What effect consumers intention to buy counterfeit luxury brands the moderating role of product involvement and product knowledge evidence from vietnam

35 Figure 4.2: Moderation effect of product involvement on the relationship between value-expressive function and counterfeit purchase intention .... 40 Figure 4.3: Moderation effect of

-

Nguyễn Hạo Nhiên

WHAT EFFECT CONSUMERS’ INTENTION TO BUY

COUNTERFEIT LUXURY BRANDS? THE MODERATING ROLE

OF PRODUCT INVOLVEMENT AND PRODUCT KNOWLEDGE:

EVIDENCE FROM VIETNAM

Ho Chi Minh City – Year 2015

-

Nguyễn Hạo Nhiên

WHAT EFFECT CONSUMERS’ INTENTION TO BUY

COUNTERFEIT LUXURY BRANDS? THE MODERATING ROLE

OF PRODUCT INVOLVEMENT AND PRODUCT KNOWLEDGE:

EVIDENCE FROM VIETNAM

ID: 22130049

SUPERVISOR: Dr Ngô Viết Liêm

Ho Chi Minh City – Year 2015

Thirdly, I would like to acknowledge all the professors and staffs of International School of Business, University of Economics Ho Chi Minh City for providing me the best opportunities to study and develop deep understanding about research and business

Finally, I would like to express my gratitude to my dear family and friends This thesis would have never come true without their encouragements and supports

TABLE OF CONTENTS

ABSTRACT 1

CHAPTER 1: INTRODUCTION 3

1.1 Research background 3

1.1.1 The emerging luxury market 3

1.1.2 Counterfeiting 4

1.2 Existing studies on counterfeiting 5

1.3 Research objectives 6

1.4 Scope of the study 6

1.5 Research significance 7

1.6 Research structure 7

CHAPTER 2: LITERATURE REVIEW 9

2.1 Luxury brands 9

2.2 Counterfeits 9

2.3 Counterfeit purchase intention 10

2.4 Value-expressive and social-adjustive function 10

2.5 Product involvement 12

2.6 Product knowledge 13

2.7 Summary 14

CHAPTER 3: RESEARCH METHODOLOGY 17

3.1 Data collection 17

3.2 Measurement 18

3.3 Luxury brands 20

3.4 Measurement validation 20

3.5 Hypotheses tests 22

3.6 Summary 24

CHAPTER 4: DATA ANALYSIS 26

4.1 Pilot study 26

4.2 Data collection result and demographics 28

4.3 Measurement validation 28

4.3.1 Cronbach‘s alpha 28

4.3.2 Exploratory factor analysis 32

4.3.3 Confirmatory factor analysis 33

4.3.4 Composite reliability and average variance extracted 36

4.3.5 Divergent reliability and Pearson correlation 36

4.3.6 Final scales 37

4.4 Hypotheses tests 39

4.4.1 Model A 39

4.4.2 Model B 41

4.4.3 Model C 42

4.5 Discussion 43

4.6 Summary 45

CHAPTER 5: CONCLUSION 47

5.1 Overview 47

5.2 Managerial implications 47

5.3 Limitations and future research 48

References 51

Appendix A 55

Appendix B 56

Appendix C 57

Appendix D 60

Appendix E 61

Appendix F 67

Appendix G 68

LIST OF TABLES

Table 2.1: Summary of definitions 14

Table 2.2: Summary of hypotheses 15

Table 3.1: Measurement scales 19

Table 3.2: Summary of research criteria 24

Table 4.1: In-depth interview results 26

Table 4.2: Demographics 28

Table 4.3: Cronbach‘s alpha 29

Table 4.4: Item-total statistics for product knowledge scale 29

Table 4.5: Item-total statistics for counterfeit purchase intention scale 30

Table 4.6: Cronbach‘s alphas after item reduction 31

Table 4.7: KMO and Bartlett‘s test 32

Table 4.8: Promax rotation with Kaiser Normalization, k=4 33

Table 4.9: Modification indices covariances 34

Table 4.10: Composite score and average variance extracted 36

Table 4.11: Square root of average variance extracted and correlations 37

Table 4.12: Modified scales 37

Table 4.13: Summary of research criteria results 38

Table 4.14: Hierarchical multiple regression – Model A 40

Table 4.15: Hierarchical multiple regression – Model B 41

Table 4.16: Hierarchical multiple regression – Model C 42

Table 4.17: Hypothesis test results 45

LIST OF FIGURES

Figure 2.1: Conceptual model 16 Figure 3.1: Research process 17 Figure 4.1: Modified confirmatory factor analysis model 35 Figure 4.2: Moderation effect of product involvement on the relationship between value-expressive function and counterfeit purchase intention 40 Figure 4.3: Moderation effect of product knowledge on the relationship between social-adjustive function and counterfeit purchase intention 43

ABSTRACT

In recent years, luxury market is growing fast worldwide—particularly in Asia However, along with the growth is the emerging threat from counterfeits Despite many works on counterfeiting, more works are needed to investigate the impacts of psychological aspects on counterfeit purchasing intention Furthermore, the interaction effects of the psychological factors and two important constructs in consumer behavioral studies—product involvement and product knowledge—were also examined to further understand the problem

In order to achieve the research objectives, five respondents took part in in-depth interviews to check the wordings of a 25-item questionnaire The revised questionnaire was used in the main survey 248 respondents participated in the survey, and 201 usable answers were retained Tests were conducted to examine the validity and reliability of the measurement scales, which resulted in the elimination of an item The hypotheses were then tested with hierarchical multiple regression method

The results from the tests indicated that social-adjustive function significantly affected counterfeit purchase intention (p < 001) On the contrary, the effect of value-expressive function was found to be positive (β = 078) but insignificant (p = 271) Regarding to the moderation effects, the interaction between product involvement and value-expressive function negatively and significantly affected counterfeit purchase intention (p = 029), while the impacts of the interaction between product involvement and social-adjustive function turned out to be insignificant (p = 268) The results also claimed that social-adjustive function was found to be negatively moderated by product knowledge (p = 042)

It can be inferred from the findings that luxury brands focusing on communicating the social-adjustive functions are more likely to face the risk of counterfeiting than those focusing on the value-expressive functions In order to reduce customers‘ intention to buy counterfeits, two marketing combinations—value-expressive function focused combined with campaigns strengthening product involvement, and social-adjustive focused combined

with campaigns strengthening product knowledge—are advised to be taken into consideration seriously, while other combinations have not been proven to be effective yet and require further investigation

Chapter 1 INTRODUCTION

This chapter tends to introduce the situation and characteristics of luxury brands, as well as their war against counterfeits Chapter 1 also summarizes previous studies on the subject, as well as the gaps should be filled Research objectives, research questions and research scope are stated Research structure is also described

1.1 Research background

This section investigates the growing market of luxury brands and products worldwide, especially in Asian countries It also gives an overview of battle against counterfeiting of luxury brands

1.1.1 The emerging luxury market

In recent years, the market for luxury products is expanding faster and faster Though there are still disagreements in estimating luxury market size due to different methods, researchers all agree on the fast increase of the market (Heine & Phan, 2011; Ho, Moon, Kim, & Yoon, 2012; Truong, McColl, & Kitchen, 2008) According to Truong et al (2008), this fast increase of the luxury market is due to two main reasons: Firstly, the economy has been better recently with improved business environment, resulting in lower unemployment rates (which lead to higher income and consumption) and lower production costs (which understandably lead to production expansion) Secondly, the market of luxury products now does not contain only wealthy consumers, but also lower-class ones Truong et al (2008) summarize that the low-class consumers nowadays tend to purchase luxury products to imitate the high-class, to gain good feelings for the purpose of self-rewarding, or simply due

to the high-quality This results in a new kind of luxury brand—which Truong et al (2008) call the masstige—a portmanteau of mass market and prestige market, indicating that the luxury market is now open to the mass Moreover, Cavender and Kincade (2013) also state the role of lower entry barriers Thanks to globalization, the level of managing and conducting business raises consequently, which enables potential players to join the luxury-

brand game—a game used to be available to a limit number of companies who possess the best technologies and resources to produce luxury products

The emerging market of luxury products does not solely happen in Western countries, but also in the East, for example China (Langlois & Barberio, 2013), India (Gupta, 2009), Korea (Ho et al, 2012) and many other Asian countries Therefore, the need

of conducting researches on luxury products, especially in Asian countries, is never ceased

to grow

However, rapid growth has its own flip side It creates new serious challenges for the luxury brands Hennigs, Wiedmann, Behrens, and Klarmann (2013) raise the question about the balance between the brand expansion and its rarity, since as the brand grows, it becomes more common to the mass, reducing its ―luxury‖ state Furthermore, luxury brand managers

do not only have to balance their own quality and quantity, but also have to struggle with a special competitor: the counterfeits Hennigs et al (2013) point out the threat of counterfeits

to luxury brands, as the costs to produce counterfeits are getting lower and lower recently Counterfeits threaten the brands‘ revenue, profit, and especially flood the market with products, lowering brands‘ rarity and make them less ―luxury.‖

1.1.2 Counterfeiting

Counterfeiting has become a critical issue in branding, especially studies of luxury brands as mentioned in the previous section According to Hardy (2014), the value of counterfeit market will increase to 1.7 million USD in 2015, ranging from casual goods to luxury brands Despite many efforts to enforce intellectual laws as well as anti-counterfeiting activities, counterfeiting is still an international problem in various regions—China, Sub-Saharan African, Southeast Asian, and Western Balkans (Meiring, 2014; Mijatovic, Marenovic, Kliska, & Kompari, 2014; Phau & Teah, 2009; Vachanavuttivong et

al, 2014)

In luxury categories, brands have to face bigger threats from counterfeits, as they are easily produced with low costs (Phau & Teah, 2009) Staake, Thiesse and Fleisch (2009) claim that companies in many categories experience sharp decreases in revenue, which in

turn puts great pressure on the company financial situation and sets high entry barriers to the industry Furthermore, consumers of highly-counterfeited brands consider the brands less attractive (Staake et al, 2009) To make it worse, many luxury brand consumers intentionally buy counterfeits as an inexpensive alternative to real products, despite the risks they may face, as well as the efforts brands have made to distinguish their products from counterfeits (Perez, Castaño, & Quintanilla, 2010) Therefore, it is essential to investigate counterfeit consumers‘ motivations, in order to minimize the threats of luxury brand counterfeits

1.2 Existing studies on counterfeiting

According to the review of Staake et al (2009), many researches have tried to figure out the factors affecting consumers‘ intention to buy counterfeits in many aspects—especially supply-side However, little has been done to investigate demand-side reasons, as many researchers expect counterfeit purchasing activities to be the results of untruthful sellers (Penz & Stöttinger, 2005) Nonetheless, as stated above, in many cases, consumers knowingly buy counterfeits (Perez et al, 2010), hence demand-side aspects must be investigated more carefully

Among the works examining demand-side aspects of counterfeit purchasing, little has been done focusing on the psychological aspects of intentional counterfeit purchasing—particularly value-expressive and social-adjustive aspects of the activities (also called self-expressing and self-presenting)—the most important values of luxury brands (Nia & Zaichkowsky, 2000) Wilcox, Kim and Sen (2009) have conducted research on the value-expressive and social-adjustive function, however, their research does not focus on investigating the significance of the impacts of the two functions on counterfeit purchase intention

Furthermore, although works have been done on the psychological aspect of counterfeit purchasing, little examines the moderating effects of many important factors—specifically product involvement and product knowledge, two widely recognized variables

in predicting consumers‘ behaviors (Bian & Moutinho, 2011) Bian and Moutinho (2011)

even state the importance of product involvement as follow: ―product involvement is a central framework, vital to understanding consumer decision-making behavior and associated communications‖ (p.195) Hence, it is worth studying what are the effects of these two important concepts in consumer behavior study in our model of counterfeit purchase intention

1.3 Research objectives

Due to the need of researches examining psychological aspect of counterfeit purchasing intentions, this research aims to study the effects of value-expressive function and social-adjustive function on counterfeit purchase intentions Furthermore, this paper also tries to investigate the moderating effects of product involvement on the relationships among value-expressive function, social-adjustive function and luxury counterfeit purchase intention; and the moderating effect of product knowledge on the relationship between social-adjustive function and luxury counterfeit purchase intention

Hence the research questions:

1 To what extent do value-expressive function and social-adjustive function influence counterfeit purchase intention in the cases of luxury brands?

2 To what extent does product involvement moderate the relationships among value-expressive function, social-adjustive function, and luxury counterfeit purchase intention?

3 To what extent does product knowledge moderate the relationship between social-adjustive function and luxury counterfeit purchase intention?

1.4 Scope of the study

This research focuses on investigating the impact of psychological function on counterfeit purchase intention, as well as the moderation effects of product involvement and product knowledge of the consumers Please be well informed that this research studies consumers in Ho Chi Minh City and in fashion-related luxury field only due to resource limitations

1.5 Research significance

This study contributes to the understandings of luxury market and its struggle against counterfeits, especially in the Vietnamese context Moreover, the findings will help luxury brand managers to wisely choose better marketing combinations to reduce the intention to purchase counterfeits of their customers

Chapter 2: Literature review

Chapter 2 defines clearly the constructs (value-expressive function, social-adjustive function, product knowledge, product involvement and counterfeit purchase intention) as well as other definitions (counterfeits, luxury brands) and digs deeper into previous works Hypotheses and conceptual model are also developed in this chapter with detail justifications

Chapter 3: Research methodology

Chapter 3 describes the method to conduct the research, including data collection (method and measurement) and data analysis Details of the process of building the questionnaire, as well as the collection methods and tools of testing the data reliability and validity are also given in this chapter

Chapter 4: Data analysis

Chapter 4 includes the results and findings inferred from the collected data Measurement and hypothesis tests are run and presented in details, using Statistical Package for Social Science (SPSS) 16.0 and AMOS 20.0 of International Business Machines Corporation (IBM) The results of the tests are also summarized for further discussion in Chapter 5

Chapter 5: Conclusion

Chapter 5 discusses further the results and findings from Chapter 4 to draw out managerial implications applicable in real life Limitations of this paper are also clearly stated to suggest further researches in the future

Chapter 2 LITERATURE REVIEW

This chapter gives a clear view about the important constructs of the paper, including the definition of luxury brand, counterfeit, counterfeit purchase intention, value-expressive and social-adjustive function, product knowledge and product involvement Five hypotheses are also proposed after being argued based on previous works

2.1 Luxury brands

The concept of luxury has been defined in various works, however, there is no clear criteria to distinguish luxury from non-luxury brands, as ‗luxury‘ concept is subjective (Wiedmann, Hennigs, & Siebels, 2007) and strongly depends on the context (Vigneron & Johnson, 2004)—for example, a brand may appear to be luxurious to this consumer, but not

to another one However, there are common characteristics that different works agree on, including high level of rarity or uniqueness, high-quality, high-price and high level of self-expressing as well as social meaning (Miller & Mills, 2012; Vigneron & Johnson, 2004; Wiedmann et al, 2007) Furthermore, Nia and Zaichkowsky (2000) state that the main difference between luxury and non-luxury brands is the role of psychological in comparison

to functional value In luxury brands, the psychological value is relatively high, and is the most important factor affecting purchasing decision

Therefore, in this research, luxury brands are defined as brands bought mainly to serve consumers‘ psychological needs—specifically self-expressing and self-presenting—that are rare or unique, high-quality and highly priced

2.2 Counterfeits

Counterfeits can be defined as products that bear a fake or indistinguishable trademark, logo, name or design of another product, in order to illegally take advantage of the brand value of the real product (Bian & Moutinho, 2009; Phau & Teah, 2009; Staake et

al, 2009; Wilcox et al, 2009) Wilcox et al (2009) also state that the quality of counterfeits is

usually low, however, it has been rising gradually and in some cases, has reached the same level of that of the real products

Counterfeits are categorized into deceptive and non-deceptive (Bian & Moutinho, 2009; Wilcox et al, 2009) Deceptive counterfeits are products that are designed to intentionally confuse the buyers, and buyers buy them without knowing that they are counterfeits On the contrary, non-deceptive counterfeits are products that can be distinguished from the real products easily by the buyers—based on various types of differences, for example in product appearances, in trademarks, or in distribution channels—and consumers purchase them intentionally This research focuses on the second type

2.3 Counterfeit purchase intention

A behavioral intention is the willingness and the effort of an individual to perform that behavior (Ajzen, 1991) Based on Ajzen‘s suggestions, counterfeit purchase intention can be defined as the willingness and the effort of an individual to purchase counterfeits—in this case counterfeits of luxury brands Ajzen (1991) also states behavioral intention is good indicator of actual action, hence, counterfeit purchase intention is a good proxy of counterfeit purchase action in the model

Furthermore, The Theory of Planned Behavior (Ajzen, 1991) claims that people cannot have the intention to perform any action without being able to control that action Therefore, counterfeit purchase intention cannot exist in the case of deceptive counterfeits That is the reason why this research focuses on the intention to purchase non-deceptive counterfeits only

2.4 Value-expressive and social-adjustive function

According to Grewal, Mehta, & Karrdes (2004), Katz (1960), Wilcox et al (2009), in

a consuming context, a person holds attitudes toward something in order to serve one of several psychological functions that they may benefit from, including knowledge function (easing the decision making process), utilitarian function (maximizing received value from products), value-expressive function (communicating or expressing one‘s values), and

social-adjustive function (presenting oneself and interacting with other people) On the other hand, various works applying the theory of planned behavior have proved that attitudes toward counterfeit purchasing have significant impacts on counterfeit purchase intention (Chiu, Lee, & Won, 2014; Koklic, 2011; Phau, Teah, & Lee, 2009) Therefore, the functions

of the products should affect consumers‘ counterfeit purchase intention

This research aims to investigate the social-side functions of the luxury products only—which mean value-expressive function and social-adjustive function Wilcox et al (2009) suggest that people who are driven by value-expressive function tend to buy because they want to express their values, beliefs or personalities On the other hand, social-adjustive-function-driven consumers purchase products because they help them to fit in social situations In order words, value-expressive function satisfies the need of self-expressing—i.e dealing with self-judgments, whereas social-adjustive function serves the need of self-presenting—i.e dealing with social judgments Hence, in this study, value-expressive function is defined as function of a product that helps consumers to communicate

or express their values Social-adjustive function is defined as function of a product that helps consumers to fit in social situations, present themselves and interact with other people

Many works suggest that value-expressive function affect consumer‘s decisions to purchase counterfeits Wiedmann, Hennigs and Klarmann (2012) propose that this influence

is negative Due to the nature of the function, consumers who are driven by expressive function cannot enjoy the same level of satisfaction buying and using counterfeits, since they themselves are able to fully perceive the differences Respondents of Perez et al (2010) also find themselves uncomfortable purchasing counterfeits One of them, whose name is Cristina, says: ―You don‘t feel the same way that you do with the original,

value-you‘re always going to know that‖ (Perez et al, 2010, p 226)

Similarly, social adjustive function is proposed to have impacts on counterfeit purchase intention (Wiedmann et al, 2012) Perez et al (2010) find out that counterfeits—especially super-non-deceptive counterfeits, i.e non-deceptive counterfeits that cannot be easily distinguished from real products by other people—provide the same social effects

Consumers tend to purchase counterfeits, especially those of trendy brands—brands containing high level of social-adjustive function (Perez et al, 2010) Hence, our hypotheses are stated as follows:

H1: Value-expressive function has a negative effect on counterfeit purchase intention

H2: Social-adjustive function has a positive effect on counterfeit purchase intention

2.5 Product involvement

Product involvement can be defined as the perceived level of importance of a particular product category in consumer‘s life, because it meets the needs or values of the consumer (Bian & Moutinho, 2011; Quester & Ai, 2003; Te'eni-Harari & Homik, 2010) In other words, consumer‘s response to products that meet their needs or values creates high product involvement

The level of product involvement reflects two aspects: (1) the importance of the product in consumer‘s life, and (2) the interest of the consumer in the product, or the amount of pleasure the consumer gains from the product (McQuarrie & Munson, 1992; Vaughn, 1986; Zaichkowsky, 1987) Therefore, the higher the product involvement, the more benefit the consumer gains from choosing the right product This makes consumers tend to put more effort in choosing high-involvement products (Bian & Moutinho, 2011)

As a result, in the case of high-involvement products, the risk of choosing wrong product devalues the benefits that counterfeits may bring to its consumers Furthermore, the importance of high-involvement products may strengthen the effect of self-deceiving feeling

on the counterfeit purchase intention Snyder and DeBono (1985, as cited in Wilcox et al, 2009) also found that value-expressive-function-driven consumers care more about quality-related problems—one of the main disadvantages of counterfeits (Wilcox et al, 2009) The higher the product involvement, the more important the product is to the consumer, and the quality-related problems would appear to be of bigger concern than those in the case of low-involvement products

Besides, as proposed by Bian and Moutinho (2011), when the product involvement is high, consumers think more carefully before making the purchase, which means they have

to bring into consideration the social-adjustive benefit that real products may bring in comparison to counterfeits Moreover, the risk of being discovered consuming counterfeits has more impacts in the case of high-involvement products, hence making the social benefits of counterfeits less appealing

Hence, it can be proposed that:

H3: High level of product involvement strengthens the relationship between expressive function and counterfeit purchase intention

value-H4: High level of product involvement weakens the relationship between adjustive function and counterfeit purchase intention

social-2.6 Product knowledge

Marks and Olson (1981) define product knowledge as information relating to the product stored in the memory The information is collected directly or indirectly Marks and Olson (1981) also suggest that consumers who possess high level of product knowledge maintain more complex decision making procedure while purchasing, as they have more infomation to consider

As consumers with high level of product knowledge understand more about the product, they are better aware that the counterfeits are low-quality products (Bian & Moutinho, 2011) Because of that, they are able to figure out the differences between counterfeits and real products Due to the fact that knowledgeable consumers can easily distinguish counterfeits, they perceive the risk of being discovered using counterfeits more seriously When the risk of being unmasked is high, the self-presenting benefit effect is weakened, which leads consumers towards real products (Perez et al, 2010)

H5: High level of product knowledge weakens the relationship between adjustive function and counterfeit purchase intention

social-2.7 Summary

This chapter deals with the definition of main concepts—luxury, counterfeit, counterfeit purchase intention, value-expressive function, social-adjustive function, product involvement and product knowledge The definitions employed in this paper are summarized in Table 2.1

Table 2.1: Summary of definitions

Luxury brands Brands bought mainly to serve consumers‘

psychological needs—specifically self-expressing and self-presenting—that are rare or unique, high- quality and highly priced

Miller & Mills (2012);

Nia & Zaichkowsky (2000);

Vigneron & Johnson (2004);

Wiedmann et al (2007)

Counterfeits Products that bear a fake or indistinguishable

trademark, logo, name or design of another product, in order to illegally take advantage of the brand value of the real product

Bian & Moutinho (2009);

Phau & Teah (2009);

involvement product category in consumer‘s life, because it

meets the needs or values of the consumer

Marks & Olson (1981)

The relationships among concepts are also investigated to form the following hypotheses as in Table 2.2

Table 2.2: Summary of hypotheses

H1 Value-expressive function has a negative effect on

counterfeit purchase intention

Perez et al (2010); Wiedmann et al (2012)

H2 Social-adjustive function has a positive effect on

counterfeit purchase intention

Perez et al (2010); Wiedmann et al (2012)

H3

High level of product involvement strengthens the relationship between value-expressive function and counterfeit purchase intention

Bian & Moutinho (2011); Perez et al (2010); Snyder and DeBono (1985, as cited in Wilcox

Bian & Moutinho (2011); Perez et al (2010); Wiedmann et al (2012)

H5

High level of product knowledge weakens the relationship between social-adjustive function and counterfeit purchase intention

Bian & Moutinho (2011); Perez et al (2010)

The conceptual model of this research is presented below:

Figure 2.1: Conceptual model

H2 (+)

H3 (+) H4 (-)

H5 (-)

Chapter 3 RESEARCH METHODOLOGY

This chapter describes the method of conducting the research in each stage, including the data collection method (item generation to build the questionnaire, clarification of luxury brands chosen in this research, and data collection) and the data analysis method (measurement validation as well as hypothesis tests) The complete research process is also given

Revision

Final questionnaire

Survey (n=201)

Cronbach‘s alpha test

Exploratory factor analysis

Confirmatory factor analysis

Hierarchical multiple regression

The pilot study included five in-depth interviews The questionnaire was translated into Vietnamese Five customers were chosen to participate in in-depth interviews in order

to revise the Vietnamese questionnaire to avoid any difficulties in understanding the questions

After the pilot study, the main study with a larger scale survey was conducted Respondents—chosen by using convenient sampling method—were asked to answer the revised questionnaire in Vietnamese Every fifth customer that walked through the main entrance of Diamond Plaza (District 1, Ho Chi Minh City, Vietnam) was asked to answer the self-administered questionnaire The data collected in the main study was analyzed to test the reliability as well as the validity of the model

3.2 Measurement

The questionnaire contained 25 items measuring five constructs: value-expressive function, social-adjustive function, counterfeit purchase intention, product involvement and product knowledge

Both value-expressive function and social-adjustive function were measured by four items each The items—adopted from Wilcox et al (2009)—were seven-point Likert-type questions, ranging from 1 – ―completely disagree‖ to 7 – ―completely agree.‖

Counterfeit purchase intention construct was measured by the three items adopted from (Hung et al, 2011) The items were seven-point Likert-type questions, ranging from 1 – ―completely disagree‖ to 7 – ―completely agree.‖

Product involvement construct was measured using the Revised Personal Involvement Inventory of McQuarrie and Munson (1992), which consisted of 10 items, as this was a widely used scale measuring product involvement All the items were 7-point bipolar evaluative adjective pairs

Product knowledge construct was measured using the four items developed by Smith and Park (1992) The four items were seven-point Likert-type questions, ranging from 1 –

―completely disagree‖ to 7 – ―completely agree.‖

The final measurement scales are presented as in Table 3.1

Table 3.1: Measurement scales

Value-expressive function

Wilcox et al (2009)

VE3 Luxury brand helps me express myself

VE4 Luxury brand helps me define myself

Social-adjustive function

SA1 Luxury brand is a symbol of social status

Wilcox et al (2009) SA2 Luxury brand helps me fit into important social situations

SA3 I like to be seen wearing/using luxury brand

SA4 I enjoy it when people know I am wearing/using a luxury brand

Product knowledge

PK1 I feel very knowledgeable about fashion products

Smith and Park (1992)

PK2 If a friend asked me about fashion products, I could give them advice

about different brands

PK3 If I had to purchase fashion products today, I would need to gather very

little information in order to make a wise decision

PK4 I feel very confident about my ability to tell the difference in quality

among different brands of fashion products

Product involvement

I consider fashion products to be…

PI1 Unimportant/Important

McQuarrie and Munson (1992)

PI2 Irrelevant/Relevant

PI3 Means nothing to me/Means a lot to me

PI4 Unexciting/Exciting

PI5 Dull/Neat

PI6 Doesn‘t matter/Matter to me

PI7 Boring/Interesting

PI8 Not fun/Fun

PI9 Unappealing/Appealing

PI10 Of no concern/Of concern to me

Counterfeit purchase intention

CP1 I have strong possibility to purchase counterfeits of fashion-related luxury

product

Hung et al (2011) CP2 I‘m likely to purchase counterfeits of fashion-related luxury product

CP3 I have high intention to purchase counterfeits of fashion-related luxury

Due to the nature of the term ―luxury,‖ the luxuriousness of brands differs depending

on each respondent (Wiedmann et al, 2007) Therefore, in this research, respondents were given several widely-recognized luxury brands—specifically Rolex, Gucci, Louis Vuitton—

as references to understand more about the concept of luxury As fashion-related products are the best target of counterfeiters (Eisend & Schuchert-Güler, 2006), respondents were asked to consider fashion-related luxury products that they have bought, such as clothes, purse, watch, shoes before answering the questions

3.4 Measurement validation

The convergent validity was measured through three methods: construct reliability, item reliability, and average variance extracted Construct reliability was tested using Cronbach‘s alpha and composite reliability score According to Nunnally and Bernstein

(1994), the Cronbach‘s alpha of each group must exceed 0.70 to pass the test Molina, Montes, and Ruiz-Moreno (2007) proposed that the composite reliability score should be above 0.70 Based on Chin (1998), the loadings of each item should be no lower than 0.60

to pass the item reliability test Furthermore, as suggested by Fornell and Larcker (1981), the constructs‘ average variance extracted should be 0.50 or higher

The discriminant validity was measured by examining cross-loadings and by using square root of the average variance extracted As proposed by Chin (1998) and Fornell and Larcker (1981), the model will have discriminant validity if the square root of the average variance extracted of each construct exceeds every correlation between that construct and any other construct

Exploratory factor analysis was run during the process to collect the information about item loadings and cross-loadings As proposed by Pallant (2007), the sample size should be 150 or above, and the case-per-item ratio should be 5:1 or above (Comfrey & Lee, 1992; Hair, Anderson, Tatham, & Black, 1998) Since there were 25 items in the questionnaire, the minimum number of responses to run exploratory factor analysis was 150 (5 x 25 = 125, lower than the 150 criterion) Furthermore, Pallant (2007) also suggests that the Kaiser-Meyer-Olkin value should not be lower than 0.6, and the Bartlett‘s test of sphericity should be significant (p < 0.05)

Confirmatory factor analysis was also run to test the model fit The cut-off criteria applied in this paper were χ2/df ≤ 3:1 (Kline, 2005); CFI ≥ 0.95 (Hu & Bentler, 1999); RMSEA ≤ 0.06 (Hu & Bentler); and SRMR ≤ 0.08 (Hu & Bentler, 1999)

To figure out any multicollinearity issues that might occurred, Pearson correlation analysis was conducted According to Field (2005), all correlation should be below 0.8 to avoid multicollinearity

After measurement validation, hypotheses tests were conducted using the modified scales

3.5 Hypotheses tests

In order to test the moderation effects, product involvement and product knowledge were dichotomized using median-split method Therefore, product involvement included two levels, namely low product involvement and high product involvement Similarly, product knowledge included two levels, namely low product knowledge and high product knowledge

Particularly, to test the hypotheses proposed in Chapter 2, the following hierarchical multiple regression models were used:

Firstly, Model A was used to test the effect of value-expressive function on counterfeit purchase intention (H1), as well as the moderation effect of product involvement

on the relationship between value-expressive function and counterfeit purchase intention (H3) using hierarchical multiple regression The first step included only value-expressive function; the second step included product involvement; and in the third step, the product of value-expressive function and product involvement was added Below is the equation of Model A:

y = β0 + β1x1 + β2x3 + β3x1x3 + ε1 where:

y: counterfeit purchase intention;

second step included product involvement; and in the third step, the product of adjustive function and product involvement was added Below is the equation of Model B:

social-y = β0‘ + β1‘x2 + β2‘x3 + β3‘x2x3 + ε2 where:

y: counterfeit purchase intention;

y = β0‖ + β1‖x2 + β2‖x4 + β3‖x2x4 + ε3 where:

y: counterfeit purchase intention;

n: minimum sample size;

m: number of predictor variables

Since the number of predictor variables was four, the minimum sample size should

be 82

3.6 Summary

The data collection process consisted of two main phases: a pilot study to check if there was any problem in understanding the questionnaire, and a main survey to collect the data Due to the nature of the term ―luxury,‖ participants were asked to think about a fashion-related product that they considered luxury when answering the questionnaire

The data would go through measurement validation The criteria in measurement validation were specified in details in Table 3.2

Table 3.2: Summary of research criteria

Cronbach‘s alpha of

each group

Nunnally and Bernstein

Square-root of the average variance extracted

of each construct > correlation between that construct and any other constructs

Minimum sample size

Comfrey and Lee (1992); Hair et al (1998)

case-per-item ratio ≥ 5:1 (or ≥ 125 in this

paper)

Tabachnick and Fidell

(2001) n > 50 + 8m (or ≥ 82 in this paper)

Kaiser-Meyer-Olkin

Bartlett‘s test of

χ 2

After measurement validation tests, final data were used to test the hypotheses, Hierarchical multiple regression was employed in this stage

Chapter 4 DATA ANALYSIS

The data collected was analyzed to test the measurement model—including convergent validity and discriminant validity—as well as the hypotheses Statistical Package for Social Science (SPSS) 16.0 and AMOS 20.0 of International Business Machines Corporation (IBM) was employed in this paper The results of the data collection and data analysis procedure are presented below

4.1 Pilot study

The questionnaire was translated into Vietnamese and used to ask five respondents Respondents were read each item and asked to express how they understood the item The results are shown below

Table 4.1: In-depth interview results

this item as ―need to

gather very little

information‖ due to

this item as ―need to gather very little information‖ due to

No

―low interest in

fashion‖

―low interest in fashion‖

√ Fully and correctly understand the item

Most of the items were understood correctly, except PK3 and PI2 Respondent A and Respondent E mistakenly interpreted PK3 as ―need to gather very little information‖ due to

―low interest in fashion.‖ This mistake appeared because they overlooked the word ―wise‖ (the item turned to be ―If I had to purchase fashion products today, I would need to gather very little information in order to make decision.‖) Since the misunderstood was due to the respondents‘ recklessness, the item was not modified

Item PI2 was understood thoroughly by all the respondents However, Respondent C suggested that the question should be in more details—it should state clearly that irrelevant/relevant to something or anything As all the respondents understood this

question, the specification was not considered to be necessary, and the item was not changed

In summary, the items were not changed after the pilot study, and were used in the main survey

4.2 Data collection result and demographics

Finally 248 people agreed to take part in the research, and 201 usable responses were collected The other 47 responses were eliminated because they were not completed or not seriously answered

Below were some important demographics, which are gender and pass behavior (have bought luxury products or not) As can be seen, gender was equally distributed, while pass behavior fell more on the have-bought group due to the characteristics of the location that the data collection process was conducted:

Table 4.2: Demographics

Have not bought luxury products

Have bought luxury products

4.3.1 Cronbach’s alpha

As stated in Chapter 3, in measurement validation process, Cronbach‘s alpha test was run to eliminate any bad items to increase construct reliability This was the result of the initial Cronbach‘s alpha test:

Table 4.3: Cronbach‘s alpha

PI1, PI2, PI3, PI4, PI5, PI6,

Counterfeit

All Cronbach‘s alpha scores were above 7 However, the item-total statistics tables—particularly those of product knowledge scale and of counterfeit purchase intention scale—suggested deleting some items to improve the scores Specifically, in the scale measuring product knowledge, item PK3 was suggested to be eliminated to improve the Cronbach‘s alpha score from 817 to 852 Cronbach‘s alpha score would increase 035, which means the elimination of item PK3 would be a reasonable improvement The remaining number of items was three, which was not too low to cause scale reliability issues

as stated by Hinkin (1995) As a result, item PK3 was deleted from product knowledge scale

Table 4.4: Item-total statistics for product knowledge scale

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Corrected Total Correlation

In the construct counterfeit purchase intention, item CP3 was suggested to be eliminated to improve the Cronbach‘s alpha score from 896 to 952 However, as stated by Hinkin (1995), constructs with low number of items were reported to be unreliable Furthermore, Cronbach‘s alpha of the three-item scale was 896, nearly reached 9, which was considered excellent (George & Mallery, 2003) Moreover, the item-total correlation of CP3 to the counterfeit purchase intention scale was high—.674 Therefore, CP3 was not deleted from the scale

Table 4.5: Item-total statistics for counterfeit purchase intention scale

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Corrected Total Correlation

Table 4.6: Cronbach‘s alphas after item reduction

Correlation

Cronbach's Alpha if Item

Deleted

Cronbach’s alpha

4.3.2 Exploratory factor analysis

As stated in Chapter 3, the minimum number of responses to run exploratory factor analysis should be 150 or above The number of qualified responses was 201, hence passed the minimum threshold

Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett‘s test of sphericity was run The Kaiser-Meyer-Olkin measure of sampling adequacy score was 868, higher than the 6 criterion, and Barlett‘s Test of sphericity was strongly significant (p =.000), which means that exploratory factor analysis was valid

Table 4.7: KMO and Bartlett‘s test

Kaiser-Meyer-Olkin Measure of Sampling

Bartlett's Test of Sphericity

Approx Chi-Square 3.270E3

According to Matsunaga (2010), Promax rotation is recommended over Varimax rotation in social research Therefore, Promax rotation with Kaiser Normalization, k=4 was employed Since all the loadings were above 6, no items were eliminated

The loadings are given in Table 4.8

As can be seen in the loadings table, the differences between the highest loading and the second-highest loading of all the items were higher than 2 Therefore, it was safe to infer that there was no cross-loaded item