Lies Ekhart, 6371507
MSc in Marketing
First supervisor:
Dr. Karin Venetis
Second supervisor:
November 2014
Version 2.4
An explanatory study; the role of
product involvement
Table of content
Abstract ... 6
1. Introduction ... 7
2. Literature review ... 11
2.1 What is perceived price unfairness? ... 11
2.2 Why is perceived price unfairness important in marketing? ... 15
2.3 What are the effects of perceived price unfairness? ... 17
2.4 What is product involvement and why is it important in a perceived price unfairness context? ... 20
2.5 How does product involvement influence the behavioral reaction(s) when perceived price unfairness occurs? ... 23
2.6 Theoretical model and hypothesis ... 25
3. Research methodology ... 26 3.1 Research design ... 26 3.2 Questionnaire design ... 27 3.3 Data collection ... 29 3.4 Data analysis ... 31 4. Results ... 32 4.1 Sample ... 32 4.2 Reliability analyses ... 33 4.3 Descriptive statistics ... 37
4.3 Testing the model ... 39
5. Conclusion and discussion ... 48
6. References ... 50
List of tables and figures
Figure 1; Theoretical model
Figure 2; Estimated Marginal Means of complaining Figure 3; Estimated Marginal Means of complaining
Figure 4; Estimated Marginal Means of (tendency to negative) word-of-mouth Figure 5; Estimated Marginal Means of (tendency to negative) word-of-mouth Figure 6; Estimated Marginal Means of willingness to claim refund
Figure 7; Estimated Marginal Means of willingness to claim refund Figure 8; Estimated Marginal Means of purchase intention
Figure 9; Estimated Marginal Means of purchase intention Figure 10; Estimated Marginal Means of doing nothing Figure 11; Estimated Marginal Means of doing nothing
Table 1; Distribution of the manipulated situation Table 2; Distribution of gender
Table 3; Distribution of marital status Table 4; Distribution of nationality Table 5; Distribution of education Table 6; Reliability test of complaining Table 7; Item-Total statistics of complaining
Table 8; Reliability test of (tendency to negative) word-of-mouth Table 9; Item-Total statistics of (tendency to negative) word-of-mouth Table 10; Reliability test of willingness to claim refund
Table 11; Item-Total statistics of willingness to claim refund Table 12; Reliability test of purchase intention
Table 13; Item-Total statistics of purchase intention Table 14; Reliability test of perceived price unfairness Table 15; Item-Total statistics of perceived price unfairness Table 16; Reliability test of product involvement
Table 20; Independent samples test
Table 21; Crosstabulation of perceived price unfairness and product involvement Table 22; Levene's Test of Equality of Error Variances for the dependent variable complaining
Table 23; Test of Between-Subjects Effects of the dependent variable complaining
Table 24; Levene's Test of Equality of Error Variances for the dependent variable (tendency to negative) word-of-mouth
Table 25; Test of Between-Subjects Effects of the dependent variable (tendency to negative) word-of-mouth
Table 26; Levene's Test of Equality of Error Variances for the dependent variable willingness to claim refund
Table 27; Test of Between-Subjects Effects of the dependent variable willingness to claim refund
Table 28; Levene's Test of Equality of Error Variances for the dependent variable purchase intention
Table 29; Test of Between-Subjects Effects of the dependent variable purchase intention Table 30; Levene's Test of Equality of Error Variances for the dependent variable doing nothing
Acknowledgements
I am using this opportunity to express my sincerest gratitude to my supervisor, Ms. Dr. Karin Venetis, who has supported me throughout my thesis with her continues support, patience, knowledge and motivation while giving me the space to work on my own speed due to personal circumstances. Her guidance helped me in all the time of research and writing of this thesis.
Furthermore I would like to thank my family, friends and acquaintances for helping me filling in and spreading out my questionnaire.
Special thanks to my friend Margreet Hummel with who I could deliberate, who gave me constructive feedback, supported me and scarified a few weekends so I was ‘forced’ to work on my thesis and my partner and parents (-in-law) who supported me on many ways throughout the period I needed to finish my thesis.
Thank you, Lies Ekhart
Abstract
Purpose – This study aims to improve the understanding of the influence of product
involvement on the relationship between perceived price unfairness and the behavioral reaction(s). The research question of this study is: “What is the effect of product involvement on the behavioral reaction(s) when price unfairness in perceived?”
Research methodology – This study builds on existing literature and theories on perceived
price unfairness, as well as literature on consumer behavior and product involvement. The research method is an explanatory, deductive, quantitative experiment using one method, a questionnaire collecting cross-sectional data. In the questionnaire respondents are asked how they perceive a specific situation in which a price is increased, how they will behave and how involved they are with the product. Because the situation is manipulated in perceived fairness and a product with a wide range of involvement is used, the influence of the manipulation and the difference between high and low product involvement can be investigated.
Results – The hypothesis ‘High product involvement has a negative influence on the
relationship between perceived price unfairness and the behavioral reaction(s)’ is not confirmed. The results confirm that respondents in the ‘unfair’ condition react stronger compared to respondents in the ‘fair’ condition and suggests that women and older respondents tend to react stronger on certain behaviour compared to men and younger respondents.
Research limitations – This study is limited to a Dutch context and to one specific product
and price increase situation. Future research could involve getting further insight of the effects of product involvement, gender and age by studying different countries, product categories
1.
Introduction
The emergence of dynamic price models, price comparison sites and the current economic crisis make consumers more aware of their spending pattern. It is a tough time for many companies and some of them do not make it and become bankrupt. The price of your product or service is crucial. These days a price increase can mean the end of your company. However sometimes a price increase is cannot be avoided. It is therefore very important to know how consumers will react on a price increase or price difference of a product or service compared to those of competitors.
The price offered for a certain product or service and the reason for offering that price may lead to a fairness perception of the price. Much research has been conducted on different variables that influence customer’s judgment of what is fair and what not. Most price evaluations that lead to a fairness assessment include comparisons. Perceived price (un)fairness is inferred when a consumer evaluates an outcome with a comparative other outcome. In the case of perceived price unfairness, these outcomes are prices (Lynn 1990, Monroe 2003). The comparison of prices may lead to three different outcomes:
o Prices are equal (price equity)
o Price is higher than the reference price (disadvantaged price inequity) o Price is lower than the reference price (advantaged price inequity)
If the compared price differs with the reference price, perceived price unfairness could occur. Research has shown that perceived price unfairness is a psychological factor that has an important influence on the behavioral reactions of consumers (Bougie, Pieters and Zeelenberg 2003, Raghunathan and Pham 1999, Xia, Monroe and Cox 2004, Zeelenberg and Pieters 2004). Consumers are unwilling to pay a price (increase) that they perceive as unfair (Martins and Monroe 1994, Kahneman, Knetsch, and Thaler 1986a, b) and will take actions towards these in their eyes unfair organizations, with all the additional (negative) consequences (Okun, 1981). Consumers that perceive a price (increase) as being unfair act to two elements:
o They want to protect themselves financially and seek for compensation if necessary o They need to cope with the negative emotions.
Different reactions towards a perceived unfair price are identified and Xia, Monroe and Cox (2004) categorized them into three groups:
o No action
o Self-protection (Complaining, ask for a refund, spread negative word-of-mouth or leave the relationship)
o Revenge (‘Get back’ to the seller or organization, spreading negative word-of-mouth, reporting to the media and legal actions)
Previous studies are addressing different aspects of perceived price unfairness, however, these studies do not provide any information about the influence of product involvement. This study would like to add the factor product involvement to an existing model to see whether high or low involvement of the product influences (the nature or tension of) the behavioral reaction(s) if a price (increase) is being perceived as unfair.
Previous researchers have used product involvement as an explanatory variable in consumer behavior and have shown that the level of involvement explains the complexity, extensiveness and depth of the information processing and behavioral reaction(s) during the consumer decision making process (Bloch and Richins 1983, Chakravarti and Janiszewski 2003, Dholakia 1997 and 1998, Houston and Rothschild 1978, Kleiser and Wagner 1999, Kokkinaki 1999, Laurent and Kapferer 1985, Petty et al. 1983, Phelps and Thorson 1991, Richins and Bloch 1986 and Richins, Bloch and McQuarrie 1992).
Product involvement is mostly categorized in high and low product involvement. The higher the involvement of the consumer the more he is willing to gain knowledge about the product and the more relevant and important the product is to him (Greenwald and Leavitt, 1984). Consumers with high product involvement will be more likely to be influenced by reference groups and will reach for, for some categories of goods and services, a reasoned decision (Kassarjian, 1981). Low product development on the other hand is characterized by the lack
The cognitive dissonance theory of Festinger (1957) is a conceptual model that predicts the outcome of certain social situations. The theory states that persons prefer consistency among their cognitions and will seek for it if inconsistency (dissonance) is experienced.
Lindsey-Mullikin (2003) studied the cognitive dissonance theory in relation to reactions on unexpected price changes and found several ways in which consumers reduce the experienced dissonance. Cohen and Goldberg (1970) found that more cognitive dissonance and associated dissonance reduction efforts could be expected in situations where the purchase is perceived as important, but also found that the expected reactions are not applicable for low involved purchases.
While these studies provide an important background, they do not provide insight on the effect of the degree of product involvement on the behavioral reaction(s) when a price (increase) is being perceived as unfair. Looking at the perceived price unfairness context, product involvement is an important variable. Perceived price unfairness triggers certain behavioral reactions and the degree of product involvement also influences behavioral reactions. The main focus in this study is to unite these research areas and to investigate the interactions between product involvement and the associated behavioral reactions when perceived price unfairness occurs.
Since product importance and product involvement are closely linked, it is expected that if product involvement is high the relationship between perceived price unfairness and the behavioral reaction(s) will be less positive compared to when the product involvement is low. Customers that are high involved put much effort in making the right decision, so when they perceive that the price (increase) they paid in unfair their behavioral reaction will be weaker, because they want to reduce the dissonance. Customers that are low involved lack commitment and effort in processing and/or gaining knowledge about the product. There is expected that their reactions will be stronger compared to high involved customers.
The aim of this study is to present a theoretical model, which will be an extension of previous literature about perceived price unfairness and product involvement and to include considerations of the managerial implications of the theoretical model. This study adds a new
If a certain degree of involvement leads to other behavior, marketing managers may wish to know what the segmentation implications for individuals with different levels of involvement are and how this might influence communication decisions. It broadens their understanding of the economical and psychological factors that influence the behavioral reactions of costumers to be able to effectively manage pricing strategies and (high and low involved) product strategies.
The primary objective of this study is to investigate the influence of product involvement on (the strength of) the behavioral reactions of consumers that perceive a certain price (increase) as unfair. The research question is:
o “What is the effect of product involvement on the behavioral reaction(s) when price unfairness is perceived?”
The research question is divided in the following sub-questions: o What is perceived price unfairness?
o Why is perceived price unfairness important in marketing? o What are the effects of perceived price unfairness?
o What is product involvement and why is it important in a perceived price unfairness context?
o How does product involvement influence the behavioral reaction(s) when perceived price unfairness occurs?
This thesis presents a theoretical model with a hypothesis that is tested after which the results are discussed and the research question is answered. This research concludes with the limitations of this research and suggestions for future research.
2.
Literature review
2.1 What is perceived price unfairness?
The price offered for a certain product or service and the reason for offering that price may lead to a fairness perception of the price. Xia, Monroe and Cox (2004) defined perceived price fairness as “a consumer's assessment and associated emotions of whether the difference (or lack of difference) between a seller's price and the price of a comparative other party is reasonable, acceptable or, justifiable” (p. 3).
Most price evaluations that lead to a fairness assessment include comparisons. Perceived price fairness is inferred when a consumer evaluates an outcome with a comparative other outcome. In the case of perceived price unfairness, these outcomes are prices (Lynn 1990, Monroe 2003). Price comparisons can be explicit and implicit. An explicit price comparison is comparing one price with another (range of) price(s). An implicit price comparison is made when one price is compared to an undefined lower price that is expected. This can be because you belong to a certain group based on your income, age, etc. Furthermore price comparisons are subjective and seen from the consumer’s point of view and are biased by self-interest (Oliver and Swan, 1989a). If the compared price differs with the reference price, perceived price unfairness could occur. Maxwell (2002) found that if the price equals the reference price, consumers will perceive the price as fair.
The theory of distributive justice states that costumers in an exchange relationship with others should receive a reward in all exchanges with others that is in balance to what is invested in the relationship (Homans, 1961). In 1965 Adams presented the equity theory, which complements the theory of distributive justice by emphases the importance of equality of the input and output in a relationship. So the equity theory is broader then the theory of distributive justice, because it adds comparative others that could influence the perceived fairness of an exchange relationship.
These comparative others could be: o Other persons
o Other organizations o Previous experience(s)
Only when others are not available, a self/self comparison will be made (Major, 1994).
The comparison of prices may lead to one of the following three different outcomes:
o Prices are equal, called ‘price equality’ which does not trigger a perception of fairness (Maxwell, 2002).
o Price is higher than the reference price, called ‘disadvantaged inequality’, which triggers a perception of unfairness
o Price is lower than the reference price, called ‘advantaged inequality’, which triggers a perception that is less fair (Ordóñez, Connorry and Coughlan, 2000).
Disadvantaged inequality and advantaged inequality both found differences in the paid price compared to the reference price. Therefore both prices can be perceived as being unfair, whether you are paying the higher or lower price. Although disadvantaged price inequity is more often perceived as being unfair compared to advantaged price inequity (Adams 1965, Austin, McGinn, and Susmilch 1980, Martins 1995, Van den Bos et al. 2006, Turow, Feldman and Meltzer 2005).
Besides the comparison of prices, there are also other potential factors influencing perceived price unfairness. Procedural justice for instance focuses on the influence of underlying procedures used to determine the outcome of fairness perceptions (Thibaut and Walker, 1975).
Xia, Monroe and Cox (2004) distinguish the following four groups of factors that influence the fairness assessment, although they vary in their relative scope:
Degree of similarity between transactions
The degree of similarity between transactions is an important element in comparing the products or services. If the products or services differ too much than a comparison will not be made, so there will not be a fairness judgment. For instance the time when or location where the product or service is bought can make a huge difference in the possibility of comparing it to a product or service that is bought on another time or location. Also factors like different brands, models or terms can make it harder to compare products or services. Another element is the comparability of the other consumer. It could be hard to compare if the other consumer belongs to another group, for instance because of age (Martins, 1995) or income.
Research shows that given a price discrepancy, a comparison with a similar other consumer leads to higher unfair perceptions. Moreover, when there is no price discrepancy, a comparison with a similar other consumer leads to a higher fair perceptions than a self/self comparison does (Xia, Monroe and Cox, 2004).
Reason for the price
If known, the reason for the price or the rule used to set the price can also be of great influence on the fairness perception. Prices based on the costs are perceived as being fair while prices based on supply and demand are perceived unfair (Dickson and Kalapurakai, 1994).
The “principle of dual entitlement” suggests that a price increase resulting in higher profits for the seller is perceived as unfair, but a price increase resulting from increased cost, so the profit of the seller remains the same, is perceived as fair (Bolton, Warlop and Alba 2003, Kahneman, Knetsch, and Thaler 1986 b, Urbany, Madden and Dickson 1989). So it is fair for an organization to raise prices when faced with increased costs and it is fair for an organization to maintain prices as costs decline. It is however unfair for an organization to benefit from shifts in demand by raising prices.
Frey and Pommerehne (1993) on the other hand found in their study that a price increase based on a higher demand is fairer when supply may expand, but just making a profit from demand is unfair.
Kalapurakal, Dickson and Urbany (1991) state that the fairness perception of the dual entitlement theory is influenced by external factors like information about the seller’ costs,
Although previously discussed factors influence fairness perception consumers do not always have (sufficient) information to make a good comparison. For instance the costs of the seller (Bolton, Warlop and Alba, 2003).
Previous experience
Consumers do not always act based on one transaction. Previous experiences with the seller, in which trust plays a great deal, also influence the behavioral reaction(s). If for instance the reason for the price increase in unknown, the consumer may assume that the price increase has a legitimate reason.
Social norms and general knowledge or beliefs about the sellers’ practices or the marketplace
Price fairness perceptions are also based on economical and social norm comparisons. Many price fairness perceptions are based on what consumers think of how the seller sets the prices and whether those prices are affordable to everyone. Because information is more accessible nowadays, like easily comparing prices on the Internet or sharing experiences with others, consumers develop knowledge of marketer’s pricing strategies and the composition of the cost-profit distribution (Bolton, Warlop and Alba, 2003).
A social norm develops when many consumers engage in the same behavior. Therefore social norms can evolve over time. In the hotel industry for instance, most consumers accept dynamic pricing based on yield management.
2.2 Why is perceived price unfairness important in marketing?
The importance of perceived price unfairness and the impact of a perceived unfair price on the profitability of an organization is already acknowledged by Kahneman, Knetsch and Thaler (1986 a, b). They investigated whether it was useful to extend the standard economical model of the profit maximizing organization with perceived price unfairness, since behavioral factors are being ignored in the basic economical model.
Consumers resist unfair prices and behavior and will take actions towards unfair organizations, with all the additional (negative) consequences (Okun, 1981). Consumers want to be treated fairly. Kahneman, Knetsch and Thaler (1986 a, b) found that the willingness of organizations to enforce fairness is common, even at some own costs. Fair behavior is an instrument for an organization that wants to maximize its profit on the long run.
An increase of the price of a few percentages can have an enormous influence on the net profit of an organization. However certain price increases will be profitable on the short term but can damage the organization on the long run.
An important task for many marketing managers is to establish product-pricing strategies. They have to take into account how consumers will respond to price differences or price changes. Research shows the importance of different psychological factors that influence the reactions of consumers to a certain price (Kamen and Toman 1970, Monroe 1973). Marketing managers need to understand these economical and psychological factors that influence the behavioral reactions of costumers to be able to effectively manage pricing strategies.
Xia, Monroe and Cox (2004) also underpin the importance of price fairness by revering to different news coverage’s on pricing, like Amazon.com’s dynamic pricing strategy. A customer of Amazon.com discovered that the price of certain DVDs differed on certain occasions. This outcome was a disaster, because many customers boycotted Amazon.com. (Adamy, 2000) This illustrates that the price offered and the reason for offering a certain price can lead to an unfair price perception, with all negative consequences.
Furthermore marketing managers see major changes in the last few years looking at the market and the behavior of its consumers. The price pressure is increasing. Nowadays there are different websites were you can compare the prices of a certain product or service with a lot of competitors. The introduction of the euro made it even easier to compare prices internationally. Consumers are becoming more critical, sceptical and empowered and are more and more negotiating with retailers. Only one third of all consumers trust advertisements. Choices are preferably made based on own experiences and experiences of others. Therefore social media is crucial for the success of an organization, product or service. The influence of social media is growing rapidly. For organizations it is important to have ambassadors and to work transparently and costumer-oriented.
Only having good products and services is not enough if a high price is asked. Factors like image, environment and social responsibility are becoming more important. Consumers want to justify their expenses.
Consumers are more aware of how to spend their money. In these times of economical crises, like the current European economic crises, affordability and a good balance between price and quality is crucial. Consumers are becoming more and more price sensitive and critical and want to be taken seriously. They do not want to pay too much and are constantly seeking for the best deal.
Although perceived price unfairness is mostly based on price differences, it is not said that fair pricing means the same price for everyone or that consumers do not accept price changes or differences.
2.3 What are the effects of perceived price unfairness?
Looking at the reactions of consumers to prices, perceived price unfairness is a psychological factor that has an important influence. Consumers are unwilling to pay a price (increase) that they perceive as unfair (Martins and Monroe 1994, Kahneman, Knetsch, and Thaler 1986a, b). Consumers that perceive a price (increase) as being unfair act to two elements:
o They want to protect themselves financially and seek for compensation if necessary o They need to cope with the negative emotions.
Some responses act more to the financial element and other responses have a more emotional nature. This also depends on the financial impact or risk that is taken.
Perceived price unfairness generate negative emotions that vary in type and intensity. Together with perceived value they influence the behavioral reaction(s) (Xia, Monroe and Cox, 2004). Different emotions will lead to different actions (Bougie, Pieters and Zeelenberg 2003, Raghunathan and Pham 1999, Zeelenberg and Pieters 2004).
There is however a difference between a disadvantaged price inequality, which shows emotions like disappointment, anger or outrage (Austin, McGinn and Susmilch, 1980) and advantaged price inequality, which generates emotions like uneasiness and guilt. Bougie, Pieters and Zeelenberg (2003) state that advantaged price inequality will probably not lead to actions.
If a price is being judged as unfair the feeling(s) and action(s) associated with that perception are being directed towards the seller or the party that caused you being in that unfair situation. Not towards the other person that got a lower price.
Urbany, Madden and Dickson (1989) identified the following three factors that influence the behavioral reactions of the consumers:
o Their power o Change to succeed o Time, effort and money
But also personal characteristics have influence on the price perception and the action(s) that will be taken.
Consumers that perceive unfair prices are looking for alternatives. For several years many researchers investigated the different behavioral reactions on perceived price unfairness (Bougie, Pieters and Zeelenberg 2003, Campbell 1999, Martins 1995, Okun 1981, Raghunathan and Pham 1999, Urbany, Madden and Dickson 1989, Zeelenberg and Pieters 2004). Xia, Monroe and Cox (2004) subdivided these behavioral reactions to perceived price unfairness in three groups:
o No action o Self-protection o Revenge
No action
No actions means no significant influence on consumers planned transaction(s) with the seller. This happens mostly in advantaged situations or when the consumer feels slightly disadvantaged. Also when the costs are too high or the effort is too much compared to the perceived unfairness, no action will be taken (Urbany, Madden and Dickson, 1989).
Self-protection
The following behavioral reactions fall under the group called self-protection: o Complaining
o Ask for a refund
o Spread negative word-of-mouth o Leave the relationship
The chosen action depends on what is doing justice to the negative emotions that are being felt or what will restore equity with the least costs. These actions are mostly taken to prevent the consumer from being taken advantage of in the future.
Spreading negative word-of-mouth is an action that does not cost any money and prevents other persons in the social network from making the same mistake and it helps the consumer
Revenge
The following behavioral reactions fall into the category revenge: o Switching to the company’s direct competitor
o Also spreading negative word-of-mouth, although driven by anger purely with the goal to damage the seller. (Spreading negative word-of-mouth in the category self-protection is to cope with own feelings of unease.)
o Reporting to the media o Legal actions
The actions taken in this category are driven by strong negative feelings like anger and outrage and are mostly immediately taken. The actions mentioned are not enough to restore equity. Anger may even lead to aggressive behavior (Bougie, Pieters and Zeelenburg, 2003).
Xia, Monroe and Cox (2004) made different hypotheses out of their macro analysis. In one of these hypotheses they assume that value for money will be the major drive, as a consumer perceives a price as less fair in a disadvantaged inequality situation, for their actions. They will evaluate the costs of their actions and will take action to seek financial compensation or they will not take any action. Furthermore they expect that negative emotions will be a major drive for their actions, as a consumer perceives a price as unfair. In this situation they will deal with their negative emotions by spreading negative word-to-mouth or seek for revenge. These actions are taken to ‘get even’ with the seller and find emotional compensation.
2.4 What is product involvement and why is it important in a perceived price unfairness context?
There are many different names and definitions to describe what in this study will be called product involvement. Product involvement is also a construct that is studied in multiple disciplines, like social psychology, consumer psychology, etc. Product involvement is commonly defined as the personal relevance of the product based on inherent needs, values and interests (Zaichkowsky, 1985).
The level of involvement for a particular product or service varies across consumers and it affects the search for and processing of available information of the product or service
(Greenwald and Leavitt 1984, Houston and Rothschild 1978, Kassarjian 1981). The higher the involvement of the consumer the more he is willing to gain knowledge about the product, evaluate it critically and the more relevant and important the product is to him. Low product involvement on the other hand is characterized by the lack of commitment and effort in processing and/or gaining knowledge about the product. These consumers tend to use simple heuristics.
According to Laurent and Kapferer (1985) the degree of involvement of a consumer is influenced by four factors:
o The importance of the product
o The danger of making a wrong decision, the risk of purchasing the product o The symbolic value of the product
o The value of the pleasure the product delivers
Product involvement is mostly categorized in three types of involvement: o Situational involvement (SI)
SI is the degree of involvement that exists in certain specific situations, for instance, a purchase occasion. When a purchase occasion is near the involvement with that certain
o Enduring involvement (EI)
EI is the ongoing concern with a certain product or service. Mostly this concerns products and services that hold the consumer's thoughts and occupation without the stimulant of a direct purchase. EI transcends SI, (Houston and Rothschild 1978, Laurent and Kapferer 1985) but consumer's El with most products and services is likely to be low. There are certain products and services for which a broad range of El is applicable, for instance cars (Bloch, 1982).
o Response involvement (RI)
RI is the way in which SI and EI influence the costumer decision-making process, so how consumers respond/behave. Houstan and Rothshild (1978) defined RI as “the complexity of cognitive and other processes at various stages of the decision process”. Many persons use a computer daily but spend little time talking, learning or thinking about it, which indicates a low EI. Other persons may have a high(er) EI with computers, like gamers, web-designers, hackers and other persons who belong to certain computer communities, subscribe to computer magazines or attend computer events. Although some persons have a low EI on a daily basis with computers, their SI can be high during the purchase of a new computer. This also counts for other (high risk) products or services.
Although SI and EI are two different types of involvement with their own motivations and temporary pattern of their occurrence, they both represent a state of arousal and product interest. Because the focus of this study will be on the influence of the overall degree of product involvement after an unfair price (increase) is perceived, there will not be made a difference between SI and EI. Motivation and duration of the involvement will not be in the scope of this study. There will only be made a contradiction between high and low product involvement.
Previous researchers have used product involvement as an explanatory variable in consumer behavior and have shown that the level of involvement explains the complexity, extensiveness and depth of the information processing and behavioral reaction during the consumer decision making process (Bloch and Richins 1983, Chakravarti and Janiszewski 2003, Dholakia 1997 and 1998, Houston and Rothschild 1978, Kokkinaki 1999, Kleiser and Wagner 1999, Laurent
Kassarjian (1981) found that consumers with high product involvement would sooner reach for a reasoned decision. Petty et al (1983) complements this finding by stating that low involved consumers generate more positive thoughts to support their arguments, whereas high involved consumers elaborate more and are open for counter arguments.
One can conclude that the degree of product involvement certainly has influence and is vital in understanding how consumers process information, make decisions and how they behave. The focus in previous studies is mainly on pre-purchase decisions and behavior.
Looking at the perceived price unfairness context, product involvement is an important variable to look at. Perceived price unfairness triggers certain behavioral reactions and the degree of product involvement also influences behavioral reactions. The main focus in this study is to unite these variables and to further investigate the interactions between product involvement, perceived price unfairness and the associated behavioral reactions.
2.5 How does product involvement influence the behavioral reaction(s) when perceived price unfairness occurs?
In the previous paragraphs is discussed what perceived price unfairness is, why perceived price unfairness is important in marketing, what the effects are of an perceived unfair price and what product involvement is and how it relates to perceived price unfairness. Now the question is how product involvement influences (the strength of) the behavioral reaction(s) when perceived price unfairness occurs.
The cognitive dissonance theory of Festinger (1957) is a conceptual model that predicts the outcome of certain social situations. The theory states that persons prefer consistency among their cognitions and will seek for it if inconsistency (dissonance) is experienced. Cognitions are values, beliefs, reactions, opinions, etc. This means that consumers will minimize every difference between their experiences and expectations by adapting their perception or taking a certain action so it will be more consistent with their expectations (Balcetis and Dunning, 2007).
Lindsey-Mullikin (2003) showed in a previous study that consumers confronted with unexpected price changes react on one of the following three ways to reduce the dissonance:
o Seeking constant information that support their first belief
o Change of attitude, re-evaluating the price in comparison to the (expected) quality o Trivializing the importance of getting the best deal
This study shows the importance of cognitive dissonance to consumer behavior when talking about (unexpected) price changes.
Sweeney et al. (2000) complemented this study by identifying three conditions before such reactions in purchase decisions occur, which are:
o The importance of the purchase for the consumer o The consumer should be able to select alternatives o The decision should be irreversible
Sweeney et al. (2000) ads product importance as a condition before the reactions on an unexpected price occurs. Cohen and Goldberg (1970) found that more cognitive dissonance
linked, which could indicate that there could be a difference between low en high involvement purchases and the associated behavioral reactions. This study found that the expected reactions described by Lindsey-Mullikin (2003) are not applicable for low involvement purchases they investigated, but it is not clear to which extend this assumption can be generalized.
In this study will be investigated if the difference in low and high involvement previously found in research concerning the cognitive dissonance theory also occurs on the different behavioural reactions when an unfair price (increase) is perceived.
The next paragraph will combine previously discussed literature into assumptions, a theoretical model and a hypothesis that will be tested.
2.6 Theoretical model and hypothesis
It is proposed that product involvement (high vs. low) will influence the relation between perceived price unfairness and the behavioral reaction(s).
Looking at the cognitive dissonance theory there can be expected that high involved customers are more likely to produce cognitive behavior such as making counterarguments, derogate themselves from certain sources or trivialize the perceived unfair price than low involved customers will do, because they are more focused to restore their experienced dissonance. So whereas product involvement is high the relationship between perceived price unfairness and the behavioral reaction(s) will be less positive in contradiction to when the product involvement is low. Customers that are high involved put much effort in making the right decision, when they perceive that the price (increase) they paid in unfair their behavioral reaction will be weaker, because they want to reduce the dissonance. Customers that are low involved lack commitment and effort in processing and/or gaining knowledge about the product. It is expected that their reactions will be stronger compared to high involved customers.
Based on the literature discussed in the previous paragraphs and the expectations discussed above a theoretical model is made which is shown in figure 1 and the following hypothesis is formulated:
H1: High product involvement has a negative influence on the relationship between perceived price unfairness and the behavioral reaction(s)
3.
Research methodology
This chapter will elaborate the research design of this study, the design of the questionnaire, the way data were collected and how the data are analysed.
3.1 Research design
Saunders et al. (2009) specify three purposes of research; explanatory, exploratory and descriptive research. Because this research focuses on establishing relationships between variables based on quantitative data, the purpose of this research is explanatory.
There are different approaches. Saunders et al. (2009) distinguish two approaches; an inductive versus a deductive approach of research. The deductive approach is about testing a theory and is most used in science. An inductive approach builds a theory and tries to generalize collected data, trough data analysis, into a more general theory. This study is an example of a deductive approach. Based on literature certain expectations are formulated after which the research question is presented.
There are different strategies when choosing for a quantitative approach. Saunders et al. (2009) define seven different strategies. The choice of research strategy will be based upon the research question(s), the amount of existing knowledge, data and time available and other resources. According to Yin (2009) each strategy can be used, whether the study has an explanatory, exploratory or descriptive purpose. This study is an experiment to see the effect of manipulating perceived price unfairness, an example of studying causal links, a characteristic of an experiment. This research is a between subjects experiment.
One data collection technique was used, a questionnaire, which makes this a mono method study. The time horizon of this study is cross-sectional, because the data collected from the
3.2 Questionnaire design
When conducting a questionnaire, there is only one change to collect the data. It is not possible to ask additional information like an in-depth interview, because respondent are mostly not traceable. The design of the questionnaire is of great importance to the response rate, reliability and validity of the data that was collected. When designing a questionnaire the situation and questions used need to be adapted to the variables of the conceptual model, so there will be asked for those things that need to be measured.
The following variables will be measured:
o Perceived price unfairness, the independent variable, which will be measured using the scale of Campbell (2007).
o Product involvement, the moderating variable, which will be measured using the scale of O’ Cass (2000). O’ Cass (2000) validated a scale that measures four forms of involvement; product involvement, involvement with the purchase intention, involvement with advertisement and consumption of a product. This research will confine itself to the questions measuring product involvement, which are six items. o The behavioral responses are the dependent variables, which will be measured by the
following scales:
o Complaining, which will be measured by adapting items for behavioral responses from Zeelenberg and Pieters (2004).
o (Tendency to negative) Word-of-mouth, which will be measured using three items of Coombs and Holladay (2008).
o Willingness to claim refund, which will be measured using four items based on a scale from Srivastava (1999).
o Purchase intention, which will be measured using two items of Kamins and Gupta (1994).
o Doing nothing, which will be measured using the statement: “I do nothing”.
Next to these variables, there need to be controlled for other variables that can influence the given answers. This is done by asking demographic questions like age, gender, marital status, nationality and education. These demographic questions are also used to make sure the sample represents the Dutch population.
Dillman (2007) identified three different types of variables, opinion variables, behavioral variables and attribute variables, which all will be used in this study. The type of questions used in the questionnaire are rating questions in a matrix format using a seven-point Likert scale, list questions and open questions.
When conducting a questionnaire be aware of several biases. The participants should be equally divided over the groups taking in consideration a good balance between male and female to minimize gender effects. Subject or participant bias is decreased, because the questionnaire is self-administered. This makes it less likely that participants may say what they think is expected from them to say, in other words give favourable answers. Being anonymous when filling in the questionnaire also decreases participant bias. Observer bias is not likely, because of the use of a predefined, self-administered, Internet mediated questionnaire. Making use of this kind of questionnaire also has some disadvantages. The response rate is difficult to measure because of the use of social networks, in which it is possible for persons to forward the questionnaire without knowing this. Saunders et al. (2009) indicate a response rate of 11% or lower when using Internet. Another disadvantage of self-administered questionnaires is the fact that the participants can discuss their answers with other persons, which can cause contamination of the answers.
Because the questionnaire will be spread under the Dutch population, the items of the scales are translated into Dutch using the direct translation technique.
3.3 Data collection
The data collection process involves several steps. First a good questionnaire needs to be designed. Then a clear explanation of the purpose of the questionnaire needs to be written to introduce its topic and to increase the response rate. Then a pre-test need to be conducted, after which the questionnaire needs to be adjusted according to the finding of the pre-test.
The Dutch questionnaire can be found in appendix A. Most questions of the questionnaire are programmed on such a way that an answer must be given, so there will be almost no missing data.
Before the actual data will be collected a pre-test will be conducted to ensure that participants understand the items and meaningful answers will be generated. The translation will be checked and out of the pre-test one or two products will be selected based on their value on product involvement. One product will be selected if the product involvement has a normal distribution or two products if one is rated as a high involved product and one is rated as a low involved product. Furthermore two manipulated situations of a price increase need to be confirmed as a fair and an unfair situation. For the pre-test ten persons per group will be selected.
For the actual data collection a minimum of 100 questionnaires per group need to be gathered amongst the Dutch population. The participants are randomly assigned to one of the two situations. The questionnaire is spread out amongst family, friends and colleagues and was accessible through existing social networks on Internet. As a result a wide variety of people is reached. This type of questionnaire is called an internet-mediated questionnaire (Saunders et al, 2009). A limitation of this way of collecting data is the amount of questionnaires that is already been spread out through social networks. Nowadays many persons are contacted by friends and family members to fill in a questionnaire, so a good explanation about the purpose of the questionnaire is crucial.
Pre-test
In the first pre-test two products and two situations manipulated in their fairness were used to test their level of involvement with the different products and the level of fairness of the manipulated situations.
The two products:
• Cars – In first instance indicated by Bloch (1981) as a high involved product, but later on indicated by Bloch (1982) as a product with a normal distribution.
• Car insurance – indicated by the researcher as a low involved product.
Cars should be indicated as a high involved product or product with a normal distribution. When it is indicated as a high involved product, car insurance should be indicated as a low-involved product. The manipulated situations are based on previous studies of Campbell (1999) and Kahneman, Knetsch and Thaler (1986 a and b).
Two different questionnaires each containing two different products in two different situations were made. Twenty persons were asked to fill in a randomly selected questionnaire, so each situation was filled in by ten persons. In this case twenty persons assessed the level of involvement of the two different products and ten persons assessed the two manipulated situations. Out of this first pre-test the product involvement of cars was not indicated to be a high involved product (mean of 4.8) or product with a normal distribution (Sig.>0.05 in the test of normality), so a second pre-test was necessary to find another product that would be indicated as a high involved product or a product with a normal distribution.
In the second pre-test two other products were used in the same two manipulated situations. • Clothing - indicated by Bloch (1982) as a product with a normal distribution
• Mobile phones - indicated by White and Young (2010) as a product with a normal distribution
3.4 Data analysis
Because of the use of an Internet survey tool, the collected data could be directly uploaded in the statistical program SPSS. First the data was checked, which means looking for illegitimate codes and relationships. If this check is not done, it can cause incorrect results from which false conclusions can be drawn. After the data was checked, an overview of all respondents was made, called the sample.
The Cronbach’s alpha is used to calculate the reliability of all scales. Cronbach’s alpha is a commonly used test of internal reliability and varies between 1 (perfect internal reliability) and 0 (no internal reliability). The value of Cronbach’s alpha depends on the number of items of a scale. When a scale is measured with less then ten items, the value of Cronbach’s alpha can be quit small. Nunnally (1978) and DeVellis (2003) recommend a minimum of 0,7 for an acceptable internal reliability. All scales used in this research have less then ten items. If Cronbach’s alpha is less then 0.7 the mean inter-item correlation is calculated. Briggs and Cheek (1986) recommends a range from 0.2 to 0.4 for an optimal value of the mean inter-item correlation.
Next the descriptive statistics are generated to gain a clear overview of and compare the central tendency of all scales of the questionnaire. An independent samples t-test tests if the conditions (‘Fair’ and ‘Unfair’) significantly differ from each other. Furthermore the involvement scores of the respondents are split up to create a high and low involved group.
A two-way between-groups ANOVA test was conducted to test the theoretical model shown in figure 1 on page 24. After this test a conclusion can be drawn on how product involvement influences the behavioral reactions to perceived price unfairness.
4.
Results
4.1 Sample
In this paragraph the characteristics of the sample will be described. 860 respondents filled in a questionnaire, but 75 had to be removed because of missing values. The remaining sample consists of 685 respondents, 50.7% in the ‘fair’ situation and 49.3% in the ‘unfair’ situation as shown in table 1.
Manipulated situation
Frequency Percent Valid Percent
Cumulative Percent
Valid Unfair 347 50,7 50,7 50,7
Fair 338 49,3 49,3 100,0
Total 685 100,0 100,0
Table 1; Distribution of the manipulated situation
The average age of the respondents was 46.0 years with a standard deviation of 13 years. A large majority of the respondents were male (62%), Dutch (98%), married or living together (83%) and had finished a MBO- or HBO education (64%). The distribution of these characteristics can be found in the tables 2 to 5.
Gender
Frequency Percent Valid Percent
Cumulative Percent
Valid Male 425 62,0 62,0 62,0
Female 260 38,0 38,0 100,0
Total 685 100,0 100,0
Table 2; Distribution of gender
Marital status
Nationality
Frequency Percent Valid Percent
Cumulative Percent
Valid Dutch 671 98,0 98,0 98,0
Other/I do not know/I do not want to say
14 2,0 2,0 100,0
Total 685 100,0 100,0
Table 4; Distribution of nationality
Education
Frequency Percent Valid Percent
Cumulative Percent Valid lbo/vmbo/mavo 43 6,3 6,3 6,3 havo/vwo 80 11,7 11,7 18,0 mbo 146 21,3 21,3 39,3 hbo 292 42,6 42,6 81,9 wo 113 16,5 16,5 98,4
Other/I do not know/I do not want to say
11 1,6 1,6 100,0
Total 685 100,0 100,0
Table 5; Distribution of education
4.2 Reliability analyses
This paragraph describes the reliability analyses of the scales. Scales are considered to be reliable when Cronbach’s alpha is at least 0.70, and all item-total correlations are positive. Items with low item-total correlations can be removed if this would increase Cronbach’s alpha with at least 0.05.
Reversed items were recoded before executing the analyses, with the following scheme: 7=1, 1=7, 2=6, 6=2, 3=5, 5=3, 4=4. These items can be recognized in the tables because these names end with an R.
Complaining
The scale measuring complaining behavior is very reliable with a Cronbach’s alpha of 0.91. Cronbach’s alpha cannot be further increased by removing one or more items. See table 6 and 7.
Complaining
Cronbach’s Alpha N of Items
,906 2
Table 6; Reliability test of complaining
Complaining Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted Ik ga de verkoper om compensatie vragen 4,300 5,432 ,827 .
Ik ga klagen bij de verkoper 4,172 5,458 ,827 .
Table 7; Item-Total statistics of complaining
(Tendency to negative) Word-of-mouth
The scale measuring (tendency to negative) word-of-mouth is almost reliable with a Cronbach’s alpha of 0.69. Cronbach’s alpha can be improved to 0.77 by removing the first (reversed) item. The final scale consists of two items then. See table 8 and 9.
(Tendency to negative) Word-of-mouth
Cronbach's Alpha N of Items
,686 3
Table 8; Reliability test of (tendency to negative) word-of-mouth
(Tendency to negative) Word-of-mouth
Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted
Willingness to claim refund
The scale measuring the willingness to claim a refund is reliable with a Cronbach’s alpha of 0.94. Cronbach’s alpha cannot be increased by 0.05 or more by removing one or more items. See table 10 and 11.
Willingness to claim refund
Cronbach's Alpha N of Items
,937 4
Table 10; Reliability test of willingness to claim refund
Willingness to claim refund
Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted De waarschijnlijkheid is zeer hoog dat ik
deze winkel om een terugbetaling van het prijsverschil vraag
12,9224 39,139 ,914 ,896
Het is zeer waarschijnlijk dat ik een terugbetaling van deze winkel eis
12,5522 41,695 ,853 ,917
Het is zeer waarschijnlijk dat ik terug zal gaan naar de winkel om de lagere prijs voor de mobiele telefoon te verkrijgen
13,1537 40,435 ,852 ,917
Mijn wil om deze winkel om een
terugbetaling van het prijsverschil te vragen is zeer laag - R
13,0746 42,141 ,785 ,938
Table 11; Item-Total statistics of willingness to claim refund
Purchase intention
The scale measuring negative purchase intention, in other words how strong the intention is not to buy at this store anymore is also reliable with a Cronbach’s alpha of 0.71. Cronbach’s alpha cannot be further increased by removing one or more items. See table 12 and 13.
Purchase intention
Cronbach's Alpha N of Items
,708 2
Purchase intention Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted De kans dat ik de volgende keer wanneer ik
een mobiele telefoon nodig heb bij deze winkel zal kopen is zeer groot - R
2,9050 3,432 ,552 .
De kans is zeer klein dat ik terug zal keren naar deze winkel, wanneer ik een mobiele telefoon zoek
2,6558 2,725 ,552 .
Table 13; Item-Total statistics of purchase intention
Perceived price unfairness
The scale measuring perceived price unfairness is reliable with a Cronbach’s alpha of 0.882. Cronbach’s alpha can be improved towards 0.926 by removing the last item, but because this is not an increase of 0.05 or more, the scale will be left the way it is, with 4 items. See table 14 and 15.
Perceived price unfairness
Cronbach’s Alpha N of Items
,882 4
Table 14; Reliability test of perceived price unfairness
Perceived price unfairness
Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted De prijsstijging is oneerlijk 12,114 24,953 ,796 ,828 De prijsstijging is onacceptabel 12,001 24,757 ,836 ,811
Ik heb het gevoel dat de prijsstijging onredelijk is
12,349 24,716 ,843 ,808
Ik heb het gevoel dat de prijsstijging die ik gekregen heb hoger is dan dat anderen krijgen
11,366 31,700 ,517 ,926
Product involvement
The scale measuring product involvement is reliable with a Cronbach’s alpha of 0.90. Cronbach’s alpha cannot be further increased by removing one or more items. See table 16 and 17.
Product involvement
Cronbach's Alpha N of Items
,901 6
Table 16; Reliability test of product involvement
Product involvement Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted Mobiele telefoons betekenen veel
voor mij
23,753 52,425 ,644 ,896
Mobiele telefoons maken een belangrijk deel uit van mijn leven
23,652 52,224 ,639 ,897
Ik denk vaak aan mobiele telefoons 21,984 50,627 ,782 ,875
Ik ben erg geïnteresseerd in mobiele telefoons
22,335 48,891 ,796 ,873
Ik kan me echt identificeren met mobiele telefoons
21,634 51,922 ,726 ,884
Ik besteed veel aandacht aan mobiele telefoons
22,030 48,989 ,797 ,872
Table 17; Item-Total statistics of product involvement
4.3 Descriptive statistics
Table 18 shows the descriptive statistics of the scales. The scales are measured with a seven-point Likert scale, in which 1 stands for ‘strongly agree’ and 7 for ‘strongly disagree’. Most scales have an average score around the ‘neutral’ score 4. An exception is purchase intention, which has an average score of 2.8. This is between 2 ‘agree’ and 3 ‘slightly agree’.
Descriptive statistics
N Minimum Maximum Mean Std. Deviation
Perceived price unfairness 685 1,00 7,00 3,9822 1,68485
Product involvement 685 1,00 7,00 4,5020 1,41605
Complaining 685 1,00 7,00 4,2343 2,22875
(Tendency to negative) Word-of-mouth 685 1,00 7,00 4,1956 1,84200
The Levene’s test for equality of variances in table 20 shows that Sig. is lower than 0.05, which indicates that the variance of scores of the two groups is the same. In table 19 and 20 it can be seen that respondents in the ‘unfair’ condition rated the situation to be significantly more unfair with a mean of 3.6, which is between ‘slightly agree’ and ‘neutral’, compared to respondents in the ‘fair’ condition who had an mean score of 4.4, which is between ‘neutral’ and ‘slightly disagree’; t (683) = -6.5, p<0.001 (two-tailed).
Group Statistics
Situation N Mean Std. Deviation Std. Error Mean
Perceived price unfairness
Unfair 347 3,5790 1,59300 ,08552
Fair 338 4,3962 1,67839 ,09129
Table 19; Group statistics
Independent Samples Test
Levene’s Test for Equality
of Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Perceived price unfairness Equal variances assumed ,520 ,471 -6,537 683 ,000 -,81719 ,12500 -1,06263 -,57176 Equal variances not assumed -6,533 678,822 ,000 -,81719 ,12509 -1,06280 -,57158
Table 20; Independent samples test
To distinguish between low and high involved respondents, the involvement scores were split at the median score of 4.7. Involvement scores lower than 4.7 are labelled as ‘high involved’ and involvement scores higher than 4.7 are labelled as ‘low involved’. The distribution of the respondents over the four conditions is shown in table 21.
4.3 Testing of model
In this paragraph, the theoretical model (figure 1 on page 24) will be tested. Each behavioral reaction will be discussed.
Complaining
In table 22 can be seen that the error variance of the dependent variable ‘complaining’ is equal across groups. The assumption of equality of error variances is not violated. (Sig. > 0.05)
Levene's Test of Equality of Error Variancesa
Dependent Variable: Complaining
F df1 df2 Sig.
1,044 120 564 ,369
Tests the null hypothesis that the error variance of the dependent variable is equal across groups.
a. Design: Intercept + Condition: Perceived price unfairness + Product involvement: High/Low + Wat is uw geslacht + Wat is uw burgerlijke staat + Wat is uw nationaliteit + Wat is uw leeftijd + Wat is uw hoogst voltooide opleiding + Condition: Perceived price unfairness * Product involvement: High/Low
Table 22; Levene's Test of Equality of Error Variances for the dependent variable complaining
The two-way between-groups analysis of variance was conducted and demonstrates significant effects of the experimental condition (perceived price unfairness), gender and age. The hypothesis ‘High product involvement has a negative influence on the relationship between perceived price unfairness and the behavioral reaction(s)’ is not confirmed, because the interaction between ‘Condition: Perceived price unfairness’ and ‘Product involvement: High/Low’ is not significant.
Tests of Between-Subjects Effects
Dependent Variable: Complaining
Source Type III Sum of Squares df Mean Square F Sig.
Corrected Model 127,842a 14 9,132 1,871 ,026
Intercept 546,973 1 546,973 112,078 ,000
Condition: Perceived price unfairness 47,014 1 47,014 9,634 ,002
Product involvement: High/Low 18,384 1 18,384 3,767 ,053
Wat is uw geslacht 25,374 1 25,374 5,199 ,023
Wat is uw burgerlijke staat 15,141 3 5,047 1,034 ,377
Wat is uw nationaliteit ,191 1 ,191 ,039 ,843
Wat is uw leeftijd 34,687 1 34,687 7,108 ,008
Wat is uw hoogst voltooide opleiding 9,543 5 1,909 ,391 ,855
Condition: Perceived price unfairness * Product involvement: High/Low
2,020 1 2,020 ,414 ,520
Figure 2 demonstrates the main effect of perceived price unfairness. Respondents in the ‘unfair’ condition complain more (average score 3.7; slightly agree - neutral) compared to respondents in the ‘fair’ condition (average score 4.2; neutral - slightly disagree). The Figure also seems to indicate that highly involved respondents complain more (the blue line is below the green line), however this effect is not significant (p=0.053).
Figure 2; Estimated Marginal Means of Figure 3; Estimated Marginal Means of complaining complaining
Figure 3 shows that women tend to complain more, compared to men. Parameter estimates (not shown) indicate a negative age-coefficient, which means that older respondents complain more.
(Tendency to negative) word-of-mouth
In table 24 can be seen that the error variance of the dependent variable ‘(Tendency to negative) word-of-mouth’ is equal across groups. The assumption of equality of error variances is not violated. (Sig. > 0.05)
Levene's Test of Equality of Error Variancesa
The two-way between-groups analysis of variance was conducted and demonstrates significant effects of the experimental condition (perceived price unfairness) and gender. The hypothesis ‘High product involvement has a negative influence on the relationship between perceived price unfairness and the behavioral reaction(s)’ is not confirmed, because the interaction between ‘Condition: Perceived price unfairness’ and ‘Product involvement: High/Low’ is not significant.
Tests of Between-Subjects Effects
Dependent Variable: (Tendency to negative) word-of-mouth
Source Type III Sum of Squares df Mean Square F Sig.
Corrected Model 161,094a 14 11,507 3,570 ,000
Intercept 438,548 1 438,548 136,050 ,000
Condition: Perceived price unfairness 118,082 1 118,082 36,632 ,000
Product involvement: High/Low ,106 1 ,106 ,033 ,856
Wat is uw geslacht 15,818 1 15,818 4,907 ,027
Wat is uw burgerlijke staat 8,386 3 2,795 ,867 ,458
Wat is uw nationaliteit ,145 1 ,145 ,045 ,832
Wat is uw leeftijd 1,913 1 1,913 ,594 ,441
Wat is uw hoogst voltooide opleiding 20,314 5 4,063 1,260 ,279
Condition: Perceived price unfairness * Product involvement: High/Low
4,882 1 4,882 1,515 ,219
Error 2159,693 670 3,223
Total 14379,000 685
Corrected Total 2320,787 684
a. R Squared = ,069 (Adjusted R Squared = ,050)
Table 25; Test of Between-Subjects Effects of the dependent variable (tendency to negative) word-of-mouth Figure 4 demonstrates the main effect of perceived price unfairness. Respondents in the ‘unfair’ condition tend to express more negative word-of-mouth (average score 3.7; slightly agree-neutral) compared to respondents in the ‘fair’ condition (average score 4.6; neutral - slightly disagree). Figure 5 shows that women tend to express more negative word-of-mouth, compared to men.
Figure 4; Estimated Marginal Means of Figure 5; Estimated Marginal Means of
(tendency to negative) word-of-mouth (tendency to negative) word-of-mouth
Willingness to claim refund
In table 26 can be seen that the error variance of the dependent variable ‘willingness to claim refund’ is equal across groups. The assumption of equality of error variances is not violated. (Sig. > 0.05)
Levene's Test of Equality of Error Variancesa
Dependent Variable: Willingness to claim refund
F df1 df2 Sig.
1,084 120 564 ,273
Tests the null hypothesis that the error variance of the dependent variable is equal across groups.
a. Design: Intercept + Condition: Perceived price unfairness + Product involvement: High/Low + Wat is uw geslacht + Wat is uw burgerlijke staat + Wat is uw nationaliteit + Wat is uw leeftijd + Wat is uw hoogst voltooide opleiding + Condition: Perceived price unfairness * Product involvement: High/Low
Table 26; Levene's Test of Equality of Error Variances for the dependent variable willingness to claim refund
The two-way between-groups analysis of variance was conducted and demonstrates significant effects of the experimental condition (perceived price unfairness), gender and age. The hypothesis ‘High product involvement has a negative influence on the relationship
Tests of Between-Subjects Effects
Dependent Variable: Willingness to claim refund
Source Type III Sum of Squares df Mean Square F Sig.
Corrected Model 132,920a 14 9,494 2,190 ,007
Intercept 671,409 1 671,409 154,891 ,000
Condition: Perceived price unfairness 63,143 1 63,143 14,567 ,000
Product involvement: High/Low 3,705 1 3,705 ,855 ,356
Wat is uw geslacht 19,411 1 19,411 4,478 ,035
Wat is uw burgerlijke staat 4,872 3 1,624 ,375 ,771
Wat is uw nationaliteit ,006 1 ,006 ,001 ,970
Wat is uw leeftijd 49,096 1 49,096 11,326 ,001
Wat is uw hoogst voltooide opleiding 6,745 5 1,349 ,311 ,906
Condition: Perceived price unfairness * Product involvement: High/Low
2,383 1 2,383 ,550 ,459
Error 2904,262 670 4,335
Total 15785,382 685
Corrected Total 3037,181 684
a. R Squared = ,044 (Adjusted R Squared = ,024)
Table 27; Test of Between-Subjects Effects of the dependent variable willingness to claim refund
Figure 6 demonstrates the main effect of perceived price unfairness. Respondents in the ‘unfair’ condition are more willing to claim a refund (average score 4.0; neutral) compared to respondents in the ‘fair’ condition (average score 4.6; neutral - slightly disagree). Figure 7 shows that women are more willing to claim a refund, compared to men. Parameter estimates (not shown) indicate a negative age-coefficient, which means that older respondents are more willing to claim a refund.
Figure 6; Estimated Marginal Means of Figure 7; Estimated Marginal Means of
