AN EMPIRICAL ASSESSMENT OF THE ROLE OF COUNTRY-OF-ORIGIN
FIT AND ITS INTERACTION EFFECT WITH SERVICE QUALITY
ATTRIBUTES ON THE CONSUMER’S ATTITUDE FORMATION PROCESS:
THE CASE OF DOUBLE DEGREES IN HIGHER EDUCATION
By
LINDA ELBERSE
Newcastle University Business School and
University of Groningen
MSc. Advanced International Business Management & Marketing
3 December 2018
L.elberse@student.rug.nl L.elberse2@newcastle.ac.uk
Student number:
S2748827 (University of Groningen)
170807407 (Newcastle University Business School)
Supervisors:
Dr. Rian. Drogendijk (University of Groningen)
Dr. Elizabeth Alexander (Newcastle University Business School)
1
An empirical assessment of the role of country-of-origin fit and its interaction effect with
service quality attributes on the consumer’s attitude formation process:
The case of double degrees in higher education
Research summary: The exponential growth of international academic partnerships, in the
form of double degrees, marks out the global trend of internationalization in Higher Education (HE). The purpose of this study was to empirically test an under-examined determinant of consumer attitudes, i.e. Country of Origin (COO) fit, and its potential interaction effect with other service attributes, i.e. service quality, in the novel service context of cross-border brand alliances in HE. An experiment (n=151) was undertaken among college students in Groningen, the Netherlands. In line with prior research, results highlight the significance of COO image fit on students’ brand alliance attitude formation process in a services context. Further, our findings provide prominent insights into the relative importance of COO fit in conjunction with other service attributes initiating a new stance to an unresolved debate in the literature.
Managerial summary: In the setting of not-for-profit Higher Education Institutions (HEIs),
this study examines students’ attitudes towards international double degrees. We examine whether the partner’s level of service quality and country-of-origin (COO) image fit between partners in a double degree offering influences students’ subsequent attitudes. Results suggest that COO image fit is of similar importance as service quality attributes in partner choice. Further, we find that an increase in COO fit between partners is positively related to students’ attitudes towards that double degree. These observations do not imply that higher education (HE) managers should limit their partnerships to partner universities that delineate high perceived COO fit in students’ minds. But, it does imply some additional challenges that HE marketing managers will face when partnering with countries that lack congruence. Hence, our results provide valuable guidance for HE managers selecting partners and developing more effective marketing communication strategies.
2
Table of Contents
1. Introduction ... 4
2. Theoretical background ... 10
2.1 Service-based brand alliances ... 10
2.2 Cross-border brand alliances ... 12
3. Research framework and hypotheses development ... 14
3.1 Country-of-origin-fit ... 14
3.2 Service quality ... 17
3.3 Extension hypothesis: the moderating role of service quality ... 18
4. Methodology ... 21
4.1 Research design ... 21
4.2 Pre-test ... 22
4.3 Stimulus development for the main study ... 23
4.4 Controls in research design ... 24
4.5 Sampling, data collection and procedure ... 25
4.6 Controls for random assignment ... 27
4.7 Measures and measurements ... 29
5. Results ... 32
5.1 Assumptions ... 32
5.2 Testing main effects ... 33
5.3 Testing interaction effects ... 34
5.4 Robustness of results ... 36
5.5 Overview of findings ... 38
6. Discussion ... 40
6.1 Theoretical implications ... 40
6.2 Managerial implications ... 43
7. Limitations and future research ... 45
References ... 49
Appendix 1. Pre-test: Country-of-Origin fit (Dutch version) ... 55
Appendix 2. Pre-test: Country-of-Origin fit (English version) ... 58
Appendix 3. Main survey - stimuli HN and LN (Dutch version) ... 61
3
Appendix 5. Stimulus development ... 67
Appendix 6. QS Higher Education System Strength Ranking ... 70
Appendix 7. Histograms of the distributions ... 71
4
1. Introduction
Imperatives of the market, including global student mobility, reduced university funding, government-backed recruitment campaigns and growing competition has driven the worldwide trend of internationalization in Higher Education (HE) (Hemsley-Brown, Melewar, Nguyen, & Wilson, 2015; Altbach, 2015; Kalafatis, Ledden, Riley, & Singh, 2016). Offering transnational HE through offshore branch campuses or collaborative programs with other countries has become a central part of university systems and policies to deal with this global trend (Altbach, 2015). The projected growth of internationally mobile students demonstrates that cross-border study is, and will continue to be, “big business” (Altbach, 2013, p.38). This study will focus on one particular form of internationalization, namely, cross-border brand alliances, i.e. the short- or long-term combinations of two or more individual brands, in which the Higher Education Institutions (HEIs) involved are headquartered in different countries or markets (Simonin & Ruth, 1998; Bluemelhuber, Carter, & Lambe, 2007).
5 more preferred and common type of brand alliance compared to joint or consecutive programs (Redd, 2008). This is mainly due to the legal barriers, quality assurance and accreditation issues in granting a joint diploma (Aerden & Reczulska, 2010; Knight, 2011; Obst & Kuder, 2012). Therefore, this paper will focus on one focal HE brand alliance type, namely, double degrees, described as “a collaborative program which awards two individual qualifications at equivalent levels upon completion of the collaborative program requirements established by the two partner institutions” (Knight, 2011, p.301).
Despite the rapid growth of cross-border brand alliances, particularly double degree programs between HEIs, research on these academic partnerships remains limited (Knight, 2011; Melewar & Nguyen, 2015). Hemsley-Brown et al. in a special section of the journal of business research (2016) identified only two scholarly references that engaged with the theoretical and empirical issues of brand alliances in HE. Naidoo and Hollebeek (2016) examined dual-degree purchase intentions, and Kalafatis et al. (2016) explored students´ perceptions regarding the added value of dual degrees in HE. Surprisingly, both studies did not address the potential influence of partner’s country image or country-of-origin (COO) on respectively students´ purchase intentions or perceptions of added value. Yet, in our global economy one area that has grown significantly is international academic partnerships (Altbach, 2015; Wilkins, Butt, & Heffernan, 2018). Specifically, Knight (2011, p.304) highlighted the extant trend that over the years “double degree programs exhibit more inter-regional pairings that are remarkably international in scope”.
6 alliance" (Bluemelhuber et al., 2007, p.433), on brand alliance perceptions (Li & He, 2013; Lee, Lee, & Lee, 2013).
However, these few COO studies on brand alliances only have been conducted with respect to products. Yet, no research has been devoted to understanding COO fit in relation to a service sector context. Though, professional services, such as HE, provide a distinct research context compared to products due to its focus on people, intangibility, heterogeneity, perishability, and long-lasting relationships of continuous delivery with customers (Zeithaml, Parasuraman, & Berry, 1985; Mitchell, 1999; Hemsley-Brown & Oplatka, 2006). Extant research has verified that these distinctive service characteristics, particularly evident in the high-involvement service context of double degrees, affect consumers´ decisions and risk perceptions (Wilson, Zeithaml, Bitner, & Gremler, 2008; Kamal Basha, Sweeney, & Soutar, 2015). Within this context, Wilkins et al. (2018) argued that COO image may be used as a proxy for the institutional reputation and status of a HEI. Therefore, even low-quality colleges and universities from developed nations, such as the United Kingdom (UK) or the United States (US), may deliver value in cross-border brand alliances (Wilkins et al., 2018). The high COO image of the UK and the US regarding its HE system is likely to replace the potential low reputation that HE institutions have in those particular countries. Hence, the impact of COO fit in service settings, specifically in the growing field of international double degrees in HE, is underlined as a significant research gap that will be addressed in this study (Kalafatis et al., 2016).
7 found contradictory results. Elliot and Cameron (1994) found that the COO effect is the least important choice determinant for consumer product or service evaluation in relation to other functional or performance attributes, such as product or service quality, which the authors highlighted as the most influential attributes in guiding brand evaluations. Whereas others reveal exact opposite findings and posit that COO image, as a heuristic and more abstract basis for consumer evaluations, takes precedence over other functional or performance attributes (Hong & Wyer, 1989; Okechuku & Onyemah, 1999). Due to these contradictory results, an important question remains unanswered regarding the relative importance of COO in conjunction with other service attributes (Bluemelhuber et al., 2007). Hence, it is important to understand the relative strength of multiple pieces of either abstract or concrete information with regard to all participating brands in the consumer´s attitude formation process. Therefore, this research aims to understand the cognitive structure or factors affecting this process to provide more integrative and theoretical insights into the interaction (as well as the direct) effect of COO fit with service quality in the consumer´s brand alliance attitude formation process.
8 In addition to submitting a subset of their work to a test of replication in a non-product service context, this study will extend the literature by exploring the potential interaction effect of COO fit with certain service quality cues of the brand alliance partner. Relying on the Elaboration Likelihood Model (ELM) it is hypothesized that additional information provided about the level of service quality (i.e. rankings, accreditations) of the double degree will counterbalance the negative effects of COO fit. Furthermore, it is expected that the level of service quality cues provided will have a positive direct effect on consumers’ attitudes towards the cross-border service brand alliance.
To test these hypotheses a mixed factorial experimental design is conducted representing varying degrees of COO fit and service quality. This study’s findings show the importance of COO fit as a significant determinant of students’ attitudes towards cross-border double degrees in HE. Further, we add new findings to the extant debate in literature regarding the relative importance of COO in conjunction with other service attributes. Our results show that the interaction effect between COO fit and service quality was non-significant, indicating that the COO fit effects among treatments, with or without service quality cues, can be assumed to remain constant. Yet, this finding is contrasting with prior literature that found significant interaction effects (Hong & Wyer, 1989; Elliot & Cameron, 1994; Okechuku & Onyemah, 1999). Accordingly, this study forms a new stance to a still unresolved debate in the literature.
9 alliances (Lanseng & Olsen, 2012). This study shows the importance for managers to consider COO fit in their partner choice, since low-fit partners presumably will negatively influence the consumer’s brand alliance attitude and consequently harm the alliance (Bluemelhuber et al., 2007; Lanseng & Olsen, 2012).
10
2. Theoretical background
2.1 Service-based brand alliances
Over the past two decades brand alliances have emerged as an increasing popular brand leverage strategy, which has expanded to include both product- and service sectors (e.g. Simonin & Ruth, 1998; Bleijerveld, Gremler, & Lemmink, 2015). The worldwide service’s average contribution to GDP and value added has increased up to 65% in 2016 (Worldbank, 2018), indicating that the service sector plays a vital and growing role in our global economy. Yet, brand alliance research remains overwhelmingly focused on consumers’ attitudes towards alliances in commercial product sectors (e.g. Contractor, Kundu, & Hsu, 2003; Aaker & Keller, 1990; Simonin & Ruth, 1998; Washburn, Till, & Priluck, 2004; Thompson & Strutton, 2012; Samuelson, Olsen, & Keller, 2015). This is remarkable since the focus on people, intangibility, heterogeneity, perishability, and long-lasting relationships of continuous delivery with customers are several distinguishing characteristics that differentiates services from products that should be acknowledged (Zeithaml et al., 1985; Mitchell, 1999; Hemsley-Brown & Oplatka, 2006). The magnitude of differences between products and services can be visualised by placing them on the tangibility ‘spectrum’ of Shostack (1977) as depicted in figure 2.1.
11 The first category, located at the relatively more tangible end of the continuum, involves research-based services (e.g. checking accounts or fast-food outlets) that possess various attributes which consumers are able to evaluate prior to purchasing. The second category involves experience-based services (e.g. hairdressing or travel services/airlines) which refer to services that consumers are solely able to assess after consumption. The last category involves credence-based services (e.g. medical services or education services/teaching), referring to services at the right end of the intangibility continuum. These services are argued to be the most difficult for consumers to evaluate even after consumption (Kotler, 2003; Kamal Basha et al., 2015). In the case of HE double degrees we are dealing with a credence-based service located at the end of the continuum as the most intangible-dominant exemplar (Shostack, 1977).
It is argued that the higher the level of intangibility in services, the greater the potential divergence of outcomes will be from extant research on product-based brand alliances (Shostack, 1977). This because the distinct characteristics of services, such as its intangibility, can render brand-based signalling effects regarding perceived service quality and alter extant risk perceptions (Berentzen et al., 2008; Li & He, 2013). Therefore, it is of crucial importance to include service-based sectors, such as HE, in empirical brand alliance research in order to examine if extant effects in the literature may be stronger, weaker, non-existent or generalizable from product-based to service-based sectors.
12
2.2 Cross-border brand alliances
Next to the importance of exploring the intricacies of service-based brand alliances, branding issues in cross-border brand alliances are also highlighted in the literature as “an important but unexplored territory” (Li & He, 2013, p.94). In international management practice cross-border brand alliances have already become a popular and favoured strategy, as in today’s world it is a sensible and crucial tactic for companies to be able to compete in the increasingly globalizing market (Michailova & Ang, 2008; Li & He, 2013; Lee et al., 2013). Despite the increasing importance of cross-border brand alliances, relatively few studies have empirically addressed the consumer’s attitude formation process in this particular context (e.g. Voss & Tansuhai, 1999; Bluemelhuber et al., 2007; Lee et al., 2013). Moreover, these studies only examined this issue with regard to product-based brand alliances.
The lack of empirical research in this field is worthy of attention since cross-border brand alliances can be seen as a unique form of a brand alliance, which involves brands that are headquartered in different countries, therefore COO effects may come into play (Bluemelhuber et al., 2007). A consumer’s COO image refers to the perceptions and generalisations one uses to stereotype particular countries (Maheswaran, 1994; Lampert & Jaffe, 1998). It has been characterized as a secondary brand association or extrinsic cue that guides brand evaluation without directly influencing service (functional) performance (Keller, 1993; Peterson & Jolibert, 1995; Lee et al., 2016).
13 that attempt to enter foreign markets through a brand alliance (Agrawal & Kamakura, 1999; Magnusson et al., 2014; Lee, Lockshin, & Greenacre, 2016).
14
3. Research framework and hypotheses development
3.1 Country-of-origin-fit
The impact of COO on consumers’ perceptions regarding brand alliances can be reflected through the variable of COO fit (Bluembelhuber et al., 2007; Lee et al., 2013). Fit in brand alliances between the parent brand and the extension category has been identified in the literature as key factor to influence consumers’ perceptions and ultimately brand extensions´ success (e.g. Aaker & Keller, 1990; Simonin & Ruth, 1998; Thompson & Strutton, 2012; Samuelson et al., 2015). Brand alliance ‘fit’ can be based on both product-related- and non-product related attributes and is commonly described in terms of similarity, congruency, consistency, compatibility or match (Lanseng & Olson, 2012).
In a brand alliance, such as a double degree offering that is co-branded and promoted by both HEIs, both partners bring their, either concrete or abstract, brand associations into the alliance (Martin, Stewart, & Matta, 2005; Samuelson et al., 2015). Accordingly, the consumer’s brand alliance attitude, i.e. “the consumer’s perception of the overall quality of the brand alliance” (Aaker & Keller, 1990, p.29), is based on the simultaneous evaluation of these concurrent sets of different associations and therewith different bases of fit will emerge (Lanseng & Olson, 2012). COO fit, which will be the focus in this paper, is defined by Bluemelhuber et al. (2007, p.433) as “the consumer’s perception of compatibility of the two countries of origin involved in the brand alliance”. In line with Bluemelhuber et al. (2007) compatibility will be assessed by comparing the consumer’s perceptions of the countries’ ability to deliver high quality HE services.
15 al., 2007). In this case the information integration theory describes the process in which two (or more) different memory objects, related to consumers’ perceptions of particular countries’ HE system strength, will be combined and integrated to configure new attitudes (Anderson, 1981; Lee et al., 2016).
Based on the associative network theory, it is assumed that each country in the alliance is represented by associative schema (i.e. knowledge structures or stereotypical beliefs) in memory (Anderson, 1983). Relying on the information integration and congruence theory, COO fit originates from the degree of ‘congruence’ between these two “schemas” in consumer memory when this information is combined and integrated to make an overall judgement (Anderson, 1981; Lanseng & Olsen, 2012). When consumers perceive substantive differences or inconsistencies between the combined associative schemas related to the countries involved in the alliance, in terms of their product- or service country associations, it may negatively influence consumers’ responses towards the alliance (Bluemelhuber et al., 2007). An explanation has been provided by the congruity theory derived from an early psychological study of Osgood and Tannabaum (1955), which is a mental consistency theory in which people prefer to maintain consistency or congruence among their attitudes or thoughts (i.e. associative schema). Accordingly, Osgood and Tannenbaum (1955) found that the lack of credence for incongruous messages drives mental processes of dissonance and confusion which will consequently lead to an unfavourable view towards the offered product or service combination. Hence, the potential cognitive inconsistency between the COO images of two combined brands is an important consideration for marketing scholars and practitioners to explain the consumer’s attitude formation process towards the brand alliance.
16 These distinctive characteristics, such as its intangibility, inseparability or heterogeneity, make it more difficult for consumers to gather explicit information about the value of the service alliance to form their respective attitudes (Wilson et al., 2008). Therefore, due to the relative inability to obtain direct and certain information about particular service attributes as compared to products, consumers commonly infer their service attitudes by turning to more readily accessible cues such as COO images (Berentzen et al., 2008; Bleijerveld et al., 2015).
In this particular service context, the literature has shown that HEIs from developed nations, particularly English-speaking nations, have a strong quality association with HE services, while developing countries have a weaker association with HE system strength (Cheung, Yuen, Yuen, & Cheng, 2010; Altbach, 2013). Accordingly, based on the assumption of fit, it would be interesting to explore if consumers’ brand alliance attitudes would be effected if only the COO in the advertisement will be changed from a country with weaker associations to a country with strong associations regarding the quality of HE services (i.e. representing the degree of COO fit). Accordingly, it will be expected that low COO fit service pairings that show inconsistency and incongruence among both HEIs, will lead to high(er) risk perceptions, negative classifications, and ultimately negative overall brand alliance attitudes (Osgood and Tannabaum, 1955; Weiner, 1985; Becker-Olsen & Hill, 2009).
In line with Bluemelhuber et al.’s (2007) study on product-based brand alliances, it will be hypothesized that COO fit is an important explanatory variable for the formation of consumers’ brand alliance attitudes in a HE services context.
17
3.2 Service quality
Favourable consumer attitudes or evaluations towards products or services are tied to the formation of consumer beliefs about its perceived quality based on brand attribute cues (Aaker & Keller, 1990; Van Osselaer & Alba, 2000). Prior research found that students’ attitudes and university choices, in addition to a university’s COO, is largely dependent on various service quality cues, such as the university’s ranking, accreditations or its number of known professors (Shanka, Quintal, & Taylor, 2006; Bleijerveld et al., 2015). This available evidence indicates that positive service quality evaluations of a university (i.e. rankings, accreditations) will lead to positive attitudes towards that particular university.
Based on the information integration theory, it is expected that when two university brands become partners in a brand alliance, both universities´ quality associations will be combined and integrated by the consumer to configure their brand alliance attitudes. As outlined in section 2.1, a great amount of studies have examined this assumption by empirically testing this on product-brand alliances. However, limited evidence exists that these effects remain evident in a services context (Bourdeau, Cronin, & Voorhees, 2007; Bleijerveld et al., 2015). Coherent with prior research, it is expected that students’ attitudes towards the brand alliance is greatest when the service quality of both HEIs involved is high, somewhere in between when there is a mixed combination of a high- and low-service quality HEIs, and the lowest when both HEIs’ perceived service quality is low (Bleijerveld et al., 2015; Washburn et al., 2004).
18
3.3 Extension hypothesis: the moderating role of service quality
Due to contradictory results in the literature, an important question remains unanswered regarding the relative importance of COO fit in conjunction with other service attributes such as service quality (Bluemelhuber et al., 2007). We have reasonedthrough different theoretical frameworks, namely the associative network theory and information integration theory, that various mental processes drive brand alliance evaluations. Accordingly, we described the alliance attitude formation process in which two (or more) different memory objects of each piece of information about all alliance participants will be combined and integrated to make an overall judgement (Anderson, 1981; Lanseng & Olsen, 2012). Next to combining and integrating simple pieces of information, such as COO fit or service quality cues, the information integration theory asserts that these different attributes are further combined in a weighted assessment to configure new brand alliance attitudes (Anderson, 1981). Since we are dealing with a ‘weighted assessment’ of multiple pieces of information, it is important to determine the relative strength of each association or piece of information in engendering consumers´ attitudes of the cross-border brand alliance.
19 information, namely, the central (i.e. high cognitive elaboration) and peripheral (i.e. low cognitive elaboration) processing route (Petty & Cacioppo, 1986). It is argued that service quality cues are primary associations or internal cues that process via the central route of elaboration, while COO fit can be described as a secondary association or external cue that will be processed via the peripheral route of elaboration.
Based on simple heuristics and decision rules, such as the ELM and the hierarchy of brand associations, it will be argued that in the absence of primary brand associations (e.g. other service quality cues), students will rely on COO fit as an extrinsic cue to infer their brand alliance attitudes (e.g. Aaker & Keller, 1990; Petty & Cacioppo, 1986; Berentzen et al., 2008). Therefore, in line with hypothesis 1, it will be assumed that if no other service quality cues are provided to the student, COO fit has a significant impact on the cross-border service brand alliance attitude.
Yet, relying on the same framework and extant findings in the literature it will be argued that when consumers are provided with additional intrinsic service attribute cues, they will have more developed brand alliance associations and will be able to evaluate the combined service more objectively (Bluemelhuber et al., 2007; Lee et al., 2016). Accordingly, following the ELM and hierarchy of brand associations, it is assumed that when intrinsic service information is available, consumers will rely less on COO fit as a halo or extrinsic cue to engender their overall brand alliance attitudes (Petty & Cacioppo, 1986; Aaker & Keller, 1990).
20 attitude formation process (Anderson, 1981). Within this condition, based on the ELM and the hierarchy of brand associations, it will be hypothesized that the effect of COO fit will be dominated by the precedential and stronger effect of service quality cues upon consumer attitudes towards the brand alliance. Hence, it is assumed that additional information about the double degree’s level of service quality (i.e. accreditations, rankings) will counterbalance the presumed negative effects of COO fit in the brand alliance attitude process.
Hypothesis 3. If additional quality cues are provided to a student, the impact of COO fit on consumers’ cross-border service brand alliance attitude will diminish.
An overview of this study’s research framework is visualised in figure 3.1.
Figure 3.1 Theoretical framework for analysing consumers’ attitudes towards cross-border service-brand alliances and its interaction effect with service quality.
Country of origin fit
21
4. Methodology
4.1 Research design
This study responds to the trend of increasing cross-border brand alliances in HE and the request of Kalafatis et al. (2016) to include international double degrees for empirical assessment.
To test the hypotheses, this study adopted a 2 (COO fit: high vs. low) × 3 (Service quality: high vs. low. vs. no cues)mixed factorial design. A mixed factorial design, commonly referred to as ‘split-plot design’, combines two designs, namely a between-subjects and within-subjects design (table 4.1.1) (Verma, 2015). The inclusion of both within-within-subjects and between-subjects variables allowed to combine the benefits of each method, such as the requirement of fewer subjects and the prevention of potential boredom or fatigue effects (Verma, 2015). The partner brand’s service quality (the between-subjects variable) was manipulated by the information provided to the respondent. To manipulate COO fit (the within-subjects variable) between the home university and partner university a pre-test (n=42) was conducted.
Table 4.1.1. Visual representation of the two-way mixed factorial design
Mixed-factorial design Between-subjects design
Wi thi n -subj ect s de si
gn High Low None
High HH HL HN
Low LH LL LN
22
4.2 Pre-test
The goal of the pre-test was to select a partner country that shows high COO fit with students’ home university (the university of Groningen) for the manipulation. A total of 42 undergraduate Dutch business students were asked to fill-out the survey (Appendix 1 and 2).
The pre-test included a list of countries with varying levels of HE system strength derived from the QS Higher Education System Strength Ranking (2018) (Appendix 6). From this top 50 location ranking, five countries were selected at random from the 25 highest ranked countries of the list, and five countries were selected at random from the 25 lowest ranked countries of this list. Further, we added two English-speaking countries to check if the literature is consistent about students’ high-perceived education quality of English-speaking countries (i.e. US and UK) and see whether Dutch students perceive high COO fit with these particular countries (Obst, Kuder, & Banks, 2011). The selected countries are visualised in bold in
appendix 6.
Students were asked to rate each country on two 7-point items (1=low consistency/complementarity, 7=high consistency/complementarity), adopted from Aaker and Keller (1990), on their level COO fit with their home university in the Netherlands. To avoid confusion, each construct (i.e. consistency and complementarity) were explained in the questionnaires. Based on an averaged index of the two items, Sweden was rated with the highest fit (6.5) followed by Germany (6.18). Indonesia had the lowest fit (2.29), followed by Egypt (2.50).
23
4.3 Stimulus development for the main study
Various levels of COO fit and service quality must be established in the study to compare the effects of each level on the consumer’s brand alliance attitude formation process. The stimulus development for this study builds on the study of Bleijerveld et al. (2015) who created various stimulus/scenarios for brand alliances in HE (Appendix 5).
The stimulus format was standardized except for the manipulated variables (i.e. COO fit and service quality). A university in the Netherlands was chosen as the ‘context brand’ to increase the relevance of the survey for the respondents who were students at the university of Groningen. In line with prior literature the rankings, accreditations, known professors and age of the initiated partner institutions were used as proxy signals of their respective service quality in the stimuli (Shanka et al., 2006; Bleijerveld et al., 2015; Kalafatis, et al., 2016; Osipian, 2016). COO fit was manipulated by using the pre-test results wherein the Netherlands and Sweden represented the high COO fit pair and the Netherlands and Indonesia represented the low COO fit pair. The use of Indonesia can be seen as a more striking choice in this study’s context, since Indonesia was a Dutch colony until 1949 (Van Imhoff & Beets, 2004). Given this history between the Netherlands and Indonesia, potential lingering effects might have biased Dutch students’ perceptions. However, we assume that among the generation of students in this study’s sample (born in the period 1994-2000) this potential bias will be relatively small (Van Imhoff & Beets, 2004).
24 prize winners >. During this 1.5-year Double Degree program, students will follow courses at the partner university for six months. This program will give students the opportunity to experience two different and unique academic environments in which the strengths and perspectives of each institute will be combined.” In the questionnaire this process was repeated with another degree of COO fit. Within these stimuli variance in age, ranking, accreditations and number of Nobel prize winners represented the factor ‘service quality’ and variance in the COO of the partner university represented the factor ‘COO fit’ (Appendix 5).
4.4 Controls in research design
Regarding research design, this study controlled for potential bias through particular brand name assertions or familiarity by excluding the brand name of the given partner universities in the alliance (Rao et al., 1999). Additionally, by excluding the partner’s brand name, this study also controlled for brand fit between partners (Simonin & Ruth, 1998). Further, we controlled for product fit as both partners originate from the same industry (Simonin & Ruth, 1998). Lastly, the use of fictitious brand alliances allowed to control for students prior knowledge about the alliance and provided the freedom to choose a standardized setting with varying product characteristics for the advantages of ensuring internal validity (Li & He 2013). Although using fictitious stimuli provided the ideal setting to ensure internal validity, simultaneously this approach is known as a methodological limitation due to its lack of external validity and realism (Li & He, 2013; Van der Lans, Van den Bergh, & Dieleman, 2014). Therefore, to bolster the external validity of this study we included two stimuli (i.e. HR-LR,
table 4.5.1) representing existing brand alliances that were offered at the context university
25
4.5 Sampling, data collection and procedure
For the main survey, this study adopted a purposeful homogeneous sampling methodology to explore the study’s main hypotheses. Undergraduate Dutch business students (n=301) at the university of Groningen were approached within a normal classroom situation and asked to participate in a research study. This sample was chosen for three reasons. First, the sample included only (Dutch) business programs to reduce potential bias in differences of consumers’ brand alliance evaluations across different academic fields (Haidoo & Hollebeek, 2016). Second, to ensure familiarity with the Dutch HE system, the population comprised only Dutch students in an undergraduate degree from a Dutch university (Lanseng & Olsen, 2012). Lastly, to control for potential differences between respondent’s pre-purchase criteria versus post-purchase criteria for brand alliance evaluation, this study only included undergraduate business students to assess students’ perceptions primarily in a pre-purchase decision context (Gardial, Clemons, Woodruff, & Burns, 1994). Besides controlling for potential differences, the focus on pre-purchase evaluation stage will be of particular importance to increase the practical relevance of our research since the understanding of pre-purchase information used by service consumers is argued to be of vital importance for both marketing and strategic managers (Murray, 1991).
26 respondents had to rate their preferences for either domestic (i.e. consumer ethnocentrism) or foreign services (i.e. consumer cosmopolitism) and answer some demographic survey questions (i.e. age and gender) (Appendix 3 and 4).
A total of 301 respondents participated in the experiment and were randomly assigned to 1 of the 6 different versions of fictious or real brand alliance combinations (table 4.5.1). The sample consisted of 197 male (65.4%) and 101 female (33.6%) respondents. The average age was 19 years (Mage=19.67, SD= 1.03) and about 75% of the respondents were 19 or 20 years old.
Table 4.5.1. Descriptive information on the total sample.
Questionnaire Male Female Not listed Average age
N N (%) N (%) N (%) M SD
Without Quality cues (HN-LN)
50 64 36 - 19.62 1.16
With only high-quality cues (HH-LH)
50 72 26 2 19.74 0.94
With only low-quality cues (HL-LL)
51 60.8 37.3 2 19.55 0.97
With respectively high- and low-quality cues (HH-LL)
50 76 24 19.80 1.07
With respectively low- and high-quality cues (HL-LH)
50 60 38 - 19.36 0.99
With real quality cues (HR-LR)
50 60 40 - 19.92 1.07
Total 301 65.4 33.6 0,7 19.67 1.03
However, as mentioned in the ‘research design’ section, a mixed-factorial design, visualised in
table 4.1.1, were chosen to analyse this study’s main hypotheses. Hence, respondents of the
questionnaire combinations that were not corresponding to this particular design (i.e. HH-LL, HL-LH and HR-LR) were excluded.
27 experimental conditions of the between-subjects factor: high quality cues (n=50), low quality cues (n=51), no quality cues (n=50) (marked grey in table 4.5.1). Among them, 6 responses had 1 missing value in either the cosmopolitan or ethnocentrism scale. To adjust for these missing responses we used pairwise deletion as it maximizes all data available (Peugh & Enders, 2004). The selected sample consisted of 99 male (65.6%) and 50 female (33.1%) respondents and 2 respondents that did not identify being either male or female (1,3%). The average age was 19 years (Mage=19.64, SD= 1.02) and about 75% of the respondents were 19 or 20 years old. In the remaining of this paper this new, narrowed, sample (n=151) has been used to provide the descriptive statistics and conduct the data analyses.
4.6 Controls for random assignment
28 successful, since in quantitative research “nothing should be left to chance” (Mertler, 2015, p.93).
Consequently, to check for the robustness of our assumption of random assignment we included four extraneous variables, two demographic constructs (i.e. age and gender) and two social psychological constructs (i.e. consumer ethnocentrism, consumer cosmopolitism and social desirability bias). Regarding the demographic constructs, we controlled for the random assignment of age since Holbrook (1993) found a significant effect of age on the development of consumer preferences or attitudes. Although this study’s sample only exhibited a small variance in age (MAge=19.67, SD=1.03), as we only included only undergraduate business students, it can be argued that even small age differences have a decisive influence on consumers’ attitudes. In this sense, Schuman and Scott (1989, p.361) concluded that “memories are structured by age” and contended that ‘late adolescence’ and ‘early adulthood’ are the most influential years in life that shape people’s unique personal point of view. Accordingly, it is expected that relatively ‘older’ students may have more experience or associations with double degrees, studying abroad or particular university brands compared to younger students and therefore possess more developed brand associations. Secondly, this study controlled for potential gender differences in the attitude formation process. Kwun (2011) found a link between gender and consumers’ attitudes. Based on the selectivity theory, Kwun (2011) concluded that the general attitude formation process significantly differed between male and female consumers. Results showed that within the attitude formation process females configurate their judgements by extensively processing nearly all available cues of information, whereas males only form their attitudes by focusing on particular cues which were most prominent or available (Kwun, 2011).
Zeugner-29 Roth, Žabkar and Diamantopoulos (2015) found that both consumer ethnocentrism and cosmopolitism affect consumers’ attitudes towards domestic or foreign products. Ethnocentrism refers to a ‘anti-out group’ construct in which people display a negative stance towards buying foreign study programmes (Zeugner-Roth et al., 2015). Cosmopolitism is a ‘pro-out group’ construct that is expected to positively drive consumer preferences towards foreign study programmes (Riefler, Diamantopoulos, & Siguaw, 2012). Lastly, we controlled for potential social desirability bias related to these socio-psychological constructs, defined as “the systematic error in self-report measures resulting from the desire of respondents to avoid embarrassment and project a favourable image to others” (Fisher, 1993, p.303). We tried to avoid response bias by assuring anonymity in that ‘reactions are processed completely anonymously’, and stressing that there are no ‘right’ or ‘wrong’ answers by adding the following sentence “People have different opinions … as something positive, whereas others see this as something negative” (Appendix 3 and 4).
4.7 Measures and measurements
The independent variables, COO fit and service quality, were manipulated in the questionnaire design by pairing various developed stimuli (Appendix 5). COO fit was manipulated by showing two different stimuli in each questionnaire representing high COO fit as well as low COO fit. Service quality represented three categorical groups, namely ‘high quality cues’, ‘low quality cues’, and ‘no quality cues’. We combined varying degrees of service quality in the 3 different versions of the questionnaire as represented in table 4.7.1.
30 We measured the demographic control variables ‘Age’ as a continuous variable and ‘Gender’ as a categorical variable (1=male, 2=female, 3=not listed). Consumer ethnocentrism was measured by using a slightly adapted form to fit the study context of the five-item version of the ‘CETSCALE’ (e.g. “Dutch people should not buy foreign programmes”, “It is not right to purchase foreign programmes”), which is used and extensively validated by many different scholars (e.g. Zeugner-Roth et al., 2015; Verlegh, 2007). We used the validated and refined 11-item ‘three-dimensional C-COSMO measure’ of Riefler et al., (2012) to measure cosmopolitism, including ‘open-mindedness’, ‘diversity appreciation’ and ‘consumption transcending borders’, to measure cosmopolitism (e.g. respectively, “I like to learn about other cultures”, “Having access to products coming from many different countries is valuable to me’, “I like reading foreign newspapers and magazines”).
On each measurement scale is a reliability analyses conducted to measure the internal consistency of the chosen set of scale items. All reliabilities for each factor that exceeded the .70 criterion are to be considered acceptable (Nunnally, 1978). The cronbach’s alpha on all measurement scales were found to be reliable (α>.70) as depicted in table 4.7.1.
Table 4.7.1. Overview of all measurements and scales used in the questionnaire
Scale Representative
sources
Scale type Item CA
COO fit (manipulated)
- Dichotomous
group variable
1. High COO fit 2. Low COO fit
Service quality (manipulated)
- Categorical
group variable
1. High service quality cues 2. Low service quality cues 3. No service quality cues
Attitude toward the cross-border brand alliance (Aalliance) Osgood et al. (1957);
Simonin and Ruth (1998)
7-point Likert scale
1. I perceive this Double Degree to be favourable
2. I perceive this Double Degree to be positive
3. I perceive this Double Degree to be desirable αsweden= .835 αindonesia= .867 Consumer ethnocentrism (CE) Zeugner-Roth, Zabkar and Diamantopoulos (2015) 7-point Likert scale
1. Dutch people should not buy foreign programmes, this hurts domestic universities and causes unemployment.
31
2. it is not right to purchase foreign programmes, because this puts Dutch people out of jobs. 3. A real Dutch citizen should
always buy domestic programmes.
4. I always prefer domestic programmes over foreign ones. 5. We should follow Dutch
programmes, instead of letting other countries get rich off us. Consumer cosmopolitism (CC) Riefler et al. 2012; Zeugner-Roth et al. 2015 7-point Likert scale Open-mindedness
1. I like to learn about other cultures 2. I like having the opportunity to
meet people from many different countries
3. I like to have contact with people from different cultures
4. I have got a real interest in other countries
Diversity appreciation
5. Having access to products coming from many different countries is valuable to me
6. The availability of foreign products in the domestic market provides valuable diversity 7. I enjoy being offered a wide range
of products coming from various countries
Consumption transcending borders 8. I like reading foreign newspapers and magazines to inform myself what is happening around the world
9. I like spending my holidays in foreign countries
10. I like trying out things that are consumed elsewhere in the world. 11. I like trying original dishes from
32
5. Results
In this study a mixed design analysis with two-way ANOVA was used to test the significance of main and simple effects of service quality (between-subjects) and COO fit (within-subjects) factors (Table 5.1). The hypotheses were tested at the significance level of 0.05.
Table 5.1. Mean ratings on cross-border brand alliance attitudes as a function of COO fit and service quality
5.1 Assumptions
In order to test the suitability of using the mixed design, its required assumptions (e.g. normality, homogeneity of variance-covariance, homogeneity of variance and sphericity) were tested. Hence, the Shapiro-Wilk statistic was either significant (p≤.05) or marginally significant (P≤.10) in each of the 6 group scores, suggesting that the data is not normally distributed and thus the assumption of normality to be violated. Since any assessment should also include the evaluation of the normality of histograms (Field, 2013), it can be concluded that we are dealing with a negative skewed distribution among all six group scores (Appendix 7). Nevertheless, it is argued that with large enough sample sizes (>30) the violation of the normality assumption should not affect the outcomes in any major way (Pallant, 2007; Verma, 2015). Since, we have
2x3 mixed factorial design
C O O f it Service quality
33 a large enough sample size (>50) it remains possible to use parametric procedures such as the two-way mixed ANOVA (Pallant, 2007). However, due to the fact that we cannot assert with absolute certainty that the violation of normality will not affect our outcomes, a robustness check will be conducted wherein the dependent variables will be transformed to reduce skewness and provide a better approximation of a normal distribution (Field, 2013). The Box’s M test was non-significant(p>.001), therefore the assumption of homogeneity of variance – covariance was satisfied. The Levene’s statistic was non-significant (P>.05) in each COO fit
group (within subjects), hence the assumption of homogeneity of variance was satisfied. Lastly, the assumption of sphericity is met since the repeated measures variable has only two
levels (e.g. high and low COO fit) (Field, 2013).
5.2 Testing main effects
5.2.1 Hypothesis 1 H1 posited that high COO fit combinations (HH/HL/HN) will receive higher brand alliance attitude ratings compared to low COO fit combinations (LH/LL/LN). Since the F value for COO fit was significant (F(1,148)=35.098, P≤.05), the null hypothesis that the average attitude
score was the same in the two COO fit groups irrespective of service quality categories was rejected at the 5% level. The pairwise comparison among different COO fit groups irrespective of the service quality was done by using the Bonferroni correction. The results indicated that the subjects evaluated the high COO fit country (Sweden) more positively (M=5.50, SD=0.88) than the low COO fit country (Indonesia) (M=5.04, SD=1.07) irrespective of the service quality categories of the between-subject variable (Figure 5.1). Therefore, H1 is supported.
5.2.2 Hypothesis 2
34 combination (HN-LN), and finally by the low service quality level combination (HL-LL). In this case we choose to use the ‘no service quality’ cue condition as a substitute for the ‘in-between’ level in our analysis.
Since the F value for Service quality was marginally significant (F(2,148)=2.456, p≤.10), the null
hypothesis of no difference of mean attitude scores among different service quality categories irrespective of COO fit was partially rejected at the 10% level (Figure 5.2).
Based on this marginally significant outcome we conducted additional comparisons between the different levels of service quality by means of a simple contrast analysis in order to assess the significance of individual comparisons. Contrasts revealed that brand alliances with high service quality cues (M=5.47, SD=0.12) were significantly more desirable than brand alliances with low service quality cues (p≤.05) (M=5.09, SD=0.12). Yet, both high and low service quality cues were not significantly more (or less) desirable than brand alliances with no service quality cues (p>.05) (M=5,26, SD=0.12). Hence, with the exception of ´no quality cues´ the mean attitude difference of service quality was significant at the 0.05 level. Hence, H2 was supported in as much as students’ brand alliance attitude ratings with respect to the different service quality treatments fell in the hypothesized order (i.e. from high to low), though not every difference was found to be statistically significant.
5.3 Testing interaction effects
5.3.1 Hypothesis 3
35 did observe that the mean differences between the combinations with additional service quality cues were lower (diff=0.48 with high cues and diff=0.39 with low cues) compared to the combinations made without any service quality cues (diff=0.51). However, these mean differences were insufficient to support hypothesis 3 since the F value for the interaction (COO fit * Service quality) was not significant (P>.05). Therefore, H3 is not supported since no significant interaction existed between COO fit and service quality (Figure 5.3).
Figure 5.3 Visualisation of the interaction between COO fit and service quality cues
36
5.4 Robustness of results
5.4.1. Robustness check for normality
Since the assumption of normality was violated, a robustness check was conducted to reduce skewness and provide a better approximation of a normal distribution. Evidence of negative skewed distribution was obtained, hence, to correct for distributional problems we decided to transform the data. Through a process of trial and error as recommended by Field (2013) we decided to use a log transformation to correct for the negatively skewed distribution in each of the 6 groups (HH, HL, HN, LH, LL, LN). The Shapiro-Wilk statistic was non-significant for four groups consisting of low and no quality cues (HL, HN, LL, LN) (P>.05), but still remained significant for the two remaining groups with high quality cues (HH and LH) (P≤.05.). This suggests that the log transformed data did not provide a full normal distribution among all six groups. Nevertheless, it was able to reduce skewness and provided a better approximation of a normal distribution (Appendix 8).
37
5.4.2 Robustness check for random assignment
To check for the robustness of our assumption of random assignment we included four known extraneous variables, namely, two demographic constructs (i.e. age and gender) and two socio-psychological constructs (i.e. consumer ethnocentrism, consumer cosmopolitism) as outlined in section 4.6.
A chi-square test was performed to analyse if the frequency of males and females (i.e. gender) differs between the different levels of the between-subjects variable service quality. The relation between these variables was non-significant (𝑋2(1,N=151)=2.657, P>.05), there
was no significant relationship between the different categories of service quality and the frequency of males and females in the given sample.
38
Figure 5.4 Mean differences of consumer ethnocentrism among the different treatment groups
5.5 Overview of findings
Based on the empirical results we can provide the following conclusions for each of the suggested hypotheses (table 5.2). H1 was supported at the 5% significance level indicating that COO fit had a notable impact on student’s cross-border brand alliance attitudes. H2 was marginally supported in as much as that the brand alliance attitude ratings of the service quality categories all fell in the hypothesized order. Though, based on additional comparisons between the service quality categories (e.g. high, low and no quality cues) it can be concluded that, with the exception of the ´no quality cue´ category, the mean attitude difference of service quality was significant at the 0.05 level. Lastly, no significant evidence was found in support of H3, additional service quality cues do not counterbalance the COO fit effect as evident in H1.
39 cannot be assumed. Even when transforming the data two of the six groups remained subject to the violation of normality. Regarding the check for random assignment we can reason that results were quite robust since three of the four known extraneous variables were non-significant meaning that each experimental group was comparable with respect to the known extraneous variables. However, ethnocentrism was found to be significant (p≤.05) suggesting that notable differences exist among the three treatment groups. In this case, providing additional service quality cues has dampened a student’s response on the ethnocentrism scale. Hence, these treatments might prompted them to think about their ethnocentrism response.
Table 5.2. Overview of study’s outcomes
Hypotheses Results
H1: Country of origin fit is positively related to consumers’ attitudes about a cross-border service brand alliance.
Supported
H2: The level of service quality of the alliance partner positively influences consumers’ attitudes towards the cross-border service brand alliance.
Partially supported
H3: If additional quality cues are provided to a student, the impact of COO fit on consumers’ cross-border service brand alliance attitude will diminish.
40
6. Discussion
The purpose of this study was to empirically test an under-examined determinant of students’ attitudes (i.e. COO fit) and its potential interaction effect with other service attributes (i.e. service quality) in the novel service context of cross-border brand alliances in HE. Through delineating the mental processes behind the student attitude formations with regard to cross-border branding alliances in HE our study makes a modest but useful contribution to both brand alliance research and managerial practice.
6.1 Theoretical implications
The contribution of this study to brand alliance research is fourfold.
First, prior studies on the role of COO in brand alliances primarily focused on commercial product-based sectors (Bluemelhuber et al., 2007; Lee et al., 2013; Li & He, 2013). Our results are consistent with the findings of Bluemelhuber et al. (2007) that show that COO fit has a significant impact on consumers’ brand alliance attitudes. This result signifies the importance of acknowledging secondary brand associations in consumer and brand alliance research. Additionally, our study responded to the worldwide trend of internationalization in the HE sector and provides further understanding of brand alliances effects in international HE sectors from a student’s perspective (Hemsley-Brown & Goonawardana, 2007). All in all, this study contributed to the general brand alliance literature by suggesting the transferability of specific brand alliance effects regarding the role of COO reported in product-based sectors, to the not-for-profit and novel service context of HE (Dickinson & Barker, 2007; Naidoo & Hollenbeek, 2016).
41 be concluded that, with the exception of the ´no quality cue´ category, the mean attitude difference of service quality was significant at the 0.05 level. Within this analysis we choose to use the ‘no service quality’ cue condition as a substitute for the ‘in-between’ service quality level in our analysis. Hence, the insignificance of differences with regard to the ´no service quality cue´ condition might be due to potential differences between our chosen substitute and actual in-between level service quality cue conditions. Therefore, further research should use actual ‘in-between level service quality cues’ to test the robustness of our findings.
42 This suggests that in service contexts, in contrast with the ELM and hierarchy of brand associations, secondary brand associations are of similar importance as primary brand associations. Another possible explanation is that service quality is a much less defined concept in relation to HE, specifically for our chosen sample which are students who are in the start of their university careers. Zimitat (2008, p.1) in his study on “student perceptions of the internationalisation of the undergraduate curriculum” found significant differences between students’ perceptions and experiences in their international orientations over the years. The results showed that first year students were significantly more positive in their international orientation compared to second- and third year students (Zimitat, 2008). Hence, it is interesting to do further research into what makes students at different levels of maturity in their study careers choose for certain double degree programs over others, in order to find out which primary brand associations are relevant in this setting. All in all, this study forms an new stance to a still unresolved debate in the literature by indicating that secondary and primary brand association might be of similar importance within a services context and raises new questions for future scholars to explore.
43 well as its potential interaction effects. Moreover, in line with Keller (2003) we highlight the importance for future research to not only try to provide a more inclusive view, but also attempt to uncover a holistic perspective on the brand attitude formation process by means of integrating a multitude of viewpoints from different disciplines (e.g. psychology, economics, marketing) and ultimately strive to “synthesize the multidimensionality of brand knowledge” (Keller, 2003, p.595).
6.2 Managerial implications
Next to the theoretical implications our findings suggest some possible practical implications for strategic management and international marketing in HE.
44 low acceptance, among others by its home students, due to its low perceived fit between both institutions (Fischer, 2012). In line with this example our results show that low-fit partners negatively influence students’ brand alliance attitudes, which is likely to result in low acceptance and consequently harms the alliance.
Despite the fact that students’ demands or perspectives have not been identified as the main priority among HEIs for the establishment of double degrees, we argue that our findings still provide valuable insights for contemporary HE management practice (Obst et al. 2011). Two of the top five contemporary challenges regarding double degree programs, reported by HEIs, in the international survey of the Institute of International Education(IIE), are ensuring sustainability and the recruitment of students (Obst et al., 2011). The fact that 29% of the participants reported that they had to discontinue some of their double degree programs due to the lack of student enrolment as one of the greatest causes, show that the student’s perspective does matter for the continuation of double degrees (Obst et al., 2011). Hence, we argue that, in order for a double degree to be successful, HE managers need to balance the relevancy of particular international partnerships with their own stated strategic and academic objectives, as well as with the expectations of the market with regard to various stakeholders such as students. Hence, our results suggest some additional attention points (i.e. COO fit and service quality) for HE managers to optimize their strategic decision making regarding international partner choice for the establishment of double-degree offerings.
45 salient challenges in brand management as forced by growing competition and global student preferences (Shocker, Srivastava, & Ruekert, 1994; Altbach, 2015). Knowing the student attitude formation process in relation to different primary and secondary services attributes can guide HE brand managers in developing more effective marketing communication strategies (Berentzen et al., 2008).
6.3 Limitations and future research
This section will show the potential limitations to respectively the study’s sample, research design, results, and research context, which should be addressed in future research.
First, although this study´s sample was carefully chosen to ensure familiarity and reduce potential bias across academic fields and consumer decision making stages, it inherently limited the generalizability of our findings. Hence, in the empirical experiment conducted data was collected from a Dutch university, however, differences may exist between respondents from different countries (Bluemelhuber et al., 2007; Kalafatis et al., 2016). Future research regarding HE cross-border service alliances should gather data in countries other than the Netherlands to facilitate the generalizability of our findings. Further, Zimitat (2008) found that internationalisation of the curriculum was most positively evaluated by ‘business and law’ students compared to students from other disciplines such as ‘arts and education’ and ‘health and sciences’. Accordingly, it would be relevant to carry out similar studies with respondents from other academic disciplines (i.e. other than business) to examine the potential transferability of our findings (Naidoo & Hollenbeek, 2016).
47 Third, as noted our results should be taken with care since even after conducting a robustness check by transforming the data, we could not provide a full normal distribution among all six groups. Hence, future research could conduct a replication study with even larger sample sizes to be completely assured that the issue of non-normality would not cause major problems. Further, based on the robustness check for random assignment we can reason that the random assignment technique used in this study was quite successful since three of the four extraneous variables were non-significant. However, the socio-psychological construct ethnocentrism was found to be significant (p≤.05) suggesting that notable differences exist among the three treatment groups. In this case, providing additional service quality cues has dampened student’s response on the ethnocentrism scale, indicating that these treatments might prompted them to think about their ethnocentrism response. Future research should add some unrelated cognitive tasks or time between testing the extraneous variables and the target stimuli in order to avoid potential influences of the various treatments (Bluemelhuber et al., 2007). Finally, given the unique history of Indonesia as a former Dutch colony (Dutch East Indies), results might be biased by using the Netherlands and Indonesia as the low COO-fit country pair in the experiment. Accordingly, future research should be sensitive to their study’s context when developing their own COO fit by choosing countries that are as ‘neutral’ possible regarding the chosen home country’s history.
49
References
Aaker, D. A., & Keller, K. L. (1990). Consumer evaluations of brand extensions. The Journal of Marketing, 54(1), 27-41.
Aerden, A., & Reczulska, H. (2010). The recognition of qualifications awarded by joint programmes: An analysis of current practices by national recognition centres. European Consortium for Accreditation in Higher Education. [Accessed: 16 September 2018]. Retrieved from: http://ecahe.eu/w/images/4/47/Eca-Publication---the-recognition-of-qualifications-awarded-by-joint-programmes---2010.pdf
Agrawal, J., & Kamakura, W. A. (1999). Country of origin: A competitive advantage?. International Journal of Research in Marketing, 16(4), 255-267.
Altbach, P. G. (2013). The International Imperative in Higher Education. Rotterdam: Sense Publishers.
Altbach, P. G. (2015). Perspectives on internationalizing higher education. International Higher Education, 27, 6-8.
Anderson, N. H. (1981). Foundations of information integration theory. New York: Academic Press
Anderson, J. R. (1983). A spreading activation theory of memory. Journal of verbal learning and verbal behavior, 22(3), 261-295.
Becker-Olsen, K. L., & Hill, R. P. (2006). The impact of sponsor fit on brand equity: The case of nonprofit service providers. Journal of service research, 9(1), 73-83.
Berentzen, J. B., Backhaus, C., Michaelis, M., Blut, M., & Ahlert, D. (2008). Does “made in…” also apply to services? An empirical assessment of the country-of-origin effect in service settings. Journal of Relationship Marketing, 7(4), 391-405.
Bleijerveld, J. F., Gremler, D. D., & Lemmink, J. G. (2015). Service alliances between unequals: the apple does not fall far from the better tree. Journal of Service Management, 26(5), 807-822.
Bluemelhuber, C., Carter, L. L., & Lambe, C. J. (2007). Extending the view of brand alliance effects: an integrative examination of the role of country of origin. International Marketing Review, 24(4), 427-443.
Bourdeau, B.L., Cronin, J.J. Jr & Voorhees, C.M. (2007). Modeling service alliances: an exploratory investigation of spillover effects in service partnerships. Strategic Management Journal, 28(6), pp. 609-622.
Bourdieu, P. (1988). Homo academicus. Cambridge: Polity Press in association with Basil Blackwell.
Cheung, A., Yuen, T., Yuen, C., & Cheng, Y. C. (2010). Promoting Hong Kong’s higher education to Asian markets. International Journal of Educational Management, 24(5), 427– 447.
