Moderating Effect of Social Value Orientation on Motivation to Engage in Negative Electronic Word-of-Mouth after Anger and Regret
Master thesis
Author: Selda Ertugrul (10603190)
University of Amsterdan, Faculty of Economics and Business
June 30, 2014
Under supervision of: Dr. A. Zerres, Assistant Professor of Marketing at the University of Amsterdam.
Study Program: Msc Business Studies Specialization: Marketing
Abstract
Negative consumption experiences induce emotions such as anger and regret which are selected in this study. These negative emotions are connected with specific motivations to engage in negative electronic word-of-mouth (Wetzer et al., 2007). However, prior literature did not examine how this relationship can be moderated by the social value orientation of consumers. This study addresses this gap and also investigates if the motivations of
prosocials or proselfs differ between gender. It is suggested that angry proselfs want to vent negative feelings and regretting prosocials are motivated because of social benefits, concern for others and to help the company. Furthermore, the motivations of prosocials and proselfs without looking at which of the two emotions was felt are examined in this study. An online experiment with 164 participants was conducted to analyze the moderating effects and the generated data was thereafter tested with MANOVA. Not in line with suggestions, anger and regret did not lead to different motivations for prosocials and proselfs. Against expectation, the motivation to vent negative feelings was common for both prosocials and proselfs. Prosocials were higher motivated than proselfs for warning and helping others, helping the company, seeking advice and for extraversion/ self-enhancement. Moreover, the additional tests with gender showed differences between female proselfs' and male proselfs'
motivations. Female proselfs were higher motivated to vent negative feelings, to help and warn others and for extraversion/ self-enhancement.
Table of content
Introduction ...4
Theoretical Framework ...6
Electronic word-of-mouth ...6
Negative emotions after a consumption experience ...7
Motivations for N-eWOM ...10
Social value orientation...12
Motivations of prosocials ...15 Motivations of proselfs ...18 Emotions ...19 Anger...21 Regret ...22 Method ...24 Participants ...24 Design ...25
Materials and Procedure ...26
Emotions ...26
Social value orientation...26
Motivations for N-eWOM ...27
Pilot test ...28
Procedure ...30
Results ...31
Reliability, and means, standard deviations and correlations ...31
Manipulation check ...34
Multivariate analysis of variance ...37
Multivariate analysis of variance with gender ...39
Conclusion and Discussion ...42
General Discussion ...46
References ...52
Effects of Social Value Orientation on Motivation to Engage in Negative Electronic Word-of-Mouth after Anger and Regret
The Internet serves as a comfortable medium to share individual consumption experiences and to indicate specific emotions after a consumption experience related to products or services (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004; Wetzer, Zeelenberg, & Pieters, 2007). Sharing positive experiences can encompass recommendations while negative experiences mostly cover complaints and warnings. An increasing number of consumers spread negative experiences online without any concerns about socially negative consequences due to the anonymity of the user and the lack of a face-to-face confrontation (Verhagen, Nauta, & Felberg, 2013). The so-called negative electronic word-of-mouth (N-eWOM) can influence the purchase decisions of consumers in a destructive way for companies. This is because consumers search for information about products and services and take advantage of N-eWOM in order to minimize perceived risks and uncertainty
(Bronner & de Hoog, 2011, p. 16). Since electronic word-of-mouth (eWOM) is an important part in our modern life with advanced technology and widely spread throughout the world, it is important and interesting to know what people drives to compose N-eWOM. Especially for marketers this understanding is of great interest because N-eWOM can affect sales, product adoption and image (Chevalier & Mayzlin, 2006). Growing awareness and literature about the drivers of N-eWOM exist. Emotions and motivations affect whether or to what extent consumers deliver messages about their negative experiences. So far, it has been examined which kind of emotions and motivations lead to negative online comments but the issue of how some emotions and motivations are linked to each other has been neglected in academia. It is proven that certain emotions result in certain motivations to engage in N-eWOM (Wetzer et al., 2007). The negative emotion, anger, induces mostly the motivation to take revenge on
the company (Wetzer et al., 2007). Likewise, regret leads most frequently to the motivation to warn or help other customers (Wetzer et al., 2007). However, it is possible that this
relationship can change according to the personality traits of people. Corresponding to the motivation of people, the content of the negative online comments can differ, hence it is important to know which emotion induces which motivation and if the personality traits can affect this relationship. The social value orientation (SVO) is selected in this study as a moderating personality trait because it is expected that a detailed consideration of the SVO can moderate if and why the emotions anger and regret result in different motivations than in previous studies of Wetzer et al. (2007). Distinguishing between an other-oriented person and a self-oriented person might induce different motivations caused by a negative emotion. Furthermore, SVO is predicted to moderate the link between specific emotions and the motivation to engage in N-eWOM. The solitary understanding of the emotional or
motivational drivers is not enough for marketers for finding the most effective way handling negative online comments. A comprehensive knowledge about the drivers of N-eWOM including personality traits is essential for Marketers since N-eWOM has a harmful effect for companies. The SVO of consumers is a personality trait which can affect the motivation to engage in N-eWOM. Thus, it is important for companies to understand which online comments are written by which SVO in order to know how to react to those comments. Comments of both SVOs require different attention of companies as the motivations may differ. To study the effect, this work will address the following research question: How does social value orientation affect the relationship between negative emotions such as anger and regret after a consumption experience and the motivation to engage in negative electronic word-of-mouth? Furthermore, this study will assess if the motivations of prosocials or proselfs differ between gender. To analyze the research question and the differences between gender, this study will start with reviewing the literature and establishing a theoretical
framework along with hypotheses about the moderating effect of SVO on the relationship between motivations and the negative emotions anger or regret. Afterwards, an online experiment will be conducted to study the hypotheses, which will be followed by the conclusion. Then, a discussion with an overview about the results will follow. Finally, this study will end up with managerial implications, limitations and further research possibilities.
Theorethical Framework Electronic word-of-mouth
The lifestyle nowadays includes sharing online material by writing and distributing product reviews or sending interesting articles to friends (Berger & Milkman, 2010, p. 4). Electronic word-of-mouth (eWoM) has become an attractive medium for socially sharing emotions with consumers around the world (Wetzer et al., 2007). Lots of people engage in eWoM for different motivations. EWOM has become very important in the Marketing field because it can influence the purchase decision of consumers. A suitable definition of eWoM from Strauss (2000) is: “the positive or negative statement made by a potential, actual or former customers about a product or a company, which is made available to a multitude of people and institutions on the Internet” (p. 235). This definition is appropriate because it focuses on the positive versus negative statements made in eWOM and its far-reaching effect. The current study will also focus on the negative statements caused by negative emotions. To add, owing to the far-reaching effect it is important for Marketers to understand the rationales of negative eWOM. Drivers of negative electronic word-of-mouth (N-eWOM) will be the focal point of this study. Park and Lee (2009) claim that N-eWOM has a greater influential effect on consumers than positive eWOM. N-eWOM comprises the online communication between customers about negative consumption experiences with products and services. Customers do not disclose N-eWOM straight to a company but to a third-party (Mardhiyah, Dharmmesta, & Purwanto, 2013). Online recommendations, online reviews, sharing company
related material, rating the company and pushing the like button for company relevant information in social networks are part of eWOM. However, N-eWOM includes most often complaints and warnings about a product or service. EWOM appears in social networks where people communicate on companies' fan pages and sites or write comments with a link directed to the company through a hashtag. It also takes place in online consumer platforms and shopping websites like Amazon.com. This study focuses on instances of N-eWOM where the linguistic style and content is visible. Therefore, opinion platforms like Eopinion.com , social media websites such as Facebook and Twitter; and consumer review websites like Amazon.com will be considered. N-eWOM with negative content influences the reader by changing the attitude towards service organizations for instance (Doh & Hwang, 2009). Thus, to examine the motivations of consumers to publish N-eWOM, it requires looking at written eWOM mediums. The content of N-eWOM with the negative evaluation and the underlying negative motivation of customers is detectable in these mediums.
Negative emotions after a consumption experience
Consumers undergo consumption experiences during a purchase and when they take a purchase into consideration. In addition, consumption experiences occur during the utilization of a product and when making use of provided services. This study addresses negative
consumption experiences with all kind of service organizations which do things in exchange for money. Examples of service-organizations are restaurants, banks, airlines, and cafes (Bougie, Pieters, & Zeelenberg, 2003). Moreover, retailers such as supermarkets and department stores also belong to service organizations because their task is to assist consumers (Bougie et al., 2003). As the listing of service organizations can be extended infinitely, telecommunication companies, schools and hospitals can be classified as additional examples (Bougie et al., 2003). A consumption experience with a retailer which is also called customer experience includes the cognitive, affective, social and physical responses to a
retailer. Customer experiences can be affected by the social environment in stores, the service interface, the retail atmosphere, the assortment and the price (Verhoef, Lemon, Parasuraman, Roggeveen, Tsiros, & Schlesinger, 2009). The retail brand and former customer experiences as well as customer experiences in alternative channels for example online shopping also affect customer experiences (Verhoef et al., 2009). The situation and the customer himself can serve as a moderator between the driving factors of customer experience and the perceived customer experience as seen in the figure 1. below (Verhoef et al., 2009).
Figure 1. Conceptual Model of Customer Experience Creation (Verhoef et al., 2009)
Consumption emotions are feelings which emerge specifically during a consumption
experience or product usage (Cohen & Areni, 1991; Russell, 1979). Consumption experiences can cause consumption emotions which can be positive if the customer had a positive
experience. Similarly, a negative consumption experience induces negative consumption emotions. Regarding negative consumption experiences, the negative emotional responses
elicited are most frequently: anger, irritation, disappointment, dissatisfaction, frustration, hate and regret (Wetzer et al., 2007). A negative consumption experience can be the basis for eWOM because the developed negative emotions can motivate customers to engage in N-eWOM. According to Wetzer et al. (2007) “WoM is a form of socially sharing one’s
emotions” (p.662). The emotions frustration and anger are relatively similar in comparison to other negative emotions like discomfort (Wetzer et al., 2007, p.668). The difference is that anger is caused by the behavior of someone else and frustration is caused by a negative result such as dissatisfaction of an expected outcome (Clore, Ortony, Dienes, & Fujita, 1993). N-eWoM is driven by different motivations caused by a negative emotion. This is because of the so-called 'feeling-is-for-doing' approach which considers emotions as motivational processes and driver of certain actions. (Zeelenberg & Pieters 2006; Frijda, 1986). A certain negative emotion after an adverse consumption experience also results in a certain goal or motivation to address N-eWoM (Wetzer et al. 2007, p. 662). Because of the fact, that frustration and anger are relatively similar and generate the same motivations like taking revenge on the company (Wetzer et al., 2007), only anger will be further investigated in this study. Additionally, it has been proven that negative online complaints are a common reaction to anger (Roseman, Wiest, & Swartz, 1994). Another negative emotion which activates different motivations than anger and frustration is regret. The motivation which is mostly caused by regret is warning or helping other customers (Wetzer et al., 2007). Accordingly, anger and regret serve as representative emotions in this study because both negative emotions lead to motivations in different directions. Furthermore, anger is reported to be the first emotion felt after a negative consumption experience whereas regret is only mentioned to be the eighth (Wetzer et al., 2007). Hence, anger is a very common emotion which arises frequently and regret is a less common emotion and this makes it interesting to study the effects of these two emotions on N-eWOM.
Motivations for NeWOM
The previous researches in the literature define various motivations as drivers of eWOM. Hennig-Thurau et al. (2004), Ho and Dempsey (2008), Wetzer et al. (2007), and Bronner and de Hoog (2011) find important motivations which lead to eWOM. The most substantial motivations they found in terms of N-eWOM can be listed in four groups which embody several motivations. These are the need to be altruistic which includes warning and preventing others from making a similar purchase mistake, helping customers to make better purchase decisions and helping a company to improve products or services. The second group portray the self-interested motivations which cover self-expression, venting negative feelings and taking revenge on the company. Thirdly, the need to be in a group, which demands
strengthening social bonds, is a prominent motivation to engage in N-eWOM. Lastly, comfort search, which operates as a therapy form, is also part of the groups of motivations for N-eWOM. Hennig-Thurau et al. (2004) test the importance of existing potential motivations by building on Balasubramanian and Mahajan (2001) and specify eight motivations which are described as follows: The first motivation is platform assistance and represents the desire to transmit problems to the company through the platform without financial risk; and in a convenient and more powerful way. Especially for complaining through online comments, the online platform reduces psychological risks because it prevents from direct contact with the company (Hennig-Thurau et al., 2004). This motivation is based on the hope that the platform operator will aid and speak for the customers (Hennig-Thurau et al., 2004). Moreover, collective complaints are perceived to be more powerful than individual ones (Hennig-Thurau et al., 2004). The second motivation is venting negative feelings and is characterized by the goal to take revenge on the company; and to get rid of anger and
frustration (Hennig-Thurau et al., 2004). Concern for other consumers is the third motivation which is related to helping other consumers with their consumption decisions
(Hennig-Thurau et al., 2004). It is based on the wish to warn others of bad service organizations or save them from witnessed negative consumption experiences (Hennig-Thurau et al., 2004). Extraversion/positive self-enhancement as the fourth motivation is defined by the desire for self-expression (Hennig-Thurau et al., 2004). This motivation is driven by the thought that intelligent contributions on consumer platforms lead to recognition from others (Hennig-Thurau et al., 2004). The fifth motivation identifies the wish to receive social benefits by publishing online comments (Hennig-Thurau et al., 2004). The goal is to interact and communicate with people in the online community (Hennig-Thurau et al., 2004). Through contributions by online comments, consumers get a feeling of belonging to online
communities (McWilliam, 2000; Oliver, 1999). This motivation is also driven by the desire for social bonding. Economic incentives is characterized as the sixth motivation and
motivates consumers to engage in eWOM in order to get economic rewards or some other incentives such as Web Miles for publishing online comments (Hennig-Thurau et al., 2004). As helping the company is the next and seventh motivation, consumers are motivated to help the company if they want to give the company something back in return for a good
experience (Hennig-Thurau et al., 2004). This motivation implies that consumers wantto support the company to improve or stay successful (Hennig-Thurau et al., 2004). In the context of negative experiences, consumers want to help the company, if they are generally satisfied with it. Through the online comment, the consumers want the company to improve or change negative aspects, which they experienced. The last motivation is advice seeking (Hennig-Thurau et al., 2004). Customers search for specific feedback about their witnessed negative consumption experience (Hennig-Thurau et al., 2004). This feedback can be in form of advice, support or help for solving problems.
The motivations can differ according to emotions. The emotions,which a consumer establishes after a customer experience can cause the motivation to engage in eWoM. Wetzer
et al. (2007) argue that negative emotions are related to negative motivations for N-WOM and that these motivations may influence the content of the communication spread to others.
This study will use the eight motivational factors developed by Hennig-Thurau et al. (2004) to further analyze the moderating effect of the social value orientation of a person on the motivation to engage in N-eWOM after anger or regret.
Social Value Orientation
When making decisions such as allocating resources, individuals have to decide on how to allocate by making preferences about own outcomes and the outcomes of others. These social preferences of individuals give rise about their behavior. This behavior could be favorable or unfavorable to others while the former is costly to the agent (Murphy &
Ackermann, 2012). The social preferences are driven by the social value orientation (SVO) of a person.
SVO is about how much value individuals assign to own outcomes and to others’ outcomes. While most individuals evaluate their own outcomes positively, there are differences among individuals' evaluation of others' outcomes (Van Lange, 1999). Several models for the social value orientation exist which deal with the outcomes for the self, the outcomes for others and the equality in outcomes dimensions. The literature proposes different categories for describing social value orientations, which reach up to 10 categories (MacCrimmon & Messick, 1976). According to Messick and McClintock (1986) three categories of orientations exist and every person can represent one of these categories. These three orientations are defined in the literature as competitive, individualistic and cooperative orientations (Deutsch, 1960). Some authors such as Liebrand and McClintock (1988) also mention the altruistic orientation; altruists only want to maximize the outcome of others but results show that this orientation is represented almost never so that most studies do not consider this category. Competitors are motivated to receive the maximized difference
between the own outcome and the outcome of the other person whereas individualists try to maximize own outcomes regardless of the outcome of others (Messick & McClintock, 1986). A cooperative person is interested in an equal distribution of allocations and favors the maximization of joint outcomes (Messick & McClintock, 1986). The literature about SVO classifies competitors and individualists as proselfs and cooperatives as prosocials. This study also classifies prosocials as cooperators and proselfs as individualists and competitors, and will continue with distinguishing only between proselfs and prosocials.
Prosocials are social-oriented persons who assign positive weight to outcomes of others and the well-being of others as well as to their own well-being. Thus, prosocials are interested in fairness and in allocating resources equally (Van Lange, 1999). Proselfs who are self-oriented persons try to maximize own outcomes and are interested in always getting a better outcome or in increasing the own well-being regardless of how the others are doing. This study suggests that the social value orientation of a person can moderate the relationship between emotions and the motivations to engage in N-eWOM. Prosocials and proselfs assess the benefit of emotion adjustment strategies differently (Karagonlar &
Kuhlman, 2013, p.235). Prosocials uphold a real norm for fairness and proselfs only conform to this norm strategically when they want to maximize their self-interest (Van Dijk, De Cremer, & Handgraaf, 2004; Van Lange, 1999). It is likely that differentiating between proselfs and prosocials will lead to different results among motivations.
The Social value orientation gives rise to different behaviors because customers perceive the costs for engaging in a behavior as differently (Cameron, Brown, & Chapman, 1998). Cooperative or competitive behavior is tested in social dilemmas where cooperating deals with giving up a resource to help others without an expectation of a reward or benefit (Van Lange, 1999). For example in the ultimatum game, people can cooperate and act prosocial or defend and act proself. The ultimatum game is based on money, however in the
context of N-eWOM it is about sacrificing time and writing a helpful comment without getting any benefit because of the anonymity of the readers.
Generally, it is expected that a proself has different goals than a prosocial which differ according to emotions. According to Van Lange and Liebrand (1991), prosocials value
cooperative behavior and proselfs see cooperation as a sign of weakness and unintelligence. Cooperating refers to prosocial behavior. If cooperative behavior can be equated with the motivation to help others, then it can be suggested that proselfs do not have the motivational concern for other consumers and helping the company while prosocials do. Accordingly, non-cooperative behavior is considered as strength and intelligence by proselfs (Van Lange & Liebrand, 1991). Non-cooperation applies to dealing with own needs in order to maximize own outcomes. The proself could be motivated to vent negative feelings by publishing online comments because this motivation deals with his own needs and supports his strength and intelligence.
Moreover, adding value to the community through contributions gives utility to some customers (Hennig-Thurau et al., 2004). Because a prosocial person is interested in
maximizing joint outcomes, he strikes for adding value to the community and engages in N-eWOM. The motivations which add value to the community are concern for other consumers, helping the company, social benefits and exerting power which is part of the motivation platform assistance. (Hennig-Thurau et al., 2004).
The sixth motivation “economic incentives” deals with the consumers' appreciation of money. When writing online comments in order to receive economic incentives, no
distribution of money occurs which could affect the outcomes of others or oneself. The desire to get economic incentives also cannot influence the well-being of others in a negative or positive way. Thus, the wish to get economic incentives by writing online comments is not a matter of SVO and it is expected that the sixth motivation arises independently of SVO.
Hypothesis 1): The motivation “economic incentives” arises independently of SVO.
Motivations of prosocials.
The third motivation “concern for others” and the seventh motivation “helping the company” belong to altruistic behavior (cf. Carman, 1992). Altruism is defined by
reciprocity, social responsibility and equity (Carman, 1992). Van Lange (1999) states that the prosocial orientation connected to reciprocity. Since prosocials strike for equity through equal outcomes and therefore can feel themselves responsible for others' well-being, they can occupy the altruism motive. Accordingly, prosocials can have the motivations helping the company and concern for others to compose in N-eWOM after a negative experience because of this altruism motive. General satisfaction with a company can result from several positive experiences. If a prosocial undergoes a negative experience, he/she could be motivated to write a comment with the purpose of helping the company to improve in return for his general satisfaction. After having a negative consumption experience, prosocials could feel responsible for the well-being of other consumers who could have the same negative experience in the future. This is why prosocials could be motivated to warn consumers and help them with their own experiences. Differently, the motivations “concern for others” and “helping the company” do not serve to maximize own outcomes or well-being, they rather deal with others' outcomes. Therefore, proselfs could be lower interested than prosocials for these motivations.
Prosocials value reciprocity and therefore believe in the helpfulness of other
consumers (cf. Carman, 1992). Hence, prosocials would give advices to help other consumers and would expect it from others as well. Prosocials could be motivated for the eighth
The fourth motivation “extraversion/ self-enhancement” is about expressing feelings and writing about experiences in order to gain recognition. It is believed that intelligent comments make the writer more attractive. It is a self-related motivation which only increases the well-being or outcome of onself, therefore proselfs would have a higher motivation for “extraversion /self-enhancement”. However, self-enhancement is only achievable when writing about positive achievements, positive knowledge and positive experiences. As a consequence, writing about negative experiences and emotions would rather decrease self-enhancement. That is why it is suggested that proselfs have a very low motivation for “extraversion/ self-enhancement” after negative consumption experience. Thus, proselfs are lower motivated than prosocials for the fourth motivation or prosocials are higher motivated than proselfs.
Moreover, the motivation “social benefit” can be a motivation of prosocials because the online comments can contribute and help the people in the community. At the same time, consumers can have pleasure while chatting and get the feeling of belonging to the
community (Oliver, 1999). Thus, the motivation social benefits results in positive joint outcomes for a prosocial because he / she can add value to the community while enjoying to belong to the community. Additionally, proselfs look at outcomes which are beneficial for themselves and the fifth motivation “social benefits” provides them positive outcomes. Therefore, it is claimed that both SVOs have the fifth motivation so that it does not differ between prosocials and proselfs.
To conclude, it is claimed that prosocials could be higher motivated than proselfs to engage in N-eWOM because of the third motivation “concern for others”, the fourth motivation “extraversion/ self-enhancement and the seventh motivation “helping the
company”. The fifth motivation “social benefits” probably does not differ between prosocials and proselfs.
Hypothesis 2a): Prosocials have higher motivations than proselfs for concern for others, extraversion/ self-enhancement and helping the company to engage in N-eWOM.
Hypothesis 2b): The appearance of the motivation “social benefits” does not differ between prosocials and proselfs.
Publishing complaints in an online platform is perceived to be more powerful than single complaints and all people in the community can profit from the effect of this power. The motivation exerting power, which belongs to the first motivation “platform assistance” leads therefore to equal outcomes for the self and for others. Thus, it is proposed that prosocials are more often motivated for platform assistance but that this differs between gender. Accordingly, the consideration of this motivations needs to be done by differentiating between gender. Henley (1977) states that men are interested in social power, which is a determinant of the masculine position. This intended social power also reflects in men's speech (Henley, 1977). If men want their speech or comment to be more powerful, they will be more motivated to write an online comment in order to exert power with the assistance of the platform. Hence, it is suggested in this study that prosocials are motivated to engage in N-eWOM because of platform assistance but the amount of prosocials consists more often of men. Thus, it is proposed that a gender effect can occur for the first motivation and prosocials who are male could have a higher motivation to engage in N-eWOM because of motivation “platform assistance”.
Hypothesis 2c): Male prosocials score higher on the motivation “platform assistance” than female prosocials and proselfs.
Motivations of Proselfs.
The eighth motivation “seeking for advice” could be a wise motivation for proselfs in improve the situation. A good advice could compensate for the negative experience and therefore could maximize the own outcome.
Furthermore, proselfs could be motivated to engage in N-eWOM after they have experienced a negative emotion in order to vent negative feelings and take revenge on the company. The motivation is the intention to engage in eWOM (Wetzer et al., 2007). The social value orientation alters the relationship between behavior intention, which is favorable for the environment, and the awareness of the consequences for the environment (Gärling, Fujii, Gärling, & Jakobsson, 2003, p.8). The underlying intention of the motivation revenge is negative because it serves to harm the company. According to the Theory of Reasoned
Actions, intention leads to behavior (Fishbein & Ajzen, 1975). A missing negative intention does not lead to a negative behavior such as the articulation of revenge (Boo & Kim, 2013). Prosocials do not have negative intentions due to desired maximization of joint outcomes and because they neither want to affect the self nor the others negatively. As a consequence, prosocials do not have the motivation to take revenge on the company. However, the intention of proselfs can be negative if their own outcome can be maximized. Additionally, proselfs can maximize their own well-being by venting negative feelings regardless of how well the others do and if it harms others. Proselfs can maximize their own well-being thereby since it helps to escape from negative emotions.
To sum up, it is proposed proselfs could be higher motivated than prosocials to
compose N-eWOM because of the second motivation “venting negative feelings”. As already suggested in the motivations of prosocials, prosocials are also motivated for the eighth motivation “advice seeking”. Thus, it is claimed that both SVOs have the eighth motivation
“advice seeking” so that it does not differ between prosocials and proselfs.
Hypothesis 3a): Proselfs have higher motivations than prosocials for venting negative feelings to engage in N-eWOM.
Hypothesis 3b): The appearance of the motivation “advice seeking” does not differ between prosocials and proselfs.
The suggestion that proselfs are less motivated than prosocials to engage in N-eWOM because of their concern for others could be subjected to a gender effect. The amount of proselfs could be lower than prosocials because less male proselfs could have the motivation of concern for others. Beutel and Marini (1995) find evidence about value differences
between the genders in their studies. These studies show that women are more socio-emotional and care more about others because they feel responsible about the other's well being(Beutel & Marini, 1995). Thus, female proselfs could have the third motivation “concern for others” despite their general self-orientation.
Hypothesis 3c): Female proselfs score higher on the motivation “concern for others” than male proselfs.
Emotions
Consumption emotions arise while using a product and during a purchase or the consideration of a purchase (Cohen & Areni, 1991 ). These consumption emotions can lead to word-of-mouth communication (Nyer, 1997). For the purpose of defining which emotions induce word-of-mouth and in which manner, it is crucial to be able to distinguish between
emotions. Prior research has tried to search for differences between emotions by focusing on expressions, changes of body language, content of thoughts and feelings (Roseman et al., 1994). Roseman et al. (1994) differentiate emotions in the context of emotion research by their experiential content. This differentiation among emotions proceeds through 5
experiential categories, which are as follows: (1.) feelings, (2.) thoughts, (3.) action tendencies, (4.) actions and (5.) emotional motives (Roseman et al., 1994). Feelings are psychological or physiological impressions and effects whereas the psychological processes generate the thoughts (Bougie et al., 2003). Thoughts are a result of thinking and can be classified as ideas, arrangements of ideas, judgments, notions and plans (Bougie et al., 2003). The goodwill to practice a specific action is described as an action tendency while actions cover real behavior, which can be specific (Bougie et al., 2003). Emotional motives refer to wanted goals which are linked to distinct emotions (Bougie et al. 2003; Roseman et al., 1994). This distinction by experiential content is also used in this study's methodology part for differentiating between the emotions anger and regret. In order to clarify the five differentiating categories of emotions, the categories will be further explained by the
emotions anger and regret. Anger may concern feelings of blood rushing through one's body, thoughts about violence toward others, action tendencies such as feeling like hitting
somebody, actual behaviors like saying something nasty and emotional motives like wanting to hurt someone (Roseman et al., 1994). The emotion of regret can refer to feelings of tension in one's face, thoughts about a lost opportunity, feeling like kicking onself as an action
tendency, actions like doing something differently and emotional motives such as wanting to get a second chance (Roseman et al., 1994). Bougie et al. (2003) find evidence about the difference of the emotions anger and dissatisfaction by investigating these emotions by their experiential content. This explains and affirms the potential to distinguish anger from other emotions. Moreover, Roseman et al. (1994) find that anger and regret can be regarded as
distinct emotions.
Furthermore, an emotion can only be elicited if a person appraises an experience or an event as significant for personal welfare, which refers to the appraisal theory (Bougie et al., 2003). Despite of some differences in people's appraisals of specific events, a specific emotion occurs from a similar scheme of appraisals (Bougie et al., 2003). The appraisal theory helps to discover the antecedents of emotions since certain events are appraised in the same way producing a particular emotion. Regret differs from other negative emotions like anger in terms of experiential content, as well as its appraisals, and the behavioral
consequences (Zeelenberg, Van Dijk, Manstead, & der Pligt, (1998)
Anger.
Specific consumption emotions are systematically connected with consumer's appraisals of specific events or experiences (Ruth, Brunel, & Otnes, 2002). The emotion anger is associated with appraising an event as unpleasant and highly unfair (Ruth et al., 2002). In addition, anger occurs if a failure with a service-organization is related to others rather than to the self (Ruth et al., 2002). Thus, consumers believe that service organizations treat them intentionally unjustly (Bougie et al., 2003). Anger also results if an event is appraised as frustrating and harmful (Bougie et al., 2003). All in all, anger is an emotion, which occurs because of the failures of others and is a strong emotion because it is triggered in very negative situations. Moreover, the emotion anger is one of the strongest emotions because it highly affects social relations and the person undergoing the emotion (Bougie et al., 2003). Strong emotions produce pressure and arousal which needs to be thrown off and venting negative feelings serves to get rid of negative emotions. Wetzer et al. (2007) studied emotional effects on motivations to engage in N-eWOM. His studies reveal that people have a higher motivation to vent negative feelings and to take revenge on the company after feeling anger than feeling regret. It is supposed that Proselfs are higher motivated to write
online comments in order to vent negative feelings as suggested in Hypothesis 3a). Hence, this study proposes that proselfs are higher motivated for the second motivation “venting negative feelings” after feeling anger than after regret.
Hypothesis 4): Proselfs are higher motivated to vent negative feelings after feeling anger than after regret.
Regret.
A number of studies investigate when people feel regret because the emotion can arise in three different settings. People can either regret an action if the outcome is not favorable or the omission of an action if the outcome would have been favorable (Kahneman & Tversky, 1982; Zeelenberg, van Dijk, Manstead, & van der Pligt, 1998). These two settings belong to retrospective regret. Kahneman and Tversky (1982) argue that people feel regret more
intensively for a negative outcome stemming from action than for the same negative outcome resulting from inaction. Thirdly, regret can arise if people anticipate to feel regret when taking an action (Zeelenberg, 1999). After a negative consumption experience in which the consumer undertakes an action, regret can only arise after a negative outcome resulting from action. Thus, in the context of negative consumption experiences only the setting in which regret stems from action will be further regarded.
Regret results if an event is appraised as thinking that one had the possibility to change something about the event (van Dijk & Zeelenberg, 2002). Furthermore, consumers who feel regret attribute the failure to themselves (van Dijk & Zeelenberg, 2002). Differently than anger that refers to other-agency, regret refers to self-agency. Regret is in the literature most often related to poor decision making and Zeelenberg and Pieters (2007) describe regret as: “Regret is a comparison-based emotion of self-blame, experienced when people realize or
imagine that their present situation would have been better had they decided differently in the past” (p.4). Consumers are eager to prevent regret and try to regulate their feeling if they undertook a regrettable action (Zeelenberg & Pieters, 2006). After a regrettable action, consumers aim at improving their situation by refusing and holding down the experience or undoing the outcomes (Zeelenberg & Pieters, 2007)
Regret results less frequently in complaining behavior than other negative emotions such as disappointment because people feel responsible for the negative outcomes of their actions (van Dijk & Zeelenberg, 2002). Therefore, the motivation to vent negative feelings occurs less likely after regret because consumers attribute the responsibility rather to themselves than to the company. Consumers make the company responsible for the failure when they e.g. take revenge on the compan. In addition, consumers use the platform as an assistance to transmit problems to the company. However if the problem deals with the own, poor decision, as it is the antecedent of feeling regret, consumers would not be motivated to engage in N-eWOM because of platform assistance. Wetzer et al. (2007) find similar results in his studies; consumers have the motivation to vent negative feelings more often after anger. They also show in their studies that after feeling regret people have higher motivation to warn others and for social bonding than after feeling anger.
Prosocials are higher motivated than proselfs for motivation “concern for others” as suggested in hypothesis 2a) and are motivated because of “social benefits” as suggested in hypothesis 2b). Hence, it is proposed that prosocials are higher motivated for the third motivation “concern for others” and the fifth motivation “social benefits” after feeling anger than after regret. Moreover, since concern for others and helping the company underlie the same altruistic motive, it can be propounded that prosocials have also a higher motivation for the seventh motivation “helping the company” after feeling regret than after feeling anger.
Hypothesis 5): Prosocials are higher motivated to help and warn others, help the company and to enjoy social benefits after feeling regret than after anger.
Method Participants
N-eWOM addresses actual creation of online comments about negative emotions after consumption experiences. The experiment is not restricted to a particular research group because the Internet is open to the public and everyone can engage in N-eWOM. Thus, this study points at any Internet user and is using an online experiment on Qualitrics. The Internet users are requested to participate in the online experiment mainly through social media (Facebook), face-to-face communication and direct e-mails. The experiment was conducted in two different languages, namely English and German to be able to reach a wider set of respondents (see Appendix I for the online experiment and Appendix II for the German version). A total of 164 Internet users (108 females, 55 males) participated in the online experiment. Table 1 summarizes the demographic statistics of the participants. Each demographic question contained one to five missing answers which was ignored due to the small number (see Appendix III). The average age of the 164 participants is 23.94 years old (SD = 3.68). Except one, all participants are Social Media users, with an average of using Social Media for 3.60 hours a day (SD = 1.17). Due to the world wide use of the Internet, participants from 21 different countries partook this online experiment. Most of the
participants have the German (N=94) , Dutch (N=25) or Turkish (N=11) nationality. The full list of the nationality of participants is available in the Appendix III.
Design
The online experiment used a factorial design between subjects. The two independent variables were emotion and social value orientation. The two levels of the first independent variable (anger and regret) were combined with the two levels of the second (proself or prosocial) to produce four distinct treatments: anger and proself, anger and prosocial, regret and proself, and regret and prosocial. This is a 2x2 (two-by-two) factorial design because there are two independent variables with two levels respectively. Each participant was in only one of the four conditions and is measured using eight continuous dependent variables. The eight dependent variables were the following motivations from Hennig-Thurau et al. (2004): (1) platform assistance, (2) venting negative feelings, (3) concern for others, (4)
extraversion/positive self-enhancement, (5) social benefits, (6) economic incentives, (7) helping the company, and (8) advice seeking.
Male 55 (33.5%) 108 (65.9%) <20 26 (15.9%) 21-25 89 (54.3%) 26-30 32 (19.5%) 31-35 11 (6.7%) >36 1 (0.6%) 4 (2.4%) 67 (40.9%) 3 (1.8%) 4 (2.4%) 55 (33.5%) 27 (16.5%) 1 (0.6%) 1 (0.6%) 1 (0.6%) 35 (21.3%) 43 (26.2%) 2-3hours 37 (22.6%) 45 (27.4%) 3 (1.8%) Table 1 Demographical statistics (N = 164) number percentage *Gender Female *Age in years
*Educational level Grammar school
High school or equivalent
Vocational/technical school 2 year Some College
Bachelor's degree Master's degree Doctoral degree
Professional degree (MD, JD etc.) Use of Social Media Never
0-1 hours 1-2 hours
More than 3 hours Not daily
Materials and Procedure Emotions.
Following Roseman et al. (1994), the emotions anger and regret were
manipulated through instructions. The instructions for anger and regret were based on Bougie et al. (2003) and asked the participants to recall a consumption experience with a service organization which made them feel intense anger or regret. The instructions stimulated the participants to re-experience the consumption experiences and included examples of service organizations to facilitate the recall. The instruction for anger and regret respectively can be found in the Appendix I. Afterwards, the participants were asked to describe their negative consumption experience which made them feel intense anger or regret depending on the condition they were assigned to. The description of the consumption experience was written in an open-ended by the participants. In order to check if the manipulation of one of the emotions worked, close-ended questions based on Roseman et al. (1994) about the intensity of emotions were asked. These close-ended questions by Roseman et al. (1994) were
preferred because they measure the emotions according to their experiential content and were found to distinguish lots of emotions. These questions were answered on a 9 point Likert scale (1 = not at all to 9 = very much). Close-ended questions about both emotions anger and regret were asked together and in a mixed order to all of the participants regardless of which emotion was manipulated (see Appendix I for materials). Therewith, it was possible to test if participants with the instruction for anger scored higher on questions about the intensity of anger. The same was also applicable for the emotion regret.
Social value orientation.
The manipulation of proself or prosocial orientation were based on Weingardt, Bennet and Brett (1993) and were implemented through instructions. The instruction from Weingardt et al. (1993) focuses on the motivational orientation on group negotiation process and
outcome. However, this instruction was adjusted so that it addressed the social value orientation and two instructions for proself and prosocial orientation respectively were created. These instructions covered economical outcomes because the following measurement of the social value orientation focused on monetary distribution between another person and oneself. Furthermore, the instructions also pointed at social consequences such as the well-being of others in the prosocial instruction and the indifference to the
consequences for others in the proself instruction. That is to align the instruction with the general context of N-eWOM where the motivation to engage in N-eWOM is affected by preferred social consequences. These two instructions for proself and prosocial orientation can be retrieved in the appendix. After the instructions, the SVO of the participants was measured for the manipulation check. Therefore, the SVO Slider Measure with six items from Murphy, Ackermann and Handgraaf (2011) was used. Several other SVO measurements exist, however this SVO Slider Measure allows for more statistical power and is the most topical one between other popular measurements. For each of the six items, participants had to distribute money between an anonymous person and oneself. All of the six items have the same form and show joint payoffs with nine possibilities within which the participants had to decide for one preferred joint distribution. The decision of the participants affected how much money the other person and oneself received. The detailed description of the instructions and the six items from Murphy et al. (2011) can be found in the Appendix I.
Motivation for N-eWOM.
The eight motivations to engage in N-eWOM were measured depending on Hennig-Thurau et al. (2004). The participants responded to the question why they write comments on online platforms. Since Hennig-Thurau formulated the items about motivations for positive and negative eWOM, the items were adjusted in order to present motivations only for N-eWOM (see Table 7 or Appendix I). For each item, consisting of a reason to write online
comments grounded on a particular motivation (e.g. platform assistance, venting negative feelings, concern for others etc.), participants had to indicate to which degree this reason was representative of their motivation on a five point Likert scale (1 = strongly disagree to 5 = strongly agree). All of the items belonging to a particular motivation were asked in a mixed order to prevent affectation.
Pilot test
Before starting the actual online experiment, a pilot test was conducted to prove if the manipulations work and if the instructions and the items in the measurement are
understandable. 11 people participated in the pilot test. After a short introduction, the manipulation through an instruction for an emotion was followed. Afterwards, the participants answered close-ended questions about the emotions anger and regret as a manipulation check. Participants who were manipulated with the emotion anger mostly scored higher on questions which refer to anger in the following measurement (see Table 2). Likewise, the same held true for the emotion regret. Four participants in the anger condition scored higher in questions related to anger. Also in the regret condition, four of five
participants scored higher in questions related to regret. One participant scored equally high on regret and anger and is therefore not included in table 2. SPSS was used to test if a relationship between the manipulated emotions and the measured emotions exists. This study used an alpha level of .05 for all statistical tests. The probability of the chi-square statistic χ² = 3.600) p = 0.058 and was close to the alpha level of significance of p < .05. The null hypothesis that there is no existence of a relationship between the manipulated emotions and the measured emotions can be rejected. Since the amount of participants is low, it can be assumed that p = 0.058 shows that the relationship of
those two variables is statistically significant. Participants who were manipulated to feel an emotion actually felt this emotion.
To be able to figure out if the manipulation for the social value orientation through instructions worked, the SVO Slider Measure was used in the pilot test. The results showed that more prosocials emerged when measuring the SVO after the manipulation for prosocial orientation. After the manipulation for proself orientation, half of those manipulated persons were measured to be proselfs (see Table 3). Three of six partcipants in the proself condition were measured to be proselfs. Four of five participants in the prosocial condition were measured to be prosocials. The probability of the chi square statistic (χ² = 1.061) was p = 0.303 and was higher than the alpha level of significance of 0.05. The null hypothesis that there is no existence of a relationship between the manipulated SVO and the measured SVO can be retained. The relationship of manipulated SVO and measured SVO is statistically not significant. However, when looking at the cross tabulation in table 3, there are more
measured prosocials in the prosocial condition and equally often measured proselfs in the proself condition. Therefore, there is a tendency that the manipulation of SVO worked. The manipulation for SVO might not have worked for some participants because it is assumed that some participants did not read the instructions due to the amount and length of the instructions. This assumption was enforced by feedback in face-to-face communications about the pilot test which stated that the instructions for the emotions in the beginning are already too long so that the rest was not read by some participants. Hence, the instructions for the manipulation of emotions were shortened for the actual online experiment (see Appendix IV for the initial instructions for the manipulation of emotions in the pilot test). At the end of the online experiment, the participants were asked to report problems while completing the tasks. No problems were stated and all the answers of the respondents in the descriptive part were clear. All in all, the pilot test was successful and the instructions for the manipulations of emotion and SVO and the measurements for emotion and SVO were adopted in the actual online experiment. Thereafter, the study continued with the actual experiment. (see Appendix
IV for materials)
Procedure
Before participants started the online experiment, they were first presented a
welcoming and a short introduction. The general objective of the study was introduced here and it was explained that the online experiment consists of five parts. Following the
introduction, the participants were randomly presented an instruction for the manipulation of anger or regret. Then, participants were asked to describe a consumption experience which made them feel either intense anger or regret. The second part comprised the close-ended questions about the emotions anger and regret. Here, it was measured if the emotion which was manipulated scored higher than the other emotion. Next, the intensity of the emotion felt
3 1 50,0% 20,0% 3 4 50,0% 80,0% Total 6 5 100,0% 100,0% Table 3
Manipulation check of SVO: SVO Condition * measured SVO_type Crosstabulation
SVO (Condition) Pros elf Pros ocial SVO_type (m eas ured) Pros elf Count
% within SVO (Condition) Pros ocial Count
% within SVO (Condition) Count
% within SVO (Condition)
Anger 4 1 80,0% 20,0% Anger 1 4 20,0% 80,0% Total 5 5 100,0% 100,0% Table 2
Manipulation Check of Emotion: Emotion Condition * Measured Emotion Crosstabulation
Em otion (Condition) Regret
Em otion (m eas ured) Regret Count
% within Em otion (Condition) Count
% within Em otion (Condition) Count
during the online experience and the time of the experience was asked. In the following part three, the participants were displayed an instruction related to manipulation of proself or prosocial orientation at random. Subsequently, participants were instructed to complete the six item SVO Slider Measure task. This task was added as a manipulation check. Thereupon, the motivations to engage in N-eWOM were measured in the fourth part. Before this
measurement started, the participants were instructed to imagine to write an online comment about the negative consumption experience with the service-organization which made them feel intense anger or regret at that moment. Finally, participants were asked to answer several questions about their demographics and about the duration of social media use per day.
Results
The intensity of the emotion during the negative consumption experience was on a 10point Likert scale (0 = not intense, 10 = very intense) on average 6.29 (SD = 2.27). The negative experience happened on average three month ago. (see Appendix III)
Reliability, means, standard deviations and correlations
Except demographic questions, this study did not have missing data. The quality of the data will have an effect on the results of hypothesis testing. If the data is not reliable, the data will not be useful. Thus, testing the reliability is important and necessary for a study (Cooper dan Emory 2001). The reliability tests were leaned on a value of Cronbach's alpha (α) (Cronbach, 1951) that needs to be higher than .60 if the amount of items is small
(Schmitt, 1996). In table 4, the item for the emotion anger and its Cronbach's alpha is shown. It came out that the item of anger has been tested to be reliable. Moreover, all the questions which were asked in the online experiment and belonged to the measurement of anger are presented in table 4.
Similarly, the item of the emotion regret and its Cronbach's alpha are exhibited in table 5. All the questions which belong to regret are included in this table and the item of regret has been proven to be reliable.
Reliability testing was also attempted for the items of all motivations. The Cronbach's alpha of each motivation is shown in table 6. Furthermore, all questions which belong to each motivation are displayed in table 6. It appears that the items of motivations were tested to be reliable. Hence, all of the items in table 4, 5 and 6 were used for further testing (see Appendix V for materials related to reliability).
Anger Table 4
Cronbach's alpha of Emotion Anger
α = 0.82 Feel that you'd explode?
Think how unfair the situation was? Feel like yelling (shouting / screaming)? Say something nasty?
Strike out at someone?
Feel blood rushing through your body? Want to get back at someone?
Feel like hitting someone? Want to hurt someone?
Think of violence toward others?
Table 5
Cronbach's alpha of Emotion Regret
Regret α = 0.81
Feel tension in your face?
Think of what a mistake you made? Feel like kicking yourself? Dwell on what happened? Want to get a second chance? Do something differently? Feel like correcting your mistake? Feel a sinking feeling?
Want to improve your performance? Think about a lost opportunity?
Table 7 provides the correlation matrix of SVO, anger, regret and the eight motivations in order to estimate the relationship among all of the items. Moreover, the table shows the means and standard deviations of all the items.
Table 6
Cronbach's alpha of all Motivations
Motivation 1: Plattform assistance α = 0,68
…. it is not that costly.
…. I believe the platform operator knows the person in charge within the company and will convey my message. …. I believe companies are more accommodating when I publicize the matter.
…. it is more convenient than writing to or calling the company.
…. one has more power together with others than writing a single letter of complaint. …. the platform operator will stand up for me when speaking to the company.
Motivation 2: Venting negative feelings α = 0,76
…. the company harmed me and now I will harm the company …. I like to get anger of my chest.
…. I want to take vengeance upon the company
…. my contributions help me to shake off frustration about negative experiences
Motivation 3: Concern for others α = 0,80
…. I want to save others from having the same negative experiences as me …. I want to help others with my own negative experiences
…. I want to warn others of negative experience
Motivation 4: Extraversion / positive self-enhancement α = 0,65
…. this way I can express my feeling about a negative experience. …. I can tell others about a negative experience.
…. I feel good when I can tell others about my experiences. …. my contributions show others that I am a clever customer.
Motivation 5: Social benefits α = 0,76
…. I believe a chat among like-minded people is a nice thing. …. I meet nice people this way
…. it is fun to communicate this way with other people in the community.
Motivation 6: Economic incentives α = 0,67
…. of the incentives I receive (e.g., Web miles). [If applicable on the consumer platform] …. I receive a reward for the writing. [If applicable on the consumer platform]
Motivation 7: Helping the company α = 0,65
…. in my own opinion, good companies should be supported to improve
…. in general I am satisfied with the company that I want to help the company to improve and to be successful
Motivation 8: Advice seeking α = 0,75
…. I hope to receive advice from others that helps me solve my problems. …. I expect to receive tips or support from other users.
Manipulation check
In order to check if the manipulation of emotions worked, the data was analyzed using a chi square goodness of fit test. The relationship between the manipulated emotions and the measured emotions was tested. The null hypothesis which states that the manipulation of the emotions did not work was rejected, χ²=9.304, p = .002. The p-value achieved statistical significance (p <.05) which explains that the manipulation of emotions was successful. 158 of the total amount of participants which is 164 scored higher on one emotion. Of those 158, 68 participants were in the regret condition and were manipulated to feel regret, and 90
participants were in the anger condition. 40 participants in the regret condition actually felt regret when it was measured and 59 participants in the anger condition felt anger (see Table 8). Thus, there are more participants who felt the manipulated emotion in both conditions.
Variables: M SD 1 2 3 4 5 6 7 8 9 10 1. SVO Type 2,68 0,47 2.Anger 4,06 1,5 -,09 3.Regret 3,98 1,61 -,01 4.Motivation1 2,99 0,8 ,11 5.Motivation2 3,2 0,96 -,10 ,08 6.Motivation3 4,23 0,79 ,09 ,14 ,15 7.Motivation4 3,16 0,8 8.Motivation5 2,14 0,98 ,08 ,13 9.Motivation6 2,34 1,13 -,03 ,06 ,06 11.Motivation7 3,42 1,04 -,06 ,05 -,04 ,15 10.Motivation8 3,09 1,19 ,03 ,13 ,03 Table 7
Means, Standard Deviations, Correlations
,27** ,17* ,21** ,38** ,34** ,27** ,26** ,18* ,26** ,18* ,54** ,37** ,40** ,18* ,18* ,53** ,21** ,52** ,26** ,37** ,29** ,25** ,35** ,25** ,32** ,23** ,45** ,24** ,19* ,43** ,32** ,41** ,34** ,24** ,41**
The manipulation check of SVO was conducted by the chi square goodness of fit test. The relationship between the manipulated SVO and the measured SVO was tested. The null hypothesis which indicates that the manipulation of the SVO did not work was retained, χ² =.706, p = .783. The p-value did not reach statistical significance because p is not smaller than .05. Hence, the manipulation of SVO did not work. That is why this study continued with the data of the measured SVO to analyze the moderating effects of SVO on motivations after emotions. The table 9 shows that in each condition, prosocials and proselfs were equally distributed.
Table 10 presents the frequencies the four conditions with manipulated emotions and measured SVOs. The frequencies were as follows: proself and regret, 24 participants; prosocial and regret, 48 participants; proself and anger, 29 participants; prosocial and anger,
Table 8
Manipulation Check of Emotion: Emotion Condition * Measured Emotion Crosstabulation
Em otion (Condition)
Total Regret Anger
Regret Count 40 31 71
% within Em otion (Condition) 56,3% 43,7% 100,0%
Anger Count 28 59 87
% within Em otion (Condition) 32,2% 67,8% 100,0%
Total Count 68 90 158
% within Em otion (Condition) 43,0% 57,0% 100,0% Em otion
(m eas ured)
Table 9
Manipulation check of SVO: SVO Condition * measured SVO_type Crosstabulation
SVO (Condition)
Total Pros elf Pros ocial
Pros elf Count 26 27 53
% within SVO (Condition) 49,1% 50,9% 100,0%
Pros ocial Count 57 54 111
% within SVO (Condition) 51,4% 48,6% 100,0%
Total Count 83 81 164
% within SVO (Condition) 50,6% 49,4% 100,0% SVO_type
63 participants.
The SVOs were measured by the SVO Slider which also measured the SVO angles. Proselfs are represented until the angle 22,44 and prosocials are covered from angle 22,45. The distribution of proselfs and prosocials according to angles can be seen in figure 2. The mean of the SVO is 25,71 (SD = 13,133)
Figure 2 The distribution of the measured SVO scores of the participants from the Slider Measure as represented by angles
Total
Em otion 24 48 72
Anger 29 63 92
Total 53 111 164
Table 10
Frequencies of the four conditions with manipulated Emotion and measured SVO
Count
SVO_type Pros elf Pros ocial Regret
Multivariate analysis of variance
A multivariate analysis of variance (MANOVA) was performed to test how independent variables affected some patterning of response on the dependent variables. A multivariate testing was undertaken for the manipulated emotions and the measured SVO as fixed factors, and the eight motivations as dependent variables. Here, the moderating effect of SVO on motivations after emotions were tested. As previously said this study used an alpha level of .05 for all statistical tests. Multivariate tests showed that SVO had a significant main effect on motivations in general: F(1, 8) = 2.99, p = .004 (p < .05). Statistically significant effects arise if p < .05. Emotion had no significant main effect on motivations in general: F(1,8) = 1.03, p = .419 (p > .05). Similarly, the interaction between SVO and emotion had no significant main effect on motivations in general: F(1,8) = .81, p = .592 (p > .05).
The test of between-subjects effects indicated the effects of SVO, emotion, and the interaction between SVO and emotion on the eight individual motivations respectively which are shown in table 11. Accordingly, the interaction between SVO and emotion had no
statistically significant effect on any of the motivations (p > .05). Therefore, the effect of SVO was independent of emotions. For the moderating effect of SVO, it did not differ which emotion was experienced. Thus, the hypotheses 4) and 5) which suggested that the
moderating effect of SVO on motivations differ between emotions can be rejected. Similarly, emotion had no statistically significant effect on any of the motivations (p > .05).
SVO had a statistically significant effect on the third motivation “concern for others” with F(1,8) = 13.37, p = .000 and on the fourth motivation “extraversion / self-enhancement” with F(1,8) = 5.74, p = .018. SVO also had a statistically significant effect on the seventh motivation “helping the company” with F(1,8) = 10.68, p = .001, and on the eighth
motivation “advice seeking” with F(1,8) = 6.05, p = .015. The non-significance of SVO on the first, second, fifth and sixth motivations (p > .05) reveals that the hypothesis 3a) can be
rejected, the second motivation did not differ between proselfs and prosocials. Hypothesis 3a propounded that proselfs are higher motivated than prosocials for the second motivation “venting negative feelings”. However, the non-significance of SVO on the sixth motivation approved the hypothesis 1); the motivation “economic incentives” emerged independently of SVO. Moreover, the non-significance of SVO on the fifth motivation “social benefits” confirmed the hypothesis 2b); the emergence of the fifth motivation did not differ between prosocials and proselfs.
The statistically significant effects of SVO on the third, fourth, seventh and eighth motivations are characterized by mean differences between prosocials and proselfs which are significant at the .05 level. Prosocials show a higher mean for the third motivation (prosocial: M = 4.40, proself: M = 3.92). Also for the fourth motivation, prosocials represent a higher mean (prosocial: M = 3.27, proself: M = 2.94). Similarly, prosocials have a higher mean for the seventh motivation (prosocial: M = 3.60, proself: M = 3.04), and for the eighth motivation (prosocial: M = 3.23, proself: M = 2.75). Hence, prosocials are higher motivated than proselfs for the third, fourth, seventh and eighth motivations (see table 12). This result supports the hypothesis 2a), prosocials are higher motivated than proselfs for the third motivation
“concern for others”, the fourth motivation “extraversion /self-enhancement” and the seventh motivation “helping the company”. However, this result fails to confirm the hypothesis 3b); the appearance of the eighth motivation “advice seeking” differs between prosocials and proselfs. (see Appendix VI for materials)
Multivariate analysis of variance with gender
A multivariate testing was conducted again for the manipulated emotion, the measured SVO as fixed factors and the eight motivations as dependent variables. This time, the gender of participants was included in the fixed factors The multivariate tests demonstrated that SVO had a significant main effect on motivations in general: F(1,8) = 3.25, p = .02. Emotion with F(1,8) = 1.30, p = .248 and the interaction between SVO and emotion with: F(1,8) = 1.19, p = .309 had again no significant main effect on motivations in general. Gender of participants had a statistically significant main effect on motivations in general: F(1,8) = 2.44, p = .017. However, the interaction between SVO and gender had no statistically significant main effect on motivations in general: F(1,8) = 1.65, p = .115. Similarly, the interaction between SVO, emotion and gender had no effect on motivations in general: F(1,8) = .912, p =
SVO Emotion SVO*Emotion
.179 .711 .466 .239 .241 .280 .000* .127 .395 .018* .678 .517 .380 .608 .277 .653 .906 .637 .001* .292 .373 .015* .418 .849 Table 11
Test of between-subjects effects: “SVO“ , “Emotion“ and “interaction between SVO and emotion“ as fixed factors
Motivation 1: Plattform assistance Motivation 2: Venting negative feelings Motivation 3: Concern for others
Motivation 4: Extraversion / positive self-enhancement Motivation 5: Social benefits
Motivation 6: Economic incentives Motivation 7: Helping the company Motivation 8: Advice seeking
* Significant at p < .05 level
Table 12
Pairwise comparisons: effects of “SVO“ on motivations (Means)
Prosocials Proselfs Difference
Motivation 1 3,04 2,86 0,18 Motivation 2 3,18 3,32 -0,14 Motivation 3* 4,4 3,92 0,48* Motivation 4* 3,27 2,94 0,33* Motivation 5 2,18 2,04 0,14 Motivation 6 2,32 2,4 -0,08 Motivation 7* 3,6 3,04 0,56* Motivation 8* 3,23 2,75 0,48*
