CONSUMER ETHNOCENTRISM AND COUNTRY-OF-ORIGIN EFFECTS:
PITFALLS AND ACCELERATORS TOWARDS BUILDING BRAND EQUITY
IN AN INTERNATIONAL SERVICES INDUSTRY
AN EMPIRICAL ASSESSM ENT OF THE GERMAN AIR TRAVEL MA RKET
by Julius Passon
Rijksuniversiteit Groningen Newcastle University Business School
CONSUMER ETHNOCENTRISM AND COUNTRY-OF-ORIGIN EFFECTS: PITFALLS AND ACCELERATORS TOWARDS BUILDING BRAND EQUITY IN AN
INTERNATIONAL SERVICES INDUSTRY
AN EMPIRICAL ASSESSMENT OF THE GERMAN AIR TRAVEL MARKET
by Julius Passon
Dual Award M. Sc. Advanced International Business Management and Marketing Rijksuniversiteit Groningen
Faculty of Economics and Business
P.O. Box 800, 9700 AV Groningen, The Netherlands Newcastle University
Business School
5 Barrack Road, Newcastle upon Tyne, NE1 4SE, England
Julius Passon
Student Number: s2197820 b110482006
Am Hengstbach 5, 63263 Neu-Isenburg, Germany Tel.: +49 (0) 176 / 615 245 72
E-mail: julius.passon@gmx.de
Supervisors: Drs. H.A. Ritsema Prof. Markus Blut
and checks for potential country-of-origin effects within a specific service industry – the German air travel market. Insights generated through the analysis are utilized to identify potential accelerators and pitfalls towards building consumer-based brand equity in an international services industry, with particular focus on foreign service firms operating in Germany. Conceptualising constructs from a consumer perspective gives us insightful information on how consumers actually evaluate service offers. Predicting consumer behaviour enables marketers to provide consumers with the necessary means to make coherent choices and, in turn, generate sales.
Results show that consumer ethnocentrism and the country-of-origin effect are present, but that the levels are not excessive. The level of consumer ethnocentrism (Mean=41.49), in and of itself, not significant enough to symbolize a pitfall for foreign firms and accelerator for domestic firms. The level of consumers ethnocentric tendencies varies with demographic aspects and its impact is increased in combination with a country-of-origin effect. The country-of-origin effect in this service setting is present, with a relative significance score of 16.60 per cent, but it ranks lower than the factors of price and number of stops regarding importance of attributes consumers use to evaluate a flight. Evidence verifies that the strength and direction of a country-of-origin effect depends on the country s level of economic development, cultural similarity to the host country, as well as its image in the eyes of the consumer.
A journey of a thousand miles starts with a single step – Lao Tzu
With this in mind I began to tackle the most demanding challenge of my academic career. From the very beginning in June 2013, I realised that this was not going to be a linear process from start to finish, but rather a conjunction of notions and approaches that might also lead to dead ends and pressure to refocus and possibly shift objectives and measures. What sustained me was a combination of passion for the topic at hand, patience, perseverance, pressure and support from supervisors, discipline in times when ) didn t want to spend another day in front of the notebook, as well as friends and family who distracted me when necessary. All of these things helped me construct an academic research project from scratch and follow it through until the end.
Driven by the passion for marketing and management issues in an international context, I felt very comfortable choosing a topic which was similarly important for most product and service categories and furthermore seemed to be present in any market. Over the past months, the vague starting idea of doing something on the COO effect has eventually evolved into a complete master thesis that I hope you will enjoy reading. Along the way some people have greatly contributed to this thesis, and I would like to take this chance to show my appreciation. Hence, I would like to thank both of my supervisors Drs. H.A. Ritsema from the RUG and Prof. M. Blut from the NUBS. Both guided me from the start and provided me with valuable advice and constructive feedback whenever I encountered serious obstacles. Moreover, I would like to thank my fellow students with whom I have spent the past year and who supported me whenever necessary. Cheers!
Also I would like to give special thanks to my family. With relentless perseverance and understanding they have been constantly encouraging me, not only through the process of my thesis but throughout my entire academic career. Hence, I dedicate this thesis to my parents, in acknowledgement of their support and as a sign of my deep gratitude. Neu-Isenburg, 6 December 2013
TABLE 3.1 COUNTRIES OPERATIONALISED IN THE PREFERENCE RANKING ... 22
TABLE 3.2 SERVICE ATTRIBUTES, LEVELS AND DESCRIPTIONS ... 25
TABLE 3.3 TARGET POPULATION FOR THE ANALYSIS ... 29
TABLE 4.1 CETSCALE MEAN VALUES FROM PREVIOUS STUDIES ... 34
TABLE 4.2 CETSCALE SCORES BY DEMOGRAPHIC VARIABLES ... 36
TABLE 4.3 UTILITY VALUES BY SERVICE ATTRIBUTE LEVELS ... 38
LIST OF FIGURES FIGURE 3.1 THE FOUR PHASES OF THE CONJOINT ANALYSIS ... 24
FIGURE 3.2 STIMULI USED IN THE CONJOINT ANALYSIS ... 26
FIGURE 4.1 RELATIVE IMPORTANCE OF SERVICE ATTRIBUTES ... 38
LIST OF ABBREVIATIONS
CD Cultural Distance
CE Consumer Ethnocentrism
CETSCALE Consumers Ethnocentric Tendencies Scale
COO Country-of-Origin
COSO Country of Service Origin
LDC Less Developed Country
MDC More Developed Country
OLS Ordinary Least Squares
RBV Resource-based View
ACKNOWLEDGEMENTS ... IV LIST OF TABLES ... V LIST OF FIGURES ... V LIST OF ABBREVIATIONS ... V CHAPTER 1 INTRODUCTION ... 1 1.1BACKGROUND PROBLEM ... 1 1.2PROBLEM STATEMENT ... 3 1.3RESEARCH QUESTIONS ... 3
1.4THEORETICAL AND MANAGERIAL RELEVANCE ... 4
1.5STRUCTURE AND SCOPE ... 5
CHAPTER 2 THEORETICAL FRAMEWORK ... 6
2.1COUNTRY-OF-ORIGIN EFFECT ... 6
2.1.1 The Country-of-Origin Effect in Service Settings ... 8
2.1.2 A Distinction between More and Less Developed Countries ... 10
2.1.3 The Impact of Cultural Similarity on Country-of-Origin Effects ... 10
2.1.4 Deficiencies of the Country-of-Origin Effect ... 11
2.2CONSUMER ETHNOCENTRISM ... 12
2.2.1 Consumer Ethnocentrism on the Individual Level ... 13
2.2.2 Measuring Consumers’ Ethnocentric Tendencies ... 14
2.2.3 The Impact of Demographic Variables on Consumer Ethnocentrism ... 15
2.2.4 Origin, Extension, and Practical Relevance ... 16
2.3CONSUMER-BASED BRAND EQUITY... 17
2.4COO AND CE:INTERDEPENDENCIES,RELATIONS, AND IMPACT ON CBBE ... 19
CHAPTER 3 RESEARCH METHODOLOGY ... 21
3.1RESEARCH DESIGN ... 21
3.2PREFERENCE RANKING ... 21
3.3CONSUMERS ETHNOCENTRIC TENDENCIES SCALE ... 22
3.4CONJOINT ANALYSIS ... 23
3.4.1 Attributes, Levels, and Description ... 24
3.4.2 Research Design ... 25
3.4.3 Evaluation of the Stimuli ... 26
3.4.4 Estimation of the Part-worth Functions ... 27
3.4.5 Limitations of the Conjoint Analysis ... 28
3.5SAMPLING ... 28
3.6SURVEY STRUCTURE AND PROCEDURE ... 29
CHAPTER 4 RESULTS ... 31
4.1DATA SET ... 31
4.2.3 Education ... 32
4.2.4 Nationality ... 32
4.2.5 Country of Domicile ... 32
4.3PREFERENCE RANKING ... 32
4.4CONSUMERS ETHNOCENTRIC TENDENCIES SCALE ... 33
4.4.1 Reliability and Internal Consistency ... 33
4.4.2 Findings ... 33
4.4.3 Demographics ... 35
4.5CONJOINT ANALYSIS ... 36
4.5.1 Reliability and Validity ... 37
4.5.2 Findings ... 37
4.6HYPOTHESES TEST ... 39
4.6.1 The Impact of COO on Consumer Service Evaluation ... 39
4.6.2 The Impact of Economic Development on Consumer Preferences ... 40
4.6.3 The Impact of Cultural Similarity on Consumer Preferences ... 41
4.6.4 The Impact of Consumer Ethnocentrism on Consumer Preferences ... 42
4.6.5 The Relationship between CE and Demographic Variables... 42
CHAPTER 5 DISCUSSION ... 45
5.1SUMMARY AND INTERPRETATION OF FINDINGS ... 45
5.2COMBINED EFFECTS OF CE AND COO ... 46
5.3POINTS OF INTERSECTION WITH CBBE ... 47
CHAPTER 6 CONCLUSION ... 49 6.1CONTRIBUTION TO LITERATURE ... 50 6.2MANAGERIAL IMPLICATIONS ... 50 6.3LIMITATIONS ... 51 REFERENCES ... 53 APPENDICES ... 65 1 PREFERENCE RANKING ... 65 2 SURVEY GERMAN ... 66 3 SURVEY ENGLISH ... 74
4 CETSCALE 17-ITEM ORIGINAL ... 81
5.9CETSCALE BY GENDER ... 93
5.10CETSCALE BY EDUCATION ... 93
5.11CETSCALE BY NATIONALITY ... 94
5.12CETSCALE BY COUNTRY OF DOMICILE ... 95
5.13CETSCALE BY OCCUPATION ... 95
5.14CETSCALE BY INCOME ... 96
5.15CONJOINT ANALYSIS:ORTHOGONAL DESIGN ... 97
5.16CONJOINT ANALYSIS:PLAN- AND SIMULATION CARDS ... 97
5.17CONJOINT ANALYSIS:RELIABILITY ... 98
CHAPTER1 INTRODUCTION 1.1 BACKGROUND PROBLEM
One of the oldest concerns of international marketers is whether the foreignness of a product will affect how it is perceived by consumers and whether this foreignness affects the consumers purchase intention Schooler, . (ence, when entering or operating in foreign markets, firms face uncertainties that are extremely difficult to anticipate and overcome (Craig, Greene and Douglas, 2005). Major concepts within the field of international marketing that address these matters are the country-of-origin effect and consumer ethnocentrism (CE). Liabilities resulting from the foreignness of the firm potentially affect the consumer s perception of a product or service. This in turn may influence a firm s performance, either positively or negatively. Marketers are aware of this and, thus may utilize a firm s country of origin COO . (owever, the significance and direction of the effect vary with a firm s particular COO, product/service category, and host market.
Through an application of these constructs, information about the presence and direction of a potential COO effect, as well as the general societal tendency towards purchasing foreign products and services, can be identified. This knowledge is vital for a firm as it can exploit COO information that has a positive influence or avoid damage by disguising the true origin, if it would have negative impact. The diverse effects a firm might potentially experience in a foreign market similarly display a threat and an opportunity. Hence, knowledge about the presence and direction of these concepts is vital, as they display opportunities that may easily be exploited in order to maximize a brand s performance. (owever, the COO construct in particular, with its alleged impact on product evaluation and purchase intention, has been questioned in academic research. Arguing on basis of their UK study, Balabanis and Diamantopoulos (2008, p. , accentuated that the true importance of COO information could be significantly
overestimated in extant COO research. Other scholars have made similar assertions,
including Samiee (1994, p. 586), who stated that consumers are typically uninformed
about the COO of products . Examinations of the extent to which consumers are able to
level of correct COO brand identification was found by Balabanis and Diamantopoulos (2008), who placed it at only 22 per cent.
However, while the importance of COO information has come under scrutiny in the academic arena, the use of COO image marketing is still a common marketing practice (Magnusson, Westjohn and Zdravkovic, 2011). One pertinent and current example is the Lufthansa TV commercial that has been airing in Germany and other European countries since May 2013. The commercial employs some of the most common stereotypical character traits usually associated with Germans, viewed from the perspective of a French customer visiting Germany for a business trip. Throughout his business trip he evaluates the services provided (hotel, taxi, airport check-in, etc.) with a sarcastic undertone, but in the end he cannot conclude anything other than just perfect . Another example is Singapore Airlines, which uses Singapore Girl to build a brand and country image that is warm, tender, and friendly (Chattalas, Kramer, and Takada, 2008). These examples indicate that COO after all may have a significant influence on consumers perceptions of quality and purchase decision.
Hence, within this thesis, it is examined whether a COO effect is actually present and, if so, whether it is favourable or unfavourable in the particular setting. Furthermore, the CE construct helps indicate a general societal tendency towards purchasing foreign products and services. The combination of these concepts allows for a comprehensive coverage of the most important aspects and potential pitfalls firms may face when entering a foreign market. Particular effects resulting from a firm s foreignness affect the equity of a brand, which in turn affects the brand s performance. Brand equity, considered through a consumer s perspective, is operationalized here as consumer-based brand equity CBBE and is, according to Chabowski, Samiee and (ult 2013), together with market share and purchase likelihood one of the three main determinants of brand performance. Insights generated within the analysis will then be exploited to identify strategic implications for the firm to maximize its brand s consumer based equity.
economic importance of services, most international marketing literature focuses on the COO effect with regard of products and largely ignores the effects on services (Ahmed, Johnson, Ling, Fang and Hui, 2002). A rather small number of authors have considered and empirically examined COO effects in service settings (Ahmed et al., 2002; Berentzen, Backhaus, Michaelis, Blut and Ahlert, 2008; Bruning, 1997; Michaelis, Woisetschläger, Backhaus and Ahlert, 2008; Thelen, Honeycutt, and Murphy, 2010). This study contributes to the understanding of COO effects in services by examining the German air travel industry, with an analysis of the consumers perceptions of a flight provided by a hypothetical airline with a particular origin.
1.2 PROBLEM STATEMENT
Even though a vast amount of literature on the many aspects of the COO effect, CE, and brand equity exists, there has hardly been any research on the combinational effects these measures have on the consumer and on the equity of a service brand. Hence, this study applies a consumer perspective in an attempt to broaden the examination of the effects a service firm with a particular origin must be aware of when operating in a foreign market and attempting to build brand equity. Thus, the problem statement can be formulated as follows:
To what extend does the foreign origin of a firm affect consumers’ evaluation of the service offered? How can these insights be exploited to increase the equity of a
service brand? 1.3 RESEARCH QUESTIONS
The aforementioned problem statement can be broken down into particular components and captured in more specific research questions:
(1) In a world of globalised products, services, and brands, do concepts like CE and the COO
effect actually affect consumers’ perception of a service?
(2) Does the COO of a service firm affect the consumers’ perception of quality of the
service?
(3) Is the level of consumer ethnocentrism in the German market significant enough to
(4) How much of an actual liability’ is the foreignness of a service firm in the German
market nowadays?
(5) How can the insights generated in the analysis be exploited to increase the equity of a
(foreign) service brand operating in Germany?
1.4 THEORETICAL AND MANAGERIAL RELEVANCE
When evaluating the quality of a service in advance of a purchase, consumers rely on cues to support them in their decision-making process. Though the provided services are completely similar, foreign firms may evoke different reactions in consumers minds. (ence, firms that recognise the consumers attitude and their appeal towards them have an information advantage that may, through incorporation in the marketing strategy, potentially be transferred into a competitive advantage.
Hence, this thesis contributes to the existing research on consumer behaviour literature by combining two of the most essential constructs used to generate insights about a firms acceptance in a foreign market and the consumers attitudes towards purchasing foreign products and services. The insights generated allow for a better prediction of consumer behaviour and, in turn, positively influence firm performance. Hence, the research gap identified is threefold: First the demonstration of the presence or absence of a COO effect in the chosen setting, second the determination of the level of consumer ethnocentrism within the German market, and third the acknowledgement of these effects and their consideration within the service provider s marketing strategy to increase brand equity and thus brand performance. The combination with CE allows for generating synergies that can be exploited to increase brand performance by maximizing brand equity. Hence, identification of pitfalls and accelerators toward building brand equity is of major importance.
the objective of increasing brand performance, the COO might be emphasized to make the consumer more aware of the brand s origin, or it might be disguised in case of an unfavourable effect. As stated by Laroche, Papadopoulos, Heslop, and Mourali (2005), the presence or absence of a COO effect has significant strategic implications, both for domestic and foreign firms. Similarly, consumers who display a significant level of ethnocentrism are best targeted with an adapted marketing programme. Therefore, marketing managers operating in an international context must be aware of the sources of CBBE and understand the importance of incorporating these (Pappu, Quester and Cooksey, 2006). Only if that understanding is present are marketers capable of providing consumers with the required means to make coherent choices. In an attempt to resolve this issue, the present research provides marketers with first-hand information on how and in which settings to utilize COO cues to reinforce CBBE, service quality perception, and purchase intention.
In sum, it is important for researchers and marketers alike to know, if and how COO triggers consumers service quality perception and purchase intention as well as the impact these concepts may have on CBBE.
1.5 STRUCTURE AND SCOPE
CHAPTER 2 THEORETICAL FRAMEWORK
In the present chapter, findings of the existing literature as well as the evolution and state-of-the-art of the three concepts COO, CE, and CBBE are discussed. The examination of prevalent perspectives, conceptualisations and theories will further serve as a theoretical foundation upon which concise hypotheses may be drawn. Before considering common ground and relationships between the constructs, an individual review of each construct is provided, allowing for a more comprehensive understanding before developing the hypotheses and considering them jointly.
2.1 COUNTRY-OF-ORIGIN EFFECT
One of the major effects accompanying globalisation is the increased availability of international products and services in domestic markets. Consumers are confronted with an ever-expanding portfolio of domestic and foreign product and service offerings. However, this excessive amount of alternatives does not make the consumers decision any easier. Rather, they rely on a multitude of cues that enable them to compare the products with respect to their subjective criteria, thereby accelerating the decision-making process. The country of origin (COO) is one such clue that influences consumers in their evaluation of a product s or service s performance (an, ; Bruning, 1997). Studies have shown a general bias against foreign products and services (Peterson and Jolibert, 1995; Verlegh and Steenkamp, 1999), leading us to conclude that the COO effect on consumers evaluations and perceptions is of vital importance for international marketers. Tan and Farley s , p. 540) statement that the effect of COO on product evaluation is the most researched international aspect of consumer behaviour shows the significance within the research stream. Similarly, Peterson and Jolibert (1995) accentuate that this effect has been one of the most widely-studied phenomena in international business, marketing, and consumer behaviour literature.
Dichter , p. . Schooler conducted the first empirical test on Dichter s notion, finding significant differences in consumers evaluation of products that were, except for the name of the country stated on the Made in label, identical in all respective characteristics. Schooler s paper formed the basis for almost 50 years of intense examination of and research on the COO effect. COO effects have been examined regarding a variety of methodologies, product and service categories, source countries, and host countries with the prevailing conclusion that COO does indeed affect product and service evaluation as well as purchase decision (Al-Sulaiti and Baker, 1998; Bilkey and Nes, 1982; Liefeld, 1993; Papadopoulos and Heslop, 2003; Peterson and Jolibert, 1995; Pharr, 2005; Srinivasan and Jain, 2003; Verlegh and Steenkamp, 1999). When evaluating a product, consumers often face difficulties in accessing intrinsic information (e.g. taste, design, fit) in advance of purchase. Consequently, consumers are often forced to rely on extrinsic cues when evaluating a product (Bredahl, 2004). The COO of a product, typically operationalized through the phrase Made in… with the country name (Jaffe and Nebenzahl, 2001), is such an intangible, extrinsic attribute that serves as an informational cue and is distinct to the tangible, physical product attributes (Peterson and Jolibert, 1995). According to Liefeld (1993), extrinsic information cues (e.g. reputation, price, brand name, warranties, COO) arise from aspects external to the product. On the one hand, these extrinsic cues are recognized by the consumer as consistent and credible predictors of value and quality (Dodds, 1991; Kardes, Cronley, Kellaris, and Posavac, 2004). On the other hand, they have the characteristic of reducing the risk perceived by the consumer (Lim and Darley, 1997; Thorelli, Lim, and Ye, 1988). More generally speaking, they provide a cognitive shortcut, when intrinsic cues are not readily available, when the consumer seeks to expedite the decision-making process or when the motivation to understand intrinsic cues is lacking (Bredahl, 2004). Therefore, Cialdini , p. states: To deal with [environmental complexitiy], we need shortcuts.
[…] We must very often use our stereotypes, our rules of thumb, to classify things according to a few key features and then to respond without thinking when one or another of these trigger features is present.
for consumers regarding quality, dependability, and value for money (Han and Terpstra, ; (ong and Wyer, . The second component reinforces consumers sense of national identity by relating directly to one s group affiliation Bruning, . The COO, therefore, evolves into a badge that represents the background and history of the organization to host country constituents (Moeller, Harvey, Griffith, and Richey, 2013). Hence, at the same time, the COO of a product or service can act as a favourable signal appreciated by host country constituents, or it may tarnish the image of the firm. That means that whether consumers will like or dislike the product/service depends, at least partly, on the feelings toward the sourcing country of the product (Laroche et al., 2005). Nevertheless, Peterson and Jolibert (1995) state, the COO construct itself has been interpreted and operationalized in divergent, partly contrary, ways that led to a relatively high degree of ambiguity. Consequently, the literature lacks a generally agreed-upon definition thereof. However, in line with Jaffe and Nebenzahl (2001), country of origin here is defined as:
The country which a consumer associates with a certain product or brand as being its source, regardless of where the product is actually produced.
This definition is consistent with Johansson, Douglas, and Nonaka (1985, p. 289), who define COO as the country in which corporate headquarters of the company marketing,
the … brand is located and Usunier (2006), who accentuates that the COO is considered
to be that country consumers typically associate with a product or brand. 2.1.1 The Country-of-Origin Effect in Service Settings
Over the past decades the service sector, including international trade in services, has been characterised by significant and constant growth and has evolved into a driving force of global trade (Javalgi et al., 2003). However, whereas most research concerning the COO effect is conducted with respect to products, little research has broached the issue of potential COO effects on services (Ahmed et al., 2002, Berentzen et al., 2008; Lin and Chen, 2006; Michaelis et al., 2008; Pecotich, Pressley, and Roth, 1996). According to Michaelis et al. (2008), a major priority of service firms entering foreign markets is the establishment of trust to lower the potential risk associated with the offered services. In their study on initial trust, Michaelis et al. (2008) found that COO affects trust only in the case of a risky service and concluded that initial trust is an important precondition for
(1988) found that perceived risk is determined by several factors such as uncertainty, potential adverse consequences, as well as probability and cost of loss. In turn, the amount of perceived risk is inversely related to product and service evaluation and willingness to purchase.
A further challenge for international marketers is induced by the particular service characteristics intangibility, perishability, heterogeneity, and inseparability (Zeithaml, Parasuraman, and Berry, . Whereas a service s intangibility increases the perceived risk of a mispurchase (Berentzen et al., 2008), perishability may, in the case of excessive demand, cause quality issues (Mitchell and Greatorex, 1993). Service heterogeneity may impede the achievement of constant service quality (Guseman, 1981), and the inseparability of production and consumption demands an efficient management of supply and demand (Zeithaml, Bitner, and Gremler, 2006). Hence, according to Coulter and Coulter (2003), a service firm entering a foreign market is confronted with an even higher level of perceived risk by the consumers, because less reliable information is available and consumers have only limited access to the personal experience of others. In such a case, signalling quality and reliability to the consumer is important. Applying signaling theory, both parties the service firm and the consumer have access to different information (Connelly, Certo, Ireland and Reutzel, 2011; Soberman, 2003) and the parties must choose whether and how to communicate the required information. According to Connelly et al. (2011), the service provider typically signals service quality and reliability and the consumer as receiver must choose how to interpret it. Utilizing the service provider s origin, the COO may act as such a signal. However, the interpretation of the signal by the consumer determines the final perception of service quality and purchase intention. Hence, as services are characterised by an increased level of perceived risk, consumers must rely on other cues to signal quality and reliability. Thus, the first hypothesis on the relationship between the COO of a service provider and consumers perception of service quality is formulated as follows:
Hypothesis 1: The COO of the service provider has a significant impact on
The direction of a potential COO effect within the specific service setting is hypothesized accordingly:
Hypothesis 1a: A favourable COO image is positively related to the consumers’
perception of the quality of a service.
Hypothesis 1b: An unfavourable COO image is negatively related to the consumers’
perception of the quality of a service.
2.1.2 A Distinction between More and Less Developed Countries
An additional factor affecting the COO effect and perception of quality and risk is the COO s level of economic development. As Pecotich et al., (1996) state, consumers perceptions of service quality fluctuates depending on the service s COO as well as the COOs level of economic development. In order to reduce cognitive effort, consumers exploit categories to describe and subsequently evaluate objects (Rosch, 1978). Following categorization theory, consumers categorize a new stimulus by comparing it with a prototype. Based on that premise, countries and their level of economic development can be regarded as categories. Depending on previous experiences with prototype countries consumers tend to prefer services originating from countries in one category over services from countries in other categories. Hence, it can be expected that the level of economic development of a service provider s COO affects consumers in their service perception and evaluation. Consequently, Laroche et al. (2005) and d Astous, Voss, Colbert, Carú, Caldwell, and Courvoisier (2008) state that in contrast to highly industrialized or more developed countries (MDC), products and services originating from less developed countries (LDC) are perceived to be riskier and of lower quality. Hence, it is hypothesized that:
Hypothesis 2: Services provided by firms from MDCs are preferred compared to
similar services provided by firms from LDCs.
2.1.3 The Impact of Cultural Similarity on Country-of-Origin Effects
intention. On the whole, findings show that consumers tend to prefer products that originate from culturally similar countries than from those dissimilar to their home country (Heslop, Papadopoulos, and Bourke, 1998; Johansson et al., 1985; Wang and Lamb, 1983). However, these findings have not yet been confirmed for service settings. Hence, it is hypothesized that:
Hypothesis 3: Services stemming from countries that are culturally similar to the
country of analysis are preferred compared to services from culturally dissimilar countries.
2.1.4 Deficiencies of the Country-of-Origin Effect
While most researchers regard the COO effect on consumers evaluation of the quality of and purchase decision towards products and services as proven, a conflicting research stream has emerged, suggesting that consumers care very little about the origin of products. Several studies examining COO effects in various country and product settings did not provide evidence for a significant influence of a product s COO on quality assessment and purchase decision (Arndt, 2004; Balabanis and Diamantopoulos, 2008; Han, 1988; Liefeld, 1993; Liefeld, 2004; Samiee, 1994; Samiee, 2010; Samiee et al., 2005; Usunier, 2006; Usunier and Cestre, 2008). One of the major points of critique arising from this research stream is the limited number of variables used to test for COO effects, inevitably leading respondents to consider COO an important factor. Another commonly criticized aspect is the implicit assumption that consumers acquire or know the COO of the products and services they purchase (Liefeld, 2004). However, according to Usunier (2006) and Usunier and Cestre (2008), COO is regarded by consumers as unimportant and non-salient. Consequently, the major determinant for the importance of COO on product evaluation is, according to Liefeld (2004), the question of whether consumers acquire or know the COO of products when purchasing them.
explanation provided by Usunier (2006) is that COO information has become too difficult to obtain, resulting in consumers acting as amateur detectives (Liefeld, 1993). Combining the findings of these studies, empirical evidence suggests that consumers neither obtain much knowledge of the origins of products, nor do they intend to use such information when making a purchase decision.
Magnusson et al. (2011) take a different perspective and demonstrate that accurate COO knowledge is not a prerequisite for COO to affect brand attitude. Even without correct identification, consumers associate a product or service with a particular country. Consequently, information on consumers COO associations, correct or incorrect, is essential for international marketers. As Moeller et al. (2013, p. 95) state it: COO, …is the
anchor point for the strategic actions of the organization .
2.2 CONSUMER ETHNOCENTRISM
In the course of economic globalisation and development of multinational brands, consumers are provided with a vast variety of products and services that could have been made anywhere in the world. (owever, consumers often decide to purchase products and services that are produced in or stem from their own country. This result, an increased interaction with one s own cultural and moral identity, stems from the consumers resistance against this development and acts as a phenomenon contrary to economic globalisation (Tscheulin and Davoine, 2002; Vida and Fairhurst, 1999). Trying to explain this phenomenon, the concept of consumer ethnocentrism CE rests in the general concept of ethnocentrism introduced more than a century ago by Sumner (1906). The original concept of ethnocentrism represents the universal proclivity reflecting the view of things in which one’s own group is the centre of everything, and all
others are scaled and rated with reference to it Sumner, , p. . According to this
(Catton, 1960; Lynn, 1976; Mihalyi, 1984; Murdock, 1931; Rosenblatt, 1964; Sumner, 1906).
According to Berkman and Gilson (1978) and Markin (1974), the general applicability of the concept of ethnocentrism to the study of consumer behaviour is recognized. Hence, the concept of CE has the potential to improve understanding of how consumers compare domestic and imported products, as well as how and why their judgments may be subject to various forms of bias and error (Nisbett and Ross, 1980; John, Scott, and Bettman, . Generally, CE captures normative-based beliefs that buying domestic
products is somehow good for the country, whereas purchasing nondomestic products is deleterious to the economy, the country and fellow citizens Shimp , p. .
Consequently, Shimp and Sharma , p. define CE as a trait-like property of an
individual’s personality that
represents the beliefs held by the consumers about appropriateness, indeed morality, of purchasing foreign-made products .
2.2.1 Consumer Ethnocentrism on the Individual Level
Incorporated in the conception of CE, Shimp and Sharma (1987) accentuate that CE gives individuals a sense of identity, feelings of belongingness, and an understanding of what purchase behaviour is acceptable or unacceptable. Among the influence of affective responses towards one s own country, the ethnocentric consumer is correspondingly influenced by the normative pressure toward buying domestic products (Olsen, Granzin, and Biswas, 1993). Hence, ethnocentric individuals are characterized by a high price-inelasticity, which is marked by the tendency to disregard personal economic costs of buying a domestic product (Shankarmahesh, 2006). For highly ethnocentric consumers, purchasing imported products is a moral problem as much as it is an economic issue. As Sharma, Shimp, and Shin, (1995, p. 27) describe it, not buying foreign imports is good,
appropriate, desirable, and patriotic; buying them is bad, inappropriate, undesirable, and irresponsible. This is also reflected in Sharma et al. s (1995, p. 27) description of CEs
prejudice against imports, although it may be assumed that the overall level of consumer ethnocentricity in a social system is the aggregation of individual tendencies.
In contrast to highly ethnocentric individuals who take pride in their own values, symbols, and people, non-ethnocentric individuals evaluate products more objectively, regardless of country of origin (Netemeyer, Durvasula, and Lichtenstein, 1991). The CE concept supposedly is capable of explaining why consumers prefer domestic over imported products and similarly is able to predict consumer preferences to buy domestic (Herche, 1992). In accordance, Balabanis and Diamantopoulos (2004) examined the effect of CE towards consumer preferences for domestic and foreign products in various product categories. However, such findings have not yet been generated in service settings, thus it is hypothesized:
(ypothesis : CE has a significant impact on consumers perception of the quality of a service.
The direction of the effect CE has on consumers perceptions is hypothesized accordingly:
Hypothesis 4a: CE positively affects consumer preferences for services provided by
domestic firms.
Hypothesis 4b: CE negatively affects consumer preferences for services provided by
foreign firms.
. . Measuring Consumers Ethnocentric Tendencies
authors have assessed and approved the CETSCALE for several markets, such as the Czech Republic, Estonia, Hungary, India, the Netherlands, Poland, Russia and Spain (Bawa, 2004; Douglas and Nijssen, 2003; Durvasula, Andrews and Netemeyer, 1997; Lindquist, Vida, and Plan, 2001; Luque-Martínez, Ibanez-Zapata and Barrio-Garcia, 2000; Vida and Fairhurst, 1999).
The CETSCALE measures the appropriateness of buying domestic versus foreign products (Netemeyer et al., 1991), and thus supports the position that CE is not a matter of whether consumers are ethnocentric, but rather of how ethnocentric they are (Bawa, 2004).
2.2.3 The Impact of Demographic Variables on Consumer Ethnocentrism
Within academic research of CE, it is agreed that demographic variables impact the consumers level of ethnocentrism. Several demographic variables, such as age, gender, level of education, and income are said to trigger fluctuations of the CE level. Previous examinations (Balabanis, Diamantopoulos, Mueller and Melewar, 2001; Han, 1988; Klein and Ettenson, 1999; Nijssen, Douglas and Bressers, 2002; Shimp and Sharma, 1987) have led to dominant views, including the primary factor of age: older individuals are characterised by a higher level of CE than are younger. Moreover, the majority of researchers found that females are, on average, more consumer ethnocentric than males. In terms of education, the overwhelming consensus is that more educated individuals tend to be characterised by a lower level of CE. Findings on income, as opposed to the other demographic variables, are indecisive. While some show consumers from lower socio-ethnocentric groups to be more ethnocentric, others promote the opposite. Even though these variables have been examined in previous studies, they are of increased importance and cannot be excluded from the analysis. Over time, the society as a whole and according values, norms, and beliefs change. This, in turn, indicates the possibility of an altered effect of demographic variables on CE. Hence, it is hypothesized:
Hypothesis 5: Age and level of CE are positively correlated.
Hypothesis 6: Females exhibit greater consumer ethnocentric tendencies than do
Hypothesis 7: Level of education and level of CE are negatively correlated. Hypothesis 8: Income and level of CE are negatively correlated.
2.2.4 Origin, Extension, and Practical Relevance
Balabanis and Diamantopoulos (2004) state that in some cases consumers may actually favour foreign products over domestic ones. This notion can be explained through the application of reference group theory (Merton, 1957; Druckman, 1994) in the consumer sphere, which states that individuals are capable of recognising the home nation s flaws and therefore, include other nations in the in-group. In general, any nation is capable of becoming a positive reference group, thus supplementing the national in-group (Druckman, 1994). So the specific origin of a product or service will most likely be decisive for the variation in the consumer bias against it. The concept of CE is, therefore, inevitably linked to the COO construct. While the traditional concept of ethnocentrism acts as the basis and creates a clarification of consumers ethnocentric behaviour, the particular conception of COO can be seen as the motive to develop the CE construct (Heveroch, 2007). CE finds its origins in COO research, where in various studies the influence of country images on the purchasing behaviour of consumers was examined (Nagashima, 1970; Narayana, 1981).
2.3 CONSUMER-BASED BRAND EQUITY
The sole fact that a brand is marketed around the world is said to give it an aura of
excellence (olt, Quelch, and Taylor, . (ence, a firm s brands are often among the
first components to be extended internationally (Chabowski et al., 2013), driving globalization and affecting the proliferation of global brands (Douglas, Craig and Nijssen, 2001). Hence, branding has emerged as a top management priority in the past decade (Keller and Lehmann, 2006). Within branding, brand equity (BE) is a key marketing asset (Ambler, 2003; Davis, 2000) and its growth is a key objective (Christodoulides and de Chernatony, 2010). This is in line with previous research by Aaker (1991), who accentuated that brand equity is one of the hottest topics in management and necessary to build strong brands. A commonly accepted definition of a brand is given by Kotler (1991, p. 442), who defines a brand as
a name, term, sign, symbol, or design, or combination of them which is intended to identify the goods and services of one seller or group of sellers and to differentiate them from those
of competitors .
Moreover, Farquhar (1989) emphasizes that a brand has the ability to enhance the value of a product beyond its functional purpose. The significance of strong brands has increased constantly in the past decades, as they enable consumers to simplify decision-making, reduce risks, generate trust, and are generally tantamount with a particular level of quality (Keller, 2008). As stated by Keller (1993), knowledge that has been created about the brand in the consumers minds is perhaps a firm s most valuable asset. Traditionally, the motivation for studying BE is twofold. While one perspective considers financial aspects with the objective of measuring the value of a brand more precisely for accounting purposes, a second perspective considers the strategic effect for improving marketing activity (Christodoulides and de Chernatony, 2010; Keller, 1993). With an increase in efficiency of marketing expenses, marketers gain a more comprehensive understanding of consumer behaviour, which helps them in making better strategic decisions (Keller, 1993). Building on Aaker s (1991) conceptualisation of BE, Keller (1993) employed it from the perspective of the individual consumer and introduced the customer-based brand equity concept. Keller , p. defined it as
consumers react more favourably to the brand s marketing mix than they do to the marketing mix of another identical, unbranded product. A prerequisite for customer-based brand equity to occur is the consumers familiarity with the brand, along with their favourable, strong, and unique association with the brand (Keller, 1993). The four most important dimensions of BE from a consumer perspective are determined by Aaker (1991) as brand awareness, brand associations, perceived quality, and brand loyalty.
In line with Aaker (1991), Keller (1993), and Pappu et al. (2006), BE is conceptualized here from a consumer perspective and referred to as consumer-based brand equity (CBBE). It is defined by Aaker (1991, p. 15) as
the value consumers associate with a brand, as reflected in the dimensions of brand awareness, brand associations, perceived quality, and brand loyalty .
An overview of the CBBE conceptualisation including the four major dimensions as well as their definitions and functions can be found in table 2.1. The definitions/conceptualisations for the dimensions utilized here are based on existent definitions by Aaker (1991, p. 61), Pappu et al. (2006, p. 85), and Yoo and Donthu (2001, p. 3).
Due to CBBEs consumer orientation, the power of a brand lies in what consumers have learned, felt, seen, and heard about the brand as a result of their experiences over time (Keller, 2013). Hence, what lies in the minds and hearts of consumers determines a brand s strength. As an intangible, market-based asset, the brand provides a rich understanding of market performance (Ambler, 2003) and represents a more sustainable competitive advantage than financial and tangible assets (Hunt and Morgan, 1995).
2.4 COO AND CE: INTERDEPENDENCIES, RELATIONS, AND IMPACT ON CBBE
Brands from one country are increasingly being made available to consumers in other countries (Shocker, Srivastav,a nd Rueckert, 1994), hence, it is essential for international marketers to identify and understand the sources of equity of their brands and how a brand s performance is affected in a particular setting (Pappu et al., 2006).
As the primary focus of this work is set on how consumers perceive the foreign brand, the COO effect and the CE construct must be operationalized from a consumer perspective. The consumers level of ethnocentrism in a market determines the general attitude towards foreign products and services. However, CE only distinguishes between domestic and foreign products, not between similar products from different foreign countries. Interacting with the CE focus, the COO effect takes into account the firm s particular country of origin. The COO construct thus allows for an examination of how a particular COO is perceived by consumers of one market. Moreover, effects that are caused through the particular product or service category are likewise examined. In short, whereas CE describes the general tendency of consumers in a particular market towards purchasing foreign products or services, COO indicates the effect a particular COO has on quality perception and purchase intentions of consumers in a particular product category.
The concepts COO and CE are strongly interconnected, in such as the question whether and to which degree consumers in foreign countries evaluate products negatively, is an essential element of COO research (Heveroch, 2007). Sharma et al. (1995) describe CE as an important individual-level construct for the better understanding of country-of-origin
dynamics and Balabanis and Diamantopoulos regard CE as being of variable
CHAPTER 3 RESEARCH METHODOLOGY
To provide support for the generated hypotheses and the underlying research question, a survey was conducted which used different tools to measure the level of consumer ethnocentrism as well as the COO effect and its impact on consumers service evaluation. This chapter describes research design, data collection methods, and survey structure, and procedures. It begins with the general research design, followed by detailed information on the conceptualisation and implementation of the preference ranking, CETSCALE, and conjoint analysis. It concludes with an outline of how the theoretical foundations and measurement tools were implemented in one comprehensive survey and some information on the execution of said survey.
3.1 RESEARCH DESIGN
In a broader sense, marketing research designs can be classified as either exploratory or conclusive Malhotra, . Whereas an exploratory research design seeks to provide insights and understanding, a conclusive design aims to test specific hypotheses and scrutinize relationships. Within this thesis, a conclusive research design is chosen. Due to the specific objective of assessing cause-and-effect relationships, this project can be classified as causal conclusive. Of particular importance is if and how the manipulation of one independent variable (the combination of attribute levels) affects the dependent variable (the preference of respondent). A within-subject research design is also present, as all participants received the same survey. In order to measure consumers levels of ethnocentrism, how far they are influenced by the service provider s COO, and the relation this has to the various sections of society, the study consists of three parts: CETSCALE, conjoint analysis, and demographics.
3.2 PREFERENCE RANKING
The participants are asked to rank the 15 countries under the assumption that the flight service originating from those countries have similar attributes and features, and are sold at the same price. Rank 1 equals the most preferred COO and rank 15 the least preferred.
A preference ranking was chosen over a rating because methodological evidence suggests that rankings yield higher quality data and manifest higher discriminant
validity when evaluating a series of objects on a single scale Krosnick, , p. .
According to Balabanis and Diamantopoulos (2004), the use of rankings is consistent with advice from Liefeld (1993) that in the COO literature, where feasible, the dependent
variable measures should be choice or simulations of choice (p. 147).
3.3 CONSUMERS ET(NOCENTR)C TENDENC)ES SCALE
The CETSCALE, utilized as an instrument to measure consumers ethnocentric tendencies toward purchasing domestic versus foreign products, was introduced over a quarter century ago by Shimp and Sharma (1987). According to them, CE is recognized in the form of a general societal tendency. The scale they introduced was limited to contemporary American society. However, over time researchers have assessed the CETSCALE s validity and reliability in different settings and found it to be a reliable measure that likewise affords evidence of nomological validity in a cross-national context (Heveroch, 2007; Netemeyer et al., 1991). As stated by Shimp and Sharma (1987), the higher the score of respondents within the scale, the more likely it is that these respondents tend to choose domestic products and services over their foreign counterparts.
As the scale was originally constructed in English, the 17 items had to be translated into German for application in the German market. This was done by the researcher, a native speaker of German. Each item of the scale, along with the descriptions and instructions for the participants were then reviewed by a marketing professor at the Newcastle
University Business School whose mother tongue is German and who is also fluent in English.
The CETSCALE is a unidimensional scale that determines the scale value for each individual by summing up the rating scores of all 17 items. Response format is a 7-point-Likert-scale with the anchor points strongly agree and strongly disagree strongly agree=7; strongly disagree=1). Hence, values for each participant range from 17 – 119, whereas 119 reflects the highest degree of ethnocentric tendencies and 17 the lowest. Each statement could be evaluated with only one of the given answers from the scale, and all 17 statements had to be evaluated by the participant.
3.4 CONJOINT ANALYSIS
The conjoint analysis is an approach that aims to illustrate an individual s evaluation of a certain number of objects. )t therefore, attempts to determine the relative importance
consumers attach to salient attributes and the utilities they attach to the levels of attributes Malhotra, , p. . The traditional conjoint analysis is characterised by
a ranking of the objects according to the individual s subjective evaluation Backhaus, Erichson, Pinke, and Weiber, 2011). The stimuli used in the conjoint analysis are combinations of attribute levels determined by the researcher. In answering questions such as what influence do different components have on the total utility of an object? and which product attributes generate most satisfaction for consumers? , the conjoint analysis helps determine the utility functions of individual consumers. Whereas the service and product attributes are the independent variables, the preference of the participating individual is the dependent variable (Backhaus et al., 2011). Generally, two data sets are obtained, from which the estimation set is used to obtain the part-worth functions for the attribute levels and the holdout set is used to assess reliability and validity (Malhotra, 2010).
3.4.1 Attributes, Levels, and Description
The attributes and according attribute levels are identified by the researcher in advance. The attributes must meet several requirements. According to Backhaus et al. (2011), they must be (1) relevant, (2) suggestible by the researcher, (3) independent of one another, (4) feasible, (5) in a compensatory relationship, and (6) cannot be exclusion criteria. In general, the cognitive effort required by the respondents is significantly determined by the number of attributes and attribute levels. The attributes and according levels were identified through a pilot survey, analysis of secondary data, and discussions with an industry expert.
The three attributes identified by the researcher for an international flight within Europe are the country of origin of the airline, the price of the flight, and the number of stops during the trip. In order to examine the impact of the service provider s COO, the attribute was an obvious choice. However, the two attributes of price and number of stops were identified through an analysis of secondary data and an examination of previous studies conducted in that field (Berentzen et al., 2008; Bruning, 1997). These three attributes fulfil all requirements mentioned by Backhaus et al. (2011) as they are relevant for the service, suggestible by the researcher, independent of one another, feasible in business practice, have a compensatory relationship, and are not exclusion criteria. The feasibility characteristic has to be restrained for the attribute COO, as this cannot really be changed by the service provider: though, knowledge about the impact of a particular COO can be exploited by a firm to either disguise this information or to
emphasize it. Hence, feasibility is not existent in the form intended by Backhaus et al. (2011), but is sufficiently met for the underlying purpose.
For each attribute, three attribute levels were identified. For the service provider s COO the three levels are Switzerland, Germany, and Russia. Germany, as the home market for the analysis was a clear choice to compare the effects with foreign COOs. Switzerland and Russia were identified through a pilot survey. Switzerland was characterised by a favourable COO image, whereas Russia was characterised by a non-favourable COO image. The combination of these three COOs and their particular images allows for a comprehensive analysis and comparison. The three levels for price were € , € , and € , and the number of stops were 0, 1, and 2 stops. An overview of the attributes and attribute levels can be found in table 3.2.
3.4.2 Research Design
When designing the analysis, two major aspects concerning the composition and number of stimuli must be considered. Within the composition of the stimuli, one can choose between the pairwise approach and the full-profile method, whereas regarding the number of stimuli one must choose between a complete and a reduced design (Backhaus et al., 2011; Malhotra, 2010). The decision about both aspects is determined by the demand level the respondents are confronted with, the degree of reality within the hypothetical purchase decision, and the time respondents can and will invest. Within this thesis, the full-profile method is applied, meaning that each stimulus presented in the analysis consists of one level of each attribute. Consequently, with three attributes and three levels each, there is a total of 27 stimuli (3x3x3=27). The full-profile method
important attributes are considered and not only two as it would be the case in the pairwise approach. However, the fact that this variant is more time consuming is diminished by the option of applying a reduced design. In contrast to a complete design, a reduced design does not include all 27 possible stimuli, but rather a subset of n stimuli chosen systematically. Due to the identification of exactly three attributes and exactly three levels per attribute a symmetric design is present. This 3x3x3 symmetric design allows us to employ the Latin-Squares Method Backhaus et al., , which identifies nine of the 27 stimuli in which each attribute level is combined with every other attribute level exactly once. This results in each attribute level being represented exactly three times instead of nine in the final analysis. This reduced is preferable to a full factorial design, because a total of 27 stimuli demands much cognitive and time effort on part of the respondents and might cause them to lose focus, thereby yielding inaccurate results. An overview of the nine stimuli created through the Latin-squares method, called conjoint cards, is found in figure 3.2.
These nine conjoint cards (stimuli) represent the different offers each respondent is asked to evaluate within the analysis in accordance with his/her subjective preference. 3.4.3 Evaluation of the Stimuli
The characteristics of the traditional conjoint analysis require the participant to indicate a preference order of the stimuli according to his/her personal utility anticipations. There is a distinction between a direct and an indirect approach (Backhaus et al., 2011).
Stimulus 1 Stimulus 2 Stimulus 3
€ € €
Switzerland Germany Russia 0 Stops 1 Stop 2 Stops
Stimulus 4 Stimulus 5 Stimulus 6
€ € €
Switzerland Germany Russia 1 Stop 2 Stops 0 Stops
Stimulus 7 Stimulus 8 Stimulus 9
€ € €
However, as the indirect approach is usually applied for a rather high number of stimuli, the direct approach is applied in this study. The most common approach is the ranking, which is a nonmetric variant involving the relative evaluations of the attribute levels (Malhotra, 2010). Participants are required to rank all nine stimuli according to their subjective utility anticipations, whereby rank 1 equals the most preferred and rank 9 the least preferred combination of attribute levels. In contrast to metric approaches such as ratings, ranking data is said to reflect true consumer behaviour more accurately.
3.4.4 Estimation of the Part-worth Functions
The conjoint analysis, using the empirically generated ranking data, helps determine the part-worths for each attribute level. These utility functions are then used to obtain the metric total utility for each stimulus as well as the relative importance of the individual attributes (Backhaus et al., 2011). The basic conjoint analysis model is represented by the following formula:
∑ ∑
where
= overall utility of stimulus X
= part-worth contribution associated with the mth level of the jth attribute = 1 if the mth level of the jth attribute is present at stimulus k
= 0 otherwise
= number of levels of attribute j = number of attributes.
conjoint analysis. To analyse collected data, one can choose from a portfolio of metric and non-metric solutions, examples of which are the monotonic analysis of variance (MONANOVA) as a non-metric solution and an ordinary least squares regression (OLS) or dummy variable regression as metric alternatives. As a traditional conjoint analysis is utilized in the present thesis and respondents rank the given stimuli according to their subjective preference, a non-metric solution would be the traditional route. However, the analysis is conducted with the statistic programme SPSS, which utilizes exclusively metric measures. Hence, an OLS regression is therefore used to estimate the part-worth utilities. As the OLS regression basically provided similar results in various methodical studies, Green and Krieger (1993) recognize it as a proven and convenient approximation of non-metric techniques.
3.4.5 Limitations of the Conjoint Analysis
The conjoint analysis is an excellent tool to analyse the effect the service provider s COO has on consumers evaluation of the service: however there are some limitations which must be noted.
First, it is assumed that the important attributes can be identified by the researcher in advance and that participants make trade-offs as they evaluate the given alternatives with respect to the attributes (Malhotra, 2010). A flaw concerning the identification of important attributes was prevented through the analysis of secondary data and discussions with industry experts. Second, the conjoint assumes equal intervals by respondents when ranking the different alternatives (Backhaus et al, 2011). This is very subjective and may be applicable for some respondents, but not necessarily for others. 3.5 SAMPLING
In this particular case the sampling unit is the same as the element. As the survey is conducted online, distribution via email and social media platforms is adequate: this is not very time consuming, does not require much financial funding and has a very wide reach. Reliable data as well as a large and representative sample can be generated. To ensure an adequate level of representativeness, the sample must include a sufficient number of participants, be composed of participants of different ages, genders, levels of education, and occupations. Distribution via email and social-media platforms may lead to an overrepresentation of young participants. Hence, snowball sampling is applied to balance the sample with regard to this caveat.
3.6 SURVEY STRUCTURE AND PROCEDURE
The survey was created via the internet platform www.qualtrics.com and made accessible through a hyperlink. The hyperlink that led to the questionnaire was distributed using private email addresses and posted on social media platforms. With the objective of generating a more representative sample, snowball sampling was exploited. Due to the particular characteristics of the study, the survey language was German. Regarding the content, the survey was structured into four major sections, whereby section one was a general introduction, section two the CETSCALE, section three a conjoint analysis, and section four concerned demographic information.
The survey starts with a general introduction and lists the contact details of the researcher in case participants have further questions or wish to be informed about the results of the study. The true purpose of the study is not disclosed to the participants. Section two starts with a brief explanation of the procedure. The CETSCALE consists of
Element Sample Unit
Extent Germany and German citizens living abroad
Time October and November 2013
Target Population
Male and female individuals over 18 years who either live in Germany or have their main
domicile abroad but are German nationals
CHAPTER 4 RESULTS
In this chapter the results derived from the survey are presented, assessed and analysed. Starting with a description of the initial data set, it includes an explanation of how inconsistencies were dealt with and how the final sample was created. Section 4.1 is entirely devoted to the composition of the sample, followed by a section considering particular demographics. In the sections starting with 4.3, the actual results for the preference ranking, CETSCALE, and conjoint analysis are presented and implications for the hypotheses underlying this thesis are examined.
4.1 DATA SET
After two weeks of data collection, the survey was closed and yielded a raw data set of N=1499 respondents, of which n=638 completed the survey (survey completion rate: 42.56 per cent). The respondents who did not finish the survey were excluded from the data set opened in SPSS. All respondents who completed the survey were at least 18 years old and had their main domicile in Germany or were German nationals. The question about income was the only item in which an answer was not mandatory, and ten participants chose not to provide this information.
4.2 SAMPLE CHARACTERISTICS
In this section, characteristics of the final sample are described based on distinctive demographics such as age, gender, education, nationality, and country of residence. Please see the appendix for frequency tables and general SPSS output.
4.2.1 Age
4.2.2 Gender
Of the 638 respondents, 426 are female, accounting for 66.8 per cent of the sample and only 212 are male (33.2 per cent).
4.2.3 Education
Only one of the 638 respondents (0.2%) had not received a secondary school diploma, while 15 (2.4%) had completed secondary education (auptschule or Realschule ). The vast majority (430, or 67.4 per cent) list the A-levels as their highest academic degree. An additional respondents . % hold a bachelor s degree and 71 (11.1%) a master s degree. A small group of respondents . % report having attained a doctorate degree.
4.2.4 Nationality
The sample comprises people of 25 different nationalities, whereby 598 (93.7 per cent) are German. The second largest group contains four respondents with either a French or an US American passport. An overview of all nationalities and according frequencies from the respondents can be found in the appendix. As the survey only allowed for a single response regarding citizenship, there are no data on whether any participants hold two or more passports.
4.2.5 Country of Domicile
Because the analysis also includes Germans living abroad, a total of eight countries of residence are recorded. The vast majority of respondents (620, or 97.2 per cent) reside in Germany; however, five respondents (0.8%) currently live in Switzerland. The countries of France, the Netherlands and the USA were also mentioned, with three respondents (0.5%) each.
4.3 PREFERENCE RANKING
