY, SAT, WOM
SERVICE FRANCE PE, NE, TRUST
8 Esther Gracia,
Arnold B. Bakker, and Rosa M. Grau
Positive Emotions: The Connection between Customer Quality Evaluations and Loyalty - Study 1
LOYALTY HOSP SPAIN
9 Tronvol (2011) Journal of Service
Management, Vol. 22 Issue: 1, pp.111-134
Negative emotions and their effect on
customer complaint behaviour COMPL HOSP NORWAY ANGER, FRUST, SADNESS, FEAR, SHAME
10 Johnson et al.,
(2009) Managing Service Quality: An
International Journal, Vol. 19 Issue: 1, pp. 4-30,
Negative emotions and their effect on
customer complaint behaviour SAT HOSP NORWAY JOY, DISAP
11 Bignè et al., (2008) Journal of Services
Marketing, Vol. 22 Issue: 4, pp.303-315
The impact of experiential consumption cognitions and emotions on behavioral intentions
SAT, LOYALTY, WTPM
HOSP SPAIN PLEASED,
DISCONT
12 Menon & Dubé
(2004) Journal of Retailing 80 (2004) 229–237 Service provider responses to anxious and angry customers: different
challenges, different payoffs
SAT AIRTRAVE USA ANGER
13 Yu & Dean (2001) Negative emotions and their effect on customer complaint behaviour
The contribution of emotional
satisfaction to consumer loyalty LOYALTY, WOM, COMPLAIN , SWITICHIN , LOYALTY, WTPM EDU AUSTRALI PE 14 Liljander &
Strandvik (1996) International Journal of Service Industry
Management, 8(2), 148–169.
Emotions in service satisfaction SAT SERVICES FIN ANGER, DEPRESSED, HUMILIAT, HAPPY, SURPRISE
15 Christopher White ;
Yi-Ting Yu (2005) Journal of Services Marketing, Vol. 19
Issue: 6, pp.411-420
Satisfaction emotions and consumer
behavioral intentions WOM, COMPLAIN , SWITCHIN, WTPM, WOM
EDU SWISS ANGER, HAPPY,
HOPEFUL, SURPRISE, DEPRESSED, GULTY, DISAPPOINTED , REGRET 16 Sánchez-García, Isabel Currás-Pérez, Rafael (2 Tourism Management
32 (2011) 1397e1406 Effects of dissatisfaction in tourist services: The role of anger and regret SWITCH, COMPLAIN , SWITCH, WOM
HOSP SPAIN ANGER, REGRET
17 Kabadayi, Alan (2012) International Conference on Leadership, Technology and Innovation Management
Revisit Intention of Consumer Electronics Retailers: Effects of Customers’ Emotion, Technology Orientation and WOM Influence
WOM,
REVISIT RETAIL TURKEY PE, NE
18 Derbaix, Christian ; Vanhamme, Joëlle (2003) Journal of Economic Psychology 24 (2003) 99–116
Inducing word-of-mouth by eliciting
surprise – a pilot investigation WOM RETAIL BELGIUM JOYFUL, NE
19 Grégoire, Yany et al
(2010) Journal of the Academy of
Marketing Science (2010) 38:738–758
A comprehensive model of customer direct and indirect revenge: understanding the effects of perceived greed and customer power
NWOM,CO MPLAIN, MARKAGR E
SERVICES USA DESREVEN,
BLAME, ANGER
20 White, Christopher
J. (2010) Journal of Marketing Management, 26(5/6),
381–394.
The impact of emotions on service quality, satisfaction, and positive word-of-mouth intentions over time.
WOM, SAT EDU AUS PE, NE , BID
21 Dubè, Morgan
(1998) International Journal of Research in
Marketing 15 1998. 309–320
Capturing the dynamics of in-process consumption emotions and satisfaction in extended service transactions
SAT HEALTHCA CANADA PE, NE
22 Soscia (2007) Psychology &
Marketing, Vol. 24(10): 871–894
Gratitude, Delight, or Guilt: The Role of Consumers’ Emotions in Predicting Postconsumption Behaviors COMPLAIN NWOM REPINT PWOM REPINT
RETAIL ITALY ANGER
GUILT HAPINESS SADNESS ANGER GRATITUD PRIDE 23 Concepción
Varela-Neira et al (2008) The Service Industries Journal, 28:4, 497-512 The influence of emotions on customer's cognitive evaluations and satisfaction in
a service failure and recovery context
SAT BANK SPAIN NE
24 Girish Prayag,
Sameer Hosany1, Birgit Muskat, and Giacomo Del Chiappa
Journal of Travel
Research Understanding the Relationships between Tourists’ Emotional Experiences, Perceived Overall Image, Satisfaction, and Intention to Recommend
SAT HOSP UK/ITALY JOY
LOVE SURPRISE
Journal Title
Post-Purchase Behaviors
Industry Country Emotions Author
Data gathering
In order to estimate the effictiveness of emotion’s effects introduced through this thesis, several steps were needed to first create and build a complete database of the emotion in relation with service encounters’ findings.
During this process, the first step consisted in defining the inclusion criteria in order to combine the findings of different studies in this meta-analytical review.
The different publication candidates for inclusion were empirical studies that stipulated a relationship between emotional outcomes and post-purchase constructs. More specifically, studies who identify post-purchase behaviors as measured variables, and emotions as predictors of those behaviors. The main emotions defined are the following: Trust, Anger, Joy
Frustration, Pleased, Sadness, Hopeful, Fear, Happy, Shame, Joyful, Disappointed, Pride, Discontented, Gratitude, Depressed, Positive surprise, Guilty, Excited,Humiliated, Fulfilled, Desire of revenge , Contended, Regret, Envious
Among the main consumer outputs identified, the following: Switching providers, Complaint
Behavior, Customer Satisfaction, Customer Loyalty, Word of Mouth, Revisit intentions, Negative Word of Mouth and Willingness to pay more.
Primary studies to be used in this meta-analysis were identified thanks to a keyword search of different electronic databases: Google Scholar, UvA library, Web of Science. The key words used in this search were the following: Emotions, Customer Satisfaction, Affect, Wom,
Negative Emotions, Loyalty, Positive Emotions, Anger, Service Encounters, Switching Behavior, Retail, Hospitality, Customer Loyalty. In addition to that, researches were carried
out in regards to the references found in the accessible studies, and manual searches, of the most prominent academic journals reporting studies on customer behavior. To name some: the Journal of the Academy of Marketing Science; the Journal of Consumer Research; the Journal of Customer Satisfaction, Dissatisfaction and Complaining Behavior; the Journal of Marketing; the Journal of Marketing Research; Management Science; and Marketing Science. Finally, two professors were contacted in order to obtain additional papers on satisfaction. Unfortunately this request was not successful.
The search process started at the beginning of March 2017 and ended in mid-May, when the developed database was mature enough to be analyzed.
Only experimental studies which investigated the relationship between a given emotion and a subsequent dependent variable (behavioral outcome) were included in the collection of studies considered valuable for the meta-analysis. In addition to that, another important inclusion criteria was whether there was a clear description of the statistical methodologies behind the research and whether the objective of the study was in line with this research.
Subsequent to the collection of studies on customer satisfaction, the step was to identify a common measure of association (correlation, regression coefficient, etc.) that would allow the highest percentage of results to be incorporated into this meta-analysis. After an attentive evaluation, the correlation coefficient was designed as common effect size. The correlation coefficient presents several advantages: it is employed quite often throughout the literature to describe emotional relationships in consumer behavior, and it is the metric that allows many consumer satisfaction findings to be converted into valuable effect sizes (see McGaw, and Smith 1981). Finally, the most important characteristic of the correlation coefficient is that it preserves the continuous properties of the satisfaction measure and its correlates.
However, not all the academic publications reported correlations or similar measures that could be transformed into correlations. Due to time constrains, these articles were not included in the meta-analysis, leaving 10 articles out of the database.
The 25 studies are composed of 24 published studies and 1 dissertation. The final used database presents a total of 103 correlations (effect sizes) involving an emotion as predictor, and a behavioral outcome as outcome. Among the studies, 43 effect sizes were not coded.
The following characteristics were coded for every effect size. Yes stands for a presence of the element whilst no stands for the absence of it.
Whether the stimuli used to refer to the service encounter was:
-Recal stimuli (1 = yes, 0 = no)
-Real stimuli (1 = yes, 0 = no)
In which environment the study took place:
-Field (1 = yes, 0 = no)
-Lab (1 = yes, 0 = no)
-Online (1 = yes, 0 = no)
If the emotion was measured or manipulated (1 = yes, 0 = no)
If the emotions were measured as trait (1 = yes, 0 = no)
If the emotion was measured with regard to the feeling in that very moment but without direct
link to the service encounter (1 = yes, 0 = no)
If the emotion was explicitly measured with regard to the concrete service encounter (1 = yes, 0 = no)
If the study was an experiment (1 = yes, 0 = no)
If it was a student sample (1 = yes, 0 = no)
If a service failure was present (1 = yes, 0 = no)
Demographics:
Age - If applicable each effect size reported the medium age of the sample
Size - If applicable each effect size reported the size of the sample
Gender percentage - If applicable each effect size reported the percentage of males present in the sample, the variable is defined “P_males”
Emotions
Whether the measured emotion was Positive (1 = yes, 0 = no)
Whether the measured emotion was Negative (1 = yes, 0 = no)
Whether the measured emotion was Happy (1 = yes, 0 = no)
Whether the measured emotion was Gratitude (1 = yes, 0 = no)
Whether the measured emotion was Pride (1 = yes, 0 = no)
Whether the measured emotion was Self pity (1 = yes, 0 = no)
Whether the measured emotion was Anger (1 = yes, 0 = no)
Whether the measured emotion was Anxiety (1 = yes, 0 = no)
Whether the measured emotion was Regret (1 = yes, 0 = no)
Whether the measured emotion was Guilty (1 = yes, 0 = no)
Whether the measured emotion was Malicious Joy (1 = yes, 0 = no)
Whether the measured emotion was Frustration (1 = yes, 0 = no)
Whether the measured emotion was Humiliated (1 = yes, 0 = no)
Whether the measured emotion was Hopeful (1 = yes, 0 = no)
Whether the measured emotion was Positive Surprise (1 = yes, 0 = no)
Whether the measured emotion was Joyful (1 = yes, 0 = no)
Whether the measured emotion was Desire of revenge (1 = yes, 0 = no)
Whether the measured emotion was Fear (1 = yes, 0 = no)
Whether the measured emotion was Sadness (1 = yes, 0 = no)
Whether the measured emotion was Shame (1 = yes, 0 = no)
Whether the measured emotion was Joy (1 = yes, 0 = no)
Whether the measured emotion was Disappointed (1 = yes, 0 = no)
Whether the measured emotion was Pleased (1 = yes, 0 = no)
Whether the measured emotion was Discontent (1 = yes, 0 = no)
Whether the measured emotion was Blame (1 = yes, 0 = no)
1 if the measurement of the emotion focused on the frequency instead of the intensity (1 = yes, 0 = no)
Dependent variables:
Whether the dependent variable was Customer Satisfactio n (1 = yes, 0 = no)
Whether the dependent variable was Customer Loyalty (1 = yes, 0 = no)
Whether the dependent variable was Word of Mouth (1 = yes, 0 = no)
Whether the dependent variable was Revisit intention (1 = yes, 0 = no)
Whether the dependent variable was Negative Word of Mouth (1 = yes, 0 = no)
Whether the dependent variable was Marketplace aggression (1 = yes, 0 = no)
Whether the industry was:
Hospitality industry (1 = yes, 0 = no)
Retail industry (1 = yes, 0 = no)
Services industry (1 = yes, 0 = no)
Education industry (1 = yes, 0 = no)
Air Services industry (1 = yes, 0 = no)
Healthcare industry (1 = yes, 0 = no)
Whether there was online exposure (1 = yes, 0 = no)
Whether the geographical area was:
Europe (1 = yes, 0 = no)
USCAD (Usa + Canada) (1 = yes, 0 = no)
Asia (1 = yes, 0 = no)
Method
A meta-regression is a meta-analytical instrument that allows researchers to code a multitude of primary studies in order to compute an overall effect size evaluation (e.g., Glass, 1976; van den Noortgate & Onghena, 2003). It presents several advantages; in our case, the main advantage consists of the possibility to systematically investigate and understand how moderator variables influence the effect sizes, and to obtain a valid result regarding a specific topic. In particular, our aim is to understand how a given moderator (e.g. industry or gender) influences the overall outcome (consumer behavior). In addition to that, a meta-analytical procedure allows researcher and scholars to understand and investigate the current state of the literature, and thus, to obtain interesting information about possible research gaps and future research directions.
In order to carry out a meta-analytical process, a common value or unit of measurement has to be identified and selected. In this case, effect size is going to be the correlation coefficient.
Linear regression based on dummy variables
In order to perform our meta-regression we are going to employ a simple linear regression model.
The dependent variable Yj (named coefficient_size) consists in the different correlation coefficients that were extracted from different publications.
Generally, when in a linear meta-regression the outcome is a positive and significant coefficient means that the study characteristic is associated with a larger value of the dependent variable. On the other side, if such coefficient is negative it will decrease the dependent variable. The coefficient shows the shift in the dependent variable when the independent variable changes one unit. Dummy variables show the effect of being part of a certain category. The constant shows the expected effect when all dummy variables are zero (Hardy, 1993).
In the following model, seven variables are presented. These variables are: P MALES, that express the percentage of males present in the sample.
NON STUDENTS, Is a dummy variable that express the percentage of non students present in the
sample: 1 = non student, 0 = student
REAL, that express the percentage of studies based on field observations present in the sample, 1 = based on recall or real stimuli or , 0 = based on vignette
POSITIVE EMOTIONS that express the positive emotions coded from primary studies (further
information about the particular emotions will be included in the thesis), 1 = Positive Emotions, 0 = Negative Emotions
SATISFACTION, that express the number of customer satisfaction observation present in primary
studies, 1 = Customer Satisfaction is observed, 0 = Customer Satisfaction is not observed .
HOSPITALITY, EDUCATION and RETAIL, that express the different industries moderators
present in this analysis: 1 = The study was realized in the particular industry that defines the dummy variable, 0 = The study wasn’t realized in the particular industry that defines the dummy variable.
DATA ANALYSIS
In order to perform this meta-analysis, the technique of linear meta-regression is applied. As a first step, data is analyzed in order to understand different properties. Descriptive statistics can be found in table (I). A 95% confidence interval was established for the coefficient which plays the role of the dependent variable. No cases had missing observations.
In the following models the effect size is going to be used as a dependent variable. As previously said, the effect size is composed by correlation coefficient of the different studies. In order to test the hypothesis, and understand the relationship among the different variables, a regression is carried out.
Case Processing Summary
Cases Valid Missing
N Percent N
r 103 100,0% 0
Descriptives Statistic Std. Error
r Mean 0,2022 0,04259
95% Confidence Interval for Mean
Lower Bound 0,1177 Upper Bound 0,2866 5% Trimmed Mean 0,2154 Median 0,2900 Variance 0,187 Std. Deviation 0,43225 Minimum -0,79 Maximum 0,90 Range 1,69 Interquartile Range 0,81 Skewness -0,503 0,238 Kurtosis -0,936 0,472
Customer Satisfaction and Industry Moderators
* significant p < 0,05, ** significant p < 0,010
The variables that are introduced are the followings:
CUSTOMER SATISFACTION: Is a dummy variable that shows the presence of customer satisfaction in the effect size (1 = present, 0 = not present)
Coefficients Regression II Models Unstandardized Coefficients B Std. Error 1 (Constant) -0,012* 0,046 SATISFACTION -0,183* 0,070 POSITIVE EMOTIONS 0,592* 0,064 2 (Constant) -0,085 0,048 SATISFACTION -0,186* 0,066 POSITIVE EMOTIONS 0,585* 0,060 HOSPITALITY 0,217* 0,063 3 (Constant) -0,077 0,059 SATISFACTION -0,19* 0,068 POSITIVE EMOTIONS 0,589* 0,062 HOSPITALITY 0,209* 0,071 RETAIL -0,018 0,079 4 (Constant) -0,004 0,071 SATISFACTION -0,215* 0,069 POSITIVE EMOTIONS 0,601* 0,062 HOSPITALITY 0,138* 0,081 RETAIL -0,095* 0,090 EDUCATION -0,183* 0,103 Models R R Square 1 ,688a 0,473 2 ,729b 0,531 3 ,729c 0,531 4 0,739d 0,546
POSITIVE EMOTIONS: Is a dummy variable that shows the presence of positive emotions in the effect size (1 = present, 0 = negative emotions)
HOSPITALITY: Is a dummy variable that shows the presence of hospitality industry in the effect size (1 = present, 0 = not present)
RETAIL: Is a dummy variable that shows the presence of retail industry in the effect size (1 = present, 0 = not present)
EDUCATION: Is a dummy variable that shows the presence of the education industry in the effect size (1 = present, 0 = not present)
The first regression takes into account two main variables, customer satisfaction and positive emotions. The dummy variable “Positive emotions” has been built through the integration of different “positive emotions” as stated in the literature review. In this variable, emotions such as happiness, joy, hope, love, and the more generic Positive_emotions are reported. The second dummy variable computed in the regression is “Customer_Satisfaction”. This variable represents significant outcomes of “customer_satisfaction” presented among the different studies that are processed in this meta-regression. It is important to underline the correlation coefficients present for both positive and negative values.
The model confirms that customer emotions and positive emotions explain 48,7% of the effect size. The effect size measures the correlation between inputs and outputs, which in the analyzed case are emotional states and customer behavior or customer intentions. This first analysis found positive emotions to have a positive impact on the effect size (that explains the correlation between an input and an outcome), (B = 0,587), and customer satisfaction to have a negative impact on the effect size (B = - 0,198). This means that positive emotions increase the correlation, having a positive effect size on the dependent variable. In addition to that, it is interesting to analyze the impact of customer satisfaction. The result of this analysis might suggest that correlation is stronger for non-satisfied customer rather than satisfied customers. This result may be explained by different factors:
Negative emotions have a stronger influence on customer satisfaction rather than positive emotions.
This result might be influenced by the small number of articles included in the meta-analysis.
In order to understand more about this statement, the second phase consisted in analyzing the Industry variable. In order to do so, three main industries were analyzed (three industries that were coded, have been excluded being the number of observations too small to be computed in a meta-analysis; Bank industry = 1 observation, Healthcare industry = 2 observations). The industries took in consideration were the following:
Hospitality industry, Retail industry and Service industry. This analysis showed the following results:
1.Hospitality industry
The first industry to be analyzed was the hospitality industry. Hospitality industry includes hotels, restaurants, theme parks and different subcategories described in the literature review. This industry has both a high “human involvement” and “emotional involvement”. In line with the theoretical framework and the different publications taken into account for this analysis, the analysis should detect a positive relationship between behavioral outcomes and the hospitality industry in an emotional context.
The analysis presents the following values: r square is 0,544, that means that 54,4% of this model is explained by this regression. The coefficient analysis shows more interesting insights regarding this hypothesis. Firstly, we can see that in line with the first result, B is negative in regard to satisfaction (B = - 0,185). As previously stated, this results could be explained by the fact that customer satisfaction is influenced by negative emotions rather than positive emotions. Moreover, we see that the variable “hospitality”, that defines the industry, has a positive impact on the regression (B = 0,223). Subsequently R increases, from 0,473 to 0,531. Both results are statistically significant (p <0,05). However, only 36 effect sizes, out of 103, were coded as hospitality.
2. Retail industry
The analysis presents the following values: R square is 0,512, that means that 51,2% of this model is explained by this regression. The coefficient analysis shows more interesting insights regarding this hypothesis. Firstly, we can see that, in line with the first result, B is negative in regard to satisfaction (B = - 0,205). As previously stated, this results show that customer satisfaction is highly influenced by negative emotions rather than positive emotions. Moreover,
we see that the variable “retail”, that defines the industry, has a negative impact on the regression
(B = - 0,133). R doesn’t present any increase. Both results are statistically significant (p <0,05).
These results show that the variable “RETAIL”, that defines the industry, has a negative impact on the dependent variable. This means that the retail industry has had a small, negative influence on the relationship between emotions and consumer outcomes. However, only 23 effect sizes, out of 103, were coded as retail.
3. Education industry
The variable “Education industry” is not significant since no real effect is detected in this industry.
LOYALTY
*For this dataset only positive emotions were taken into account
P = 0,120, due to these results, the relationship is proven to be insignificant.
LOYALTY: Is a dummy variable that show the presence of loyalty in the effect size (1 = present, 0 = not present) Coefficients Regression IV Model Unstandardized Coefficients B Std. Error 1 (Constant) 0,490 0,033 LOYALTY 0,136 0,085 Model Summary
Model R R Square Std. Error of the
Estimate
NEGATIVE EMOTIONS AND THEIR OUTCOMES
*For this dataset only negative emotions were taken into account * significant p < 0,05, ** significant p < 0,010
This model is based on negative emotions. The dependent variable consists in the correlation coefficients of negative emotions underlined in the different studies included in the meta-analysis. The variables that are introduced are the followings:
COMPLAINT BEHAVIOR: Is a dummy variable that shows the presence of complaint behavior in the effect size (1 = present, 0 = not present)
NEGATIVE WOM: Is a dummy variable that shows the presence of negative WOM in the effect size (1 = present, 0 = not present)
Coefficients Regression III
Model Unstandardized Coefficients B Std. Error 1 (Constant) -0,165* 0,051 COMPLAINT BEHAVIOR 0,602* 0,121 2 (Constant) -0,22* 0,053 COMPLAINT BEHAVIOR 0,657* 0,117 NEGATIVE WOM 0,359* 0,135 3 (Constant) -0,246* 0,054 COMPLAINT BEHAVIOR 0,683* 0,116 NEGATIVE WOM 0,386* 0,134 SWITCHING BEHAVIOR 0,341** 0,195 Model Summary
Model R R Square Std. Error of the
Estimate Std. Error of the Estimate 1 ,559a 0,313 0,34804 0,34804 2 ,627b 0,393 0,33007 0,33007 3 ,654c 0,427 0,32381 0,32381
SWITCHING BEHAVIOR: Is a dummy variable that shows the presence of Switching Behavior in the effect size (1 = present, 0 = not present)
These results suggest different elements: First of all, it is underlined that a strong connection exists between negative emotions and complaint behavior ( B = 0,657 and p = 0,000 ). Therefore it is possible to affirm that negative emotional states have an influence on complaint behavior. The same relationship is underlined in NEGATIVE WOM (B = 0,359 and p = 0,011), and SWITCHING BEHAVIOR (B = 0,341 and p = 0,086), although these results present a higher significance. All the results are in line with the literature in which the emotional side of complain behavior is illustrated.
Moderators based on sample characteristics
In this model, our dependent variable is going to be the correlation coefficient underlined in the different studies included in the meta-analysis. The independent variables are: PERCENTAGE OF MALES (, NON STUDENT, REAL OBSERVATION.
* significant p < 0,05, ** significant p < 0,010
PERCENTAGE OF MALES: Is a dummy variable that shows the presence of complaint behavior in the effect size (1 = present, 0 = Female)
NON STUDENTS: Is a dummy variable that shows the presence of negative WOM in the effect size (1 = present, 0 = Students )
Coefficients Regression I Model Unstandardized Coefficients B Std. Error 1 (Constant) 0,159 0,078 PERCENTAGE OF MALES 0,079 0,163 2 (Constant) 0,041 0,095 PERCENTAGE OF MALES 0,072 0,160 NON STUDENT 0,19* 0,089 3 (Constant) 0,244 0,126 PERCENTAGE OF MALES 0,07 0,156 NON STUDENT 0,153* 0,088 REAL OBSERVATION -0,235* 0,099 Model Summary
Model R R Square Adjusted R
Square Std. Error of the Estimate 1 0,049 0,002 -0,008 0,43392 2 0,217 0,047 0,028 0,42624 3 0,316 0,100 0,072 0,41640
