.xa' any}. £8;er1. hfftfirflxiflls , av ; . any,” e(..d:.lrf..fl. 2.1 13.1. .. . J . a. 5.931.... {3... . £93 a . itsiaflafl a: 3.0 3.! to!!! e!!!) . .3,qu 43.5.)qu v it’flvhtuPIVu. 23.13éflisblx. D. Hilfié «1 l. +. .. lawn. .iwiiiitfuuxia L Edda... a .{n {.65.}. 3! .. 4.1.2.32? M flaWWxHMWu-‘uruWaomfltH .11.} .Pt I. . ‘ . 2.; - Q - . t 9% .31..|&.W~Luuv1.HIHW.NW ‘ ammuvnulshlu‘luliqmFZu énuud. film”. 5.6:! . . 4 ‘ .3- l;.....$! , 9 .I34 wflduhhflmflu! c ‘ , 4 . n)... TEEQS ’\ XUCN LIBRARY Michigan State U nlversity PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE M31 42 m URN U 6 7504‘ Ly1%r&n§ 02.2304 JUNZBZUU JUL 1 2 2005 6/01 c:/ClRC/DateDuo.p65-p.15 MODELING THE DIFFERENTIAL EFFECTS OF BRAND IMAGE AND COUNTRY-OF-MANUFACTURE ON KOREAN CONSUMERS’ PURCHASING BEHAVIOR OF HYBRID PRODUCTS By J ae-Eun Chung A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements For the degree of DOCTOR OF PHILOSOPHY Department of Human Environment and Design 2002 ABSTRACT MODELING THE DIFFERENTIAL EFFECTS OF BRAND IMAGE AND COUNTRY-OF-MANUFACTURE ON KOREAN CONSUMERS’ PURCHASING BEHAVIOR OF HYBRID PRODUCTS By Jae-Bun Chung - ' ' rtance of A comprehensive model was proposed to explaln the relative ImPO nation, purchase quality country-of-manufacture (COM) and brand image on product eval the basis of selected attitudes, and purchase intention for hybrid products on clinical P‘eSfige)' dimensions (design, performance, serviceability, brand prestige, and te d . and bran The current study also proposed models of the differential effects Of COM image on the quality perception of hybrid products using brand and COM fammafifies as moderating variables. Based on information gathered from focus group interviews With Koreans, quantitative survey instruments were designed including three products, Canon cameras made in Japan versus China, Lucky Gold Star TVs made in S. Korea versus Malaysia, and Ralph Lauren sweaters made in the 115- versus MCXiCO' Data were collected in Korea from consumers during June and JUJY 1997' Four hundred and fifly Six ”Seable questionnaires were collected, yielding an 82 percent response rate. The results of structural equation modeling (EQS 5'1) indicated that regarding TVS and cameras, brand image was more important than COM in most decision-making processes for Korean consumers. Moving production facilities to developing countries 3 however, was found to damage performance evaluation, prestigious brand and technical Images, and attitudes toward purchasing the product. For sweaters the brand ' e ’ Imag strongly influenced Korean consumers’ product evaluations, purchase attitudes and purchase intention, while COM had little influence on these processes. The current study failed t0 test the moderating effects of brand and COM familiarities on the relationships among COM, brand image and quality perception. Because the brands employed in this study were well known to Korean consumers, we were unable to identify a group of low brand familiar COnSumers within the sample. ided. International Based on these findings, managerial implications were prov ping managers in the electronic industries should be cautious about sourcing from develo aPPaIe‘ countries even for companies with favorable brand imageS. For international g‘fificm‘t for managers, on the other hand, country-sourcing considerations become less 81 . . . ' 3 {OT leltatlons of the study and recommendatlon strong and favorable apparel brands. future studies were provided. Copyright by JIUDJUUPJCHHIHVC} 2002 To My Husband, No-Wook ACKNOWLEDGEMENTS I would like to express my great appreciation to my advisor, Dr. Dawn Pysarchik, for her support and encouragement. I am very grateful to Dr. Brenda Stemquist, Dr. Patricia Huddleston, and Dr. Roger Calantone for their willingness to serve on my committee and valuable inputs. Dr. Linda Good deserves my great appreciation for her thoughtful advice and encouragement. I would also like to thank my parents and family in Korea for their love. Last but not least, I want to acknowledge my husband No-Wook Park, daughter Ariella, and son Zachary for their love and understanding. vi TABLE OF CONTENTS LIST OF TABLES ................................................................................. x LIST OF FIGURES ................................................................................ xi CHAPTER I INTRODUCTION .................................................................................. 1 Statement of the Problem ..................................................................... 2 Research Objectives ............................................................................ 5 CHAPTER 11 REVIEW OF LITERATURE ..................................................................... 6 Country Image and Quality Perception ..................................................... 6 C00 and Overall Quality Perception ....................................................... 7 C00 and Overall Quality Perception of Uninational Products: Single Cue Situations .................................................................. 7 C00 and Overall Quality Perception of Uninational Products: Multi- Cue Situations .................................................................. 8 C00 and Overall Quality Perception: Hybrid Products ................... 9 C00 and Multi-Attributes or Multi-Dimensions of Quality ............................ 16 C00 and Multi-Attributes or Multi-dimensions of Quality: Country- Image Dimensions and Quality Dimensions .............................. 17 C00 and Multi-Attributes or Multi-Dimensions of Quality of Uninational Products: Single-Cue Situations ............................ 25 C00 and Multi-Attributes or Multi-Dimensions of Quality of Uninational Products: Multiple-Cue Situations .......................... 29 C00, Brand, and Multi-Attributes or Multi-Dimensions of Quality; Hybrid products ............................................................... 3] C00 and Purchase Intention ............................................................... 38 Product Familiarity ........................................................................... 40 CHAPTER III CONCEPTUAL FRAMEWORK ............................................................... 44 Proposed Quality Dimensions across Product Categories .............................. 45 Quality Dimensions ............................................................ 45 Classification of Quality Dimensions Based on Quality Evaluation Mechanisms ................................................................... 48 Comprehensive Model of the Differential Effects of Brand Image and COM on Consumer Purchasing Behavior .......................................................... 50 vii The Relationships among COM, Brand Image, and Quality Dimensions .................................................................... 53 The Relationships among COM, Brand Image, Quality Dimensions, and Overall Quality Perceptions ............................................ 58 The Relationships among COM, Brand Image, Quality Perception, and Purchase Attitudes ....................................................... 59 The Relationships among COM, Brand Image, Purchase Attitudes, and Intention to Buy ......................................................... 61 Models of the Differential Effects of COM and Brand Image on Quality Perceptions of Hybrid Products Using Brand and COM Familiarities as Moderating Variables ...................................................................... 63 Definitions of Brand Familiarity and COM Familiarity ................... 63 Moderating Effects of Brand Familiarity and COM Familiarity on the Relationships among Brand Image, COM, Quality Dimensions and Overall Quality Perception ............................................. 65 CHAPTER IV RESEARCH METHODS ........................................................................ 72 Research Design .............................................................................. 72 Sample Selection and Data Collection: Empirical Study ............................... 73 Instrument ..................................................................................... 75 Data Analysis .................................................................................. 80 CHAPTER V RESULTS AND DISCUSSION ................................................................ 82 Demographic Characteristics of Sample ................................................... 82 Reliability Tests and Confirmatory Factor Analysis ..................................... 83 Testing the Comprehensive Models of the Differential Effects of Brand Image And COM on Consumer Purchasing Behavior ......................................... 86 Overall Model Fit ............................................................... 86 Chi-Square Difference Tests ................................................... 87 Hypotheses Tests ................................................................ 33 Testing the Moderating Effects of Brand and COM Familiarities on the Relationships among COM, Brand Image, and Quality Perception of Hybrid Products ...................................................................................... 99 Classification of Four Groups based on Brand and COM Familiarities ...................................................................... 99 Discussion .................................................................................... 1 01 CHAPTER VI SUMMARY AND CONCLUSIONS ......................................................... 105 Summary of Research Objectives and Past Literature ................................. 105 Summary of Conceptual Model ........................................................... 106 Summm of Research Methods ........................................................... 109 Sumnmry of Empirical Results ............................................................ 110 viii t Managerial Implications .................................................................... 114 Limitations and Recommendations ...................................................... 115 APPENDICES ................................................................................... 1 17 Appendix I. Questionnaire ................................................................ 1 18 Appendix II. Covariance Matrices for Confirmatory Factor Analysis .............. 131 Appendix III. Covariance Matrices for Structural Equation Model .................. 133 Appendix IV. Input SEM Models for EQS ............................................... 136 REFERENCES ................................................................................... 143 ix LIST OF TABLES Table 1. Summary of Product-Country-Image Literature: Hybrid Products ............ 10 Table 2. Quality Dimensions ................................................................... 18 Table 3 . Definitions of the Quality Dimensions and Equivalent Dimensions in C00 Studies ........................................................................... 46 Table 4. Definitions of the Quality Evaluation Mechanisms and Corresponding Quality Dimensions ................................................................... 49 Table 5. Descriptions of Questionnaire Formats ............................................. 77 Table 6. Salient Attributes and Relevant Quality Dimensions for Each Product ........ 79 Table 7. Sample Profile ......................................................................... 83 Table 8. Results of Reliability Tests ........................................................... 84 Table 9. Parameter Estimates for Confirmatory Factor Analysis .......................... 85 Table 10. Chi—Square Difference Tests for Cameras, TVs, and Sweaters ................ 87 Table ll-a. Results of SEM for Cameras ..................................................... 89 Table 1 1—b. Results of SEM for TVs .......................................................... 90 Table l 1-c. Results of SEM for a Sweaters ................................................... 91 Table l 2. Frequencies of Four Groups Based on Brand and Country F amiliarities. .. 100 Table 1 3. Results of Hypotheses Tests ....................................................... 112 Figure 1. Figure 2-a. Figure 2-b. Figure 2-c. Figure 2-d. Figure 3. LIST OF FIGURES A Comprehensive Model of the Differential Effects of Brand and COM on Korean ConsumerS’ Purchasing Intention for Hybrid Products. . . . 52 A Model of the Differential Effects of Brand and COM on Product Evaluations for Consumers with High Brand Familiarity and High COM Familiarity .............................................................. 68 A Model of the Differential Effects of Brand and COM on Product Evaluations for Consumers with High Brand Familiarity and Low COM Familiarity .............................................................. 69 A Model of the Differential Effects of Brand and COM on Product Evaluations for Consumers with Low Brand Familiarity and High COM Familiarity .............................................................. 70 A Model of the Differential Effects of Brand and COM on Product Evaluations for Consumers with Low Brand Familiarity and Low COM Familiarity .............................................................. 71 Summary of Significant Relationships by Product ........................... 92 xi CHAPTER I INTROD UCTION . . . - ' in South Korea has been recognized for Its economic expanSlon. Whleh was a shorter period of time than any other country in modern history (Holstein & Nakarmi’ 1995). Due to its strong economic growth, dynamic markets, industrialization, and stable government, the US. Departrnent of Commerce identified South Korea (hereafter referred to as “K0rea”) as one of the ten “Big Emerging Markets (BEMs) (Aguilar & Singer, 1996). For these reasons, the US. viewed S - Korea to be "vital to America's national interest" (Garten, 1997, P- XiX)- During the economic prosperity of the 19803 and early 19903, Koreans’ purchasing power increased dramatically and Koreans become passionate consumers of imported as well as domestic goods (Schuman, 1996). 9 . . . ' Korea s meteoric economic expansion, however, faced a breaking pou‘lt In early 1997, when the Asian financial crisis began (Anonymous, 1997). After experiencing a 5-8 percent decline in GDP in 1998, Korea is slowly recovering from the catastrOPhic 1997-98 economic cl‘iSis and reported GDP growth of 8.8 percent in 2000 and 2.7 percent in 200 1 (Bureau Of Economic and Business Affairs, 2002, pill-3)- Despite Korea’s current economic fluctuation, it is still a major US. export market. In 2001’ Korea was n 1998 (US. Census the eighth largest overall trade partner of the US, up from ninth i .8. Census Bureau, 2001: P-8)- Korea is the sixth largest market for US. exports (U the first Bureau, 2001’ P3) and the U-S- Provided twenty percent of Korea ’s imports for eight mOnths of 2001 (Bureau of Economic and Business Affairs, 2002, pp. 1-3). Wlth increasin trade between the U‘S ‘ and Kore ' ' ' g a, It Is Important for U- S, expat“? Wfio are targeting Korean markets. to understand Korean consumers’ purchasing behavior. In the past decade, With unprecedented technical and commanication advances. been firms have pursued global market expansion more than ever before. and also have able to create global brands. The global brand is no longer exclusively assoc1ated With one country since firms have taken advantage of lower wage rates by movmg their manufacturing or assembling locations to developing countries. They have eVen increased outsourcing of various components (or parts), designs and engineering tasks by collaborating with foreign partner firms or establiShil‘lg design centers overseas (Chao, 1998). Accordingly, these actiVitieS have contributed to the proliferation of country of origin (COO) designations, which now may include country of manufacturing (COM). country of design (COD), country of brand (COB), country of parts (COP) or country of component (COC), and country of assembly (COA). The traditional COO research paradigm, which typically assumed that a product was exclusively tied to one country, therefore, is changing in this new era 0f hybrid products. Thus, further studies are required to understand consumerS’ behaVIOI' regarding hybrid prOdUCts- This study, therefore, develops a model of Korean censumers’ purchasing behavior regarding hybrid products. The statement of the research problem is provided in the next section. Statement of the Problem . . . I L0 feld, 1993; EXIsting COO research has been criticized as being atheofetlca ( le Nebenzah], Jeffe, & Lampert, 1997; Saniiee, 1994). This criticism can be partly attributed to the lack of well-defined quality dimensions (Li & Dan't, 1997). Recent C00 studies reveal that produet evaluation iflvol ves three aspects: overall - qlla/It)’, pdel/Ct dunensrons (for example, perfomance or serviceability), and product attributes (for example, picture or SOund for TV) Many studies employed overall quality as a dependent variable (Liefeld, 1993). This construct, however, is too abstract to be “531111 in understanding the effect of C00 on quality perceptions. Examining the effect of C00 on product attributes, on the other hand, has its own deficienCY- It is ‘00 specific to construct a theory that can be generalized across different product categories. Therefore, several researchers have examined the effect of C00 011 product evaluation at the quality dimension level- (See Li & Dant, 1997.) These studies, however, lacked a set of Widely accepted general quality dimensions because these are not well defined (Thakor & Katsanis, 1 997; Li & Dant, 1997). Thus, a determination of quality dimensions that couldbe used across product classes is needed. /ln addition, COO researchers have identified that research on the hybrid product phenomenon is lacking (Johansson, 1989; Obermiller & Spangenberg, 1989; Samiee, 1994). For example, the relative importance of brand image and the various components 0f C00 in consumers’ decision making is still in question. This issue is especially important for global PTOductS because their established brand reputation could be damaged by moving their production location to places that are perceived unfavorably. The existing COO literature indicates mixed results regarding this issue- In some studies, ' ' , . 0 es It IS argued that brand “31116 might not compeHSate for the negative Image Of CO (N . dies, & Bilkey, 1993; HeSIOp, Liefeld, & Wall 1987; T36 & Gom, 1993). In other stll brand name has been found to be a more important predictor of perceived quafity and 1 9 - purchase value than C00 (Chao 89’ TSe & Lee, 1993; Ulgado & Lee, 1993)- law; more research is required to inforrrl the discussion ofthis concept / /r'(llrth,er product familiarity has been identified as an important influence on the relationship between C00 and Product evaluation. (See Johansson,1989 and Samiee, 1994 for a reView of relevant studies.) In the previous COO stUdleS product familiarity is defined as the “number of product—related experiences that have been accumulated by the consumer” (Alba & Hutchinson, 1987, p. 411) and as “the subjects’ perceived knowledge with respect to the brands in a product class” (Park & Lessig, 1981, p. 223). Such product familiarity definitions are too vague to indicate whether or not product familiarity is the same as brand familiarity. Furthermore, in the case of hybrid products, COM is not the same as COB, where the company’ S headquarters are located. Thus, brand familiarity does not necessarily mean familiarity With the country where the Therefore, the concept of product familiarity should be / product is manufactured. redefined and reexamined in the context of hybrid products, Finally, existing COO researchers have failed to provide a comprehensive VleW Of the effect 0f COO on the consumer’s decision making, That is, product evaluations or beliefs about product quality (Brodowsky, 1996', Liefeld, 1993, Samree, 1994) are the n the information-processing most frequently studied as dependent variables based 0 1989, 1990', perSpective (Erickson, JOhansson, & Chao, 1984; Hong & Wyer Jr. d the C00 effect on Obermiller & Spangenberg, 1989). Aese studies have emphasze & Wyer, 1984; Hong overall product evaluation as a cognitive process (Erickson et 81-, Jr,, 1989, 1990). In consumer behaVior research, however, there have been arguments about the overall evaluation of quality by consumers (Compeau, Grewal & many CO Cha he\ \_/‘ Monroe, 1998; Erickson et al-, 1 984; Jo hansson, Douglas, & Nona/(a, 1985,- law 33 Monroe, 1989). For exmple, a recent sde by Compeau et a]. (1998) indicates that quality evaluation has not only a cognitive component but also an affective component. With a few exceptionsi however, COO researchers have not attempted to define the construct of overall quality perception and overlooked the affective component (Erickson et al., 1984; Johansson et al., 1985; Obermiller & Spangenberg, 1989). Accordingly, a more comprehensive model of the effect of C00 on quality is needed. Research Objectives To address some of the unresolved issues in the existing literature, the Objectives of this study are to develop a model of Korean consumers’ behavior regarding l) the quality dimensions of product evaluation; 2) the effects of comm}, of manufacture (COM) and brand image on product evaluation (cognitive component), product—specific purchase attitudes (affective component), and purchase intention for hybrid pdeUCts (behaVioral °°mp0nent); and 3) the effects of brand and COM familiarity on the relationships among COM, brand image, and product evaluation of hybrid PmdUCtS- The remainder 0f the dissertation is organized as follows. Chapter 2 provides a comprehensive review of the C00, product quality, and product evaluation literature. In Chapter 3, the pr0posed theoretical models are discussed. Research design, Cl‘k‘vStionn'alil’e development, and data COllection are provided in Chapter 4, while Chapter 5 contains the reseamh findings“ Finally, in Chapter 6, the dissertation is summarized and “Phantom of the study are discussed as well as directions for future research. CHAPTER II REVIEW OF LITERATURE In this chapter, the literature is reviewed relevant to the research objects: country image, brand, quality perception, country of origin (COO), multi-attributeS/multi- dimensions of quality, purchase intention, and product familiarity. The literature I'eView is organized on the basis of studies of uninational products or hybrid products, and fimher classified on the effect of single or multiple cues on COO. More emphasis is placed on the previous studies of hybrid products since the objectives of the current Study involve the effect of country 0f manufacture (COM) for hybrid products Country Image and Quality Perception One Of the early Studies to examine country image and product perception was Nagashima’s (1970) cross-cultural survey of the image of “made in” labels among US. and Japanese businessmen, In this study, country image was defined as follows: The picture, the reputation, the stereotype that businessmen and consumers attach to products of a specific country. This image is created by .S‘mh Variables as representatwe products , national characteristics, economlc and Pelltlcal background, hIStory, and traditions (Nagashima, 1970, [3-68)- for any particular Similarly, according to Narayana (1981) “the aggregate image d with that countIY’ s country’ 5 product refers to the entire eonnotative file associate Thus the concept of country 13“)th offerings, as perceived by consumers” (p. 32). image in the ‘70s and early ‘803 was broad and general. COO researchers, however, began to narrow the concept of country image to a marketing perSpective. That is, they Viewed country image as the consumers’ general uali rce ti ns of I'Odmts made in - - . q ty pe p 0 P the country m question (BI/key & Nes, 1982; Han, 1989). Roth and Romeo (1992) defined country image by elaborating on the marketing perspective as follows : The overall perception consumers form of products from a particular country, based on their prior perceptions of the country’s production and marketing strengths and weaknesses (p. 480). Thus, based on this definition, country image is identical to the consumer’s overall evaluation of the quality of products made in a given country, and country-image dimensions are equivalent to quality dimensions. C00 and Overall Quality Perception Product quality, as a dependent variable, has been well researched in studies of uninational products as well as of hybrid products. As previously mentioned product quality can be measured at three levels: overall quality, product dimensions, and product attributes. In this section, studies of the effect of C00 on Overall quality perception in single-cue 311d multiple-cue situations for uninational and hybrid PTOd‘mts will be reviewed. C00 and Overall uali Perce tion of Uninational Product51 Sin le-Cue W - . d overall The Slmplest studies have been those in which respondents rate the dOWSky, n a single cue—COO (BIO quality of products made in a single country based 0 ositive Effect for 1996). Most of these studies found that generally COO had a P domestically-made products and a negative effect for foreign-made pI'OdUCtS 1965; Reierson, 1967; Schooler, 1971; Wall & Heslop, 1986). For instance, We“ and HeSIOP (1936) found that, in general, Canadian consumers rated Canadzan-made proa’acts (cars, wine, clothing, footwear, and home electronics) significantly higher than those made in foreign countries such as Brazfl, Taiwan, Hong Kong or fiom some European countries such as Italy, Sweden, France, USSR, Czech, and PRC. Researchers also found that the perceived quality of products made in developed countries was significantly higher than that of products made in developing countries (Gaedeke, 1973; Reierson, 1967; Wall & Heslop, 1986). For example, Gaedeke (1973) found that American consumers rated the quality of several American-made prOducts significantly higher than that of preducts made in less developed countries such as Brazil Taiwan, Argentina, Mexico, S. Korea, India, and Turkey. Single-cue COO studies, however, have been criticized because the purchasing context is not realistic. In fact, consumers are considering many intrinsic cues (such as 39/133 color, and texture) and extrinsic cues (such as brand or price) when they purchase a product. The C00 Studies with multiple cues for uninational products will be reviewed in the following section. C00 and OVerall nan Perception of Uan—Z—Mcfl Situations [— Multiple-cue StUdies are those in WhiCh respondents are asked to rate products - ' t multi le-cue based on the C00 cue as Well as other Product information'. The earhes p Stigated whether or study was done by SChooler and Wildt (1 968) in the US. They inve ”f 2} “WY manipulated product brand names as well-known versus unknown names along with the manipulation of C00 information, ““5 “’35 eOusidered to be a study of hybrid producm although the researchers did "ct identify these products as hybrid products. ’ m not a negative country image could be com ' ' pensated for b ricm . yp g a fbrelgnpfodl/ctata lower cost than a domestic product. Fu-St, respondents were asked to rate the quality of glassware based only on the “made-in” label (Japan or the United States). The mean evaluations of US. products were significantly higher than those of the Japanese products. Price Was then offered as a second cue. Respondents were given the same Price for the American product, but they were divided into six groups in which different pricing information for the Japanese product was given- When respondents were asked to rate the quality of these products, the quality ratings were not different across groups, This indicated that price had no effect on quality evaluation. C00 and Overall @ality Perception: Hybrid Products Major Studies of C00 on hybrid products are summarized in Table 1. Table 1 shows that most of these studies used multiple cues -- at least two, country and brand -- to provide product information to respondents. Since one of the interests of this study is to identify the relative importance of brand and COO for hybrid products, the effects of brand as We“ as those 0f COO are discussed. As shown in Table 1, several studies of hybrid products examined overall quality percepti on as a dependent Variable. Most of these studies measured overall quality by a ' ' . - t 1., smgle item (Ahmed & d’Astous, 1996; Golden, 1992; Iyer & Kahta, 1997, Heslop e a s discussed 1937; N68 & Bilkey. 1993; Ulgado & Lee, 1993). Each of these studies 1 below. as: :0 0050.0 u :00 8.0.00 :0 .9500 n 000 ”0:300:05: :0 .9500 u .200 ”5.000 :0 9:000 u Goo 3:05:50”: :0 9:000 u 000 .055 :0 .9500 u moo $380004 :0 .9500 u <00 :i 5:03 .0008: .0200: 00000.: 000.: £0 a 3 “000.500 .0226 .00: .3553 :05 :300 -085 :0 085.93 000::0 80 3:055:03 05 0>0E8 .0: 0.0 005.0050 .0: 60:05:08 .05 00:..000 000:0 050: 00.0 :0 0000000.0:00::< .05 05 000:..0 000 0:00: 53 05: 055 05 000 :00300: 00005005 35:00:00: .0008: 05 20:5 .50....0w.0 02 00.50.56 000000: 02:00:00 0. 00.00: 5.50:0 0.3505 mm .H: 00.0.0.3 5:000 0:00m 055 0:05 a 0: 0. 005000 000 05 05:8 3.0000 0003 05: 05.: 05 000 5:0. .m.: 00.05 ”00.00.05 02... 000 £00: 500 00 00h .30. 53 0.0.: 05 :0:3 5:. :3: 53 0.0.... 850:0:00: 00000:: 0:. 5:3 00.00:» 53 Euro :0 9:000 00:20.60 000. a 5.3 0.00: 00.0.0005 5.50.0.6 00000:: 0.3050: 05 £00.00: 0. 000:0 05 :0 050 00.0: 05 :. 0.5 00: 0:3 000 00: £053 05 :0 050 05 :. 555: 53 0. 5:3 >550“: .003 00.": 000—5 8:20:2: 050m 55 E53: 53 05:: 05 :0:3 .050.» 53 9:000 :0 .00l::0 0:... 5.00.0: .m.D :053 beguego 9:000 300: .3080 .05: 05.: 55 9:000 5.00000 05 0. 030.050 0.0.: 203 :25—5003 05 9.300030% 058:0. 00.0:0E.0-000000: b.0000 =Eo>o 0.03 300::0 000 0:00: 53 3:000:20. 0:00:00 05 00.000 00.00. .5008m. and :0 050 05 0. .05: 05.: :0 005 55 005000 5.3 9:000 000000 :5500 3:530:25 :0 000::0 0:. $5050-05 05 5:07.30 :0 00000 05 :. .0050 00.0.: 00.00. 500m 3005:0003 :. 0:00: 53 5.00803. 05053.0 $0000.: 50: .0... 00.30.0005 05 55.. on .u: 00.305200. 00000.: 055 8:0: 0?: ._0: 05: 055 05 .200 00.: 000% 803 $0055 :02: 050555 .m.: .m.D :0.m_>0_0:. 0005>0< :00... 9:000 000.90: 00 5: .000: :0 05.5 53.0.: .3 :0: 0050:0908 0: .0: Ewe: 092:. .9500 03.50: 0 .05.: 05 :03. wcoz .000: .000 m:05 5.00202. 05055.0 0: :0 0.000.. 05 :0 005m .0000 00.00. 500m 0:0::0.0... 00...: 05B 05 00.0: :0 00:000.: 05 :. 0058000 000 000 05 :0 85:09:. .m.D one“: 00:03 055 $3.: :03 05 05 00008.0 3:0 000: 000.030.... 05 :. 00.00.» 53 80:0 000 00050 00050 3.5 £050 .053 00000.: 000 00 0.0.3... $0.00: 00:_::_..:: 00.00.05 008...... 00.qu :0 09mm: .AEocooo .050: 05.5 :0 000:8. 03.50: 0 05: 0.003 .9500 090:... 30. :23. .m 05 00.50 .m:0_m:0E.0 a 8 0000000.: m:.>0E .05: 00:00 05 :0 $5050 0003 8 0000000.: 55.. 03. 5.0.0 055. 0030:. .3 038:5. 203 >030 0.5 0. 0.05.: :0 :0 «own—E 0:. 05 50 0 0:00:00 .003 o~muz 9008:0500. 055 055 6.33 2500002 0:000:55: 3 :0.:3 :. .9500 .00: 0:. 0.0 00.00.20 53 05:50 .003 .m.D .m.D 50 00:80:05.5.r0n 9:000 00 085:0. 80000...— 80035 «0.00950 0.00.5.» 30.00...— :0002 00.002 .0000...— .:0—.=2_0a 000 :uutseefiace 80:00.:— 0.0:»: "003000.: 09:07.33000 $00000: :0 b.2555 .— 030,—. 10 A‘ hurlx .0: 0 0:55 a .. .~ 0.000... 0.550 u 000 M...5.n....=_0.2 .0 00.80 n 200 0.0.80 .0 5.80 n :8 0008:0080 0.0000 n 000 0500.0 0.50 n 000 05: :0 0.00000 n :00 .0.:8000< :0 0.00000 u <00 .000 05 <00 0: 0032.0: 0.0002: 02:. 0:. :0 =0 .0: 0003 05:20: 80...... 0. 0.3.0.0 .5..0:8. .008 0:. 003 005.0000 :0..00:0_.0m .02. 000. 0: 0. 00.00::0 00.0>0. 0... .0: .9008 0003000.: 000. 0 0. 00380000 85.0000 .0 00:».000 003 .. :. 00002000 0:08. 00.00.. 0.05.: < 00:08.20. 000.02 00..00:0..0m 00.00.8002: 0:00.000 w:.0.>0.: 0: 00.00::0 003 000 .0 05.: 00.00. 0.00m 8:: 0 :0 000530.05: 0: H 000000.05 .0002: 00.0.0800 0. 0.0. .088 0.0.. 00:m 05... 0 0000.: 8...: 0.00008 0:00.. 0:. .0: 05: 05.: 0:0 <00 0: 0032.0: 5:0— memu: M0> 00.0> 0038.0: 000 $00.. 85.5.. 008800 :0 :0...0:2: .0005. 0:. 0080.98 000 40.000» 0. 00050 00050 0..:080.0< 00.00.. .055 <00 000.0<.0 .0 008:< N 000.0 00.88.00 00.00.. :0 .00:0 .5050»: 00.0.02 0. may": 0 00: 05.: 0.00 .200 05 05.: 00 :03 00 00>.» 003 00:00:20. 0.: . 000.0 000:0 0.00.00. .050 00:? 000:8.0: 00.00.. .0.:00:0:00. :0 .00:0 5.3.0:. 0. no": 0..0.:.< $.00 .30.: 80055.0 00: 0.0: 0.08.... 00 00...: 0.03 3.00 05 05.: 0.00 00:3 058.00 0.: >... 0.:000 =0.0>0 05.0. 00.. .0 00005 I 00.3.0.0 <00 05 000 083.0: 00.0. :0 000.05 0000230. 8005000 0: 003 0.0:... .00..:8.0: 00.00.. .0226 :0 0.00.00 - 0.:00: 00.000 .5050»... 00: :0 <00 05 000 8:: <00 0:0 000 0003.0: 0.00::0 53.0... 00.3.0.0 8t: 00:00.08. 0: 05 0.00::0 0.08 <00 .5008»? 0: 0.03 0.0:. .0: .0000: 5:00 0:.": .00: :0 000.03. 000 22,, 8008.2. 002... @000 .5 0.3.0 55 :00 05 82 50000.0 .00 0.: >0 . 0:2... .802: <00 .80: 8:0 80000000 000000.03 5.05 .55.. 0. 00380000 00.0.. 00000.0: 0:0 0:. 5:. 00.0.. 80:20: 08.8.0: .05.: 0 0>0: 0. 0:00: 003 00.00. N 000.0 080:5... .0000 8:000:00 50cm 0. 00380000 80.000 00:00 0 00000.: 0:. 5.3 0000.898 .002 0. 0...": 008:5... .0002: 0.3.0 8000:800 0:. :0 .0000 0>..0w0: 0 00.5.6 0. 0000: 53 05.: . 000.0 8.0.000: 05.0. .0:0.m 0305.-..03 < 00.030 000 8:50 0. 0:00: 000:. 5:. 000. 0.03 00.00. 500m 0. 02": 080.000 88:05 000 8.0.8 038800 05 .0000:800 0. 000 0.00hu 0:. 0000:8080 00:3 5:0 .. .m.D 00.0.0 85:00:.0: <00 20:. v 00.. .0 00 :. .0002: :000 0... .00 .0: 8088.030 80.0 :0: .05: 05.: 0.800000 :0 0.00. 0.00.0 .0 08.0302 .0: 0. 3:00. 05 0:08 5 .26. 052. a 03.: 203 .00.. :22 :00. 05 :00 .26: .02 es. 0.302: 30.5.8 02.22:. 5 use a 00.23 08020 55 .5202: .32 0.2880 .0. 0000 0.:00: .030. :0 05 80.0.. 0: 0. 00.3.8.0: 0.03 000.. 0. 0008 0.0002: w:..:000.:0. 00.0 000.0 0.0.0. :05. :000 0:.. w:.>0: 0.0002: 5.: 00.00.. .030. :0 05 8.0.0.. 00.00000 00.": .030... 0.. .00.. 5.30 05.0 $3.. 0: 0. 088.0: 0.03 000... 00.00000 00:0.0>00 000. 0. 0008 0.0002: 000.05 .m.D .:w::00.... :0: 02:00.0: 9:00P 00...... .0 002 _ 80:00.: 0.0035 00.89.00 0.:0..0> 000.00.: 000.2 0002 .0002: 80000:... ¥ :6? \ 50:05:30. flaw—~00 .H o—QQH ll 05: :0 0.8000 n :00 8.3.0 :0 0.8000 n 000 2300:0002 :0 0.8000 n 200 83000 :0 0.8000 n 000 80:0:800 :0 0.8000 n 000 .05.: :0 0.8000 n 000 .0.:8000< :0 0.8000 n <00 000.02 0. 00380000 0: 0. 030:0 003 .0002: 0:. 00:3 80820.0 000 03.03.: 0 0.688. 0. 0000 0: 0.000 05: .m.: .0:0.00.0.0 <00 0.003 0 08020.6 0. 0000 0: .0: 0.000 .00.: 0:. 80.: 05: 8.00 0: 08.0.8.0. 80.00.80 083...... 000 80.0 0 000.0:3 .0:. 000:8. :00 05 .000 .<00 :85 02:. :0 03:05 00000.08. 003.00.:. .50....030 0:0 ..00:0 0: 00: :00 05 <00 - 0.000.. 03000 .0: 8288.0: 3:00.. 03000 :0 .00::0 .50....030 0 00: 000 0:. .05: 008:...5 :00 .050 0:. :0 0.00::0 00000.0.:. 0:. :0 05 00 .00 .80:8. 0: 00: 000 000.02 00m": .00: :0 0080.00 000 05 000.82 0.00.... .802. =0 .082. 5.0.00.0 02 :00 .5 <00 .0... .00 >0 - 0.8... .802: <00 .00... 020 0008. 00.0.00... 858.0800 03.8. 05 00.0.00: 00020000 5.05:0. 00..0:.00< .0: 0:00: 0.03 000.: 05 000 0003.0: 0.00::0 00000.08. .50....03m 0.000000. 0.0.5.00 85:00:08 000 0.0.5.00 0.0.500. 8008.000: .0: 0.00 0000: :03. 0:0: 0.0.500... 0.03 05.: :0 0.00::0 0.08 .50....03m 0.0308050 .0: 88.8 0000: 0.0.. 0:0": 0.05 .0...:08.>.0m 000.0 0.03 000.8088 00.00.. 830 0:. :0 00.60 :0 80 :0 0.00::0 .50....03m .m.D .m... 00.0 0.002 00588:.0: 000 .30.. .50 05 E .00: 0. 00050....3. 8.0:0 05.: :0 80:8. .030. 0 00: 0000 8..: 80.58.08. 200 0:0 000 :0 80000.: 0:. 0. 0.0002: 0800 .0: 500.300.. :0 0.0>0. 30. .0 0000 00.0 0.03 0000 :00 0.05. 500.300.. :0 .0 0.000.. 00.000000 0. 0000.305. 888:8. 0.03 0000 200 0::3 .200 05 8.8 :0 0.0.0. =0 .0 00.3.0 .0:: 05.: 50:20... 5.3 0.0002: .0: 000:. 80.: 80.0::.0 0.03 00.3.0 00:08.5 0...: 200 05.: 0.: 5.3 0.0002: :0 00000020: 0.0000 00.00000 00.00.9000 0080 08.0.0 0. 000008.03 :00 .0 000:0...0000. 0.300 0. 0285:0008 0005 5:. 0.000.. 5:3: 00.0.. .0000 man: 0000. 00.0., 00300.0: .30. 0.. .30.. :0 0: 0. 08.8.0: 0.03 00...:000 00:20.00 0. 00.800858 0.0002: .md .m... 0.8.0000 0.000.. 5.30 :35 8..: 0.00.. 0... .00. 0.0000... 0.0035 8.8980 0.:0..0> 000.00.... 00002 00007. .0000... 8000095 000 50:38:80. 0.800 .. .30.. 12 Heslop, Liefeld and Wall (1987) compared Canadian views of the effect of C00 on product-quality evaluation in a single-cue versus a multiple-cue situation for three products (shirt, telephone, and wallet). In the single-cue treatment study, only the country information was manipulated (domestic [Canada], developed [U .S.], less developed [South Korea for shirt, Taiwan for wallet, Hong Kong for telephone]). On the other hand, in the multiple-cue situation, Canadian respondents were assigned to one of twelve treatment groups resulting from a two price (high/low) x 3 country (domestic [Canada], developed [U .S.], less developed [South Korea for shirt, Taiwan for wallet, Hong Kong for telephone]) x 2 brand (well-known/unknown) factorial design. Results of the single-cue situation indicated a significant effect of C00. Each of the three products from a less-developed country received a lower quality evaluation. For the shirt and telephone, Canadian respondents generally perceived Canadian products to have higher quality, but rated the quality of the Italian-made wallet higher than that of the Canadian-made. Results from the multiple-cue situation revealed no significant main effects (of C00, brand, and price) or interactions for the shirt. The overall model for quality perception of the wallet was insignificant, but the C00 main effect was Significant. In the case of the telephone, the overall model and the C00 main effect were Significant. Thus, the authors concluded that the C00 effect is greater in the single-cue Situation and that the importance of the C00 cue seems to decrease in the presence of price and brand cues. The authors also concluded, based on the result of no significant interaction among COO, price, and brand, that a negative country image might not be compensated for by brand or price. Ulgado and Lee (1993) criticized previous studies in which a significant effect of C00 was found because these studies did not consider consumers’ real purchase situations. They included only two cues, C00 and brand name, while much more information is available to consumers in reality. Thus, Ulgado and Lee (1993) conducted two studies. In the first, brand name and COM were the only information given about the product and in the second, product information on specific attributes was given in addition to the brand name and COM. They found that both brand and country information had significant effects on quality perception, when only these two stimuli were available to respondents. When they presented other intrinsic information, only brand had a significant effect on quality perception. These results imply that a well- known brand name can overcome the negative COM effect when other intrinsic attribute information is available. Similarly, Nes and Bilkey (1993) studied the effects of brand name and COO using tangible products due to the importance of intrinsic information to consumers’ product evaluation. They classified COO using three levels, low, medium, and high income based on each country’s national income. Brand names were classified into two categories, unknown and well-known. They found significant main effects of C00 and brand name on product quality and risk perception. That is, products made in less deVCIOped countries (LDC) were perceived to be riskier and of lower quality than PI‘OCluets made in developed countries. In addition, if the brand name was well known, it was perceived as being of higher quality and as less risky than if the brand name was unknOWn. According to the AN OVA results, brand name seemed to have a stronger effBCt than COO. Moreover, no significant interaction between COO and brand name 14 was found. In other words, if the products were made in LDCs, they were rated as being lower in quality and higher in risk than those from developed countries regardless of brand name. In the study by Cordell (1992), however, the interaction effect between C00 and brand name was significant. He found that preference for Timex, a well-known brand name, decreased only slightly when the product was made in Pakistan as opposed to West Germany. In the case of a watch with an unfamiliar name, Tempomax, preference declined sharply when the product was manufactured in Pakistan rather than in West Germany. Thus, this study indicates established global brands may override the COO effect, and unknown brands may improve consumers’ quality perception by moving production to developed countries. Ahmed and d’Astous ( 1996) examined the effect of a multidimensional formulation of C00, namely country of design (COD), where a product is designed or engineered, and country of assembly (COA), where a product is assembled, on consumers’ perceived quality and purchase value. Brand name and other pI‘OCmet information also were given. They found that the combined effects of COD and COA cues had a stronger impact than brand name on consumers’ evaluation of the quality and the DUI-chase value of automobiles, VCRs and shoes. Specifically, COD explained the lE'J'geSt proportion of the common variance, followed by COA and brand name for the measure of quality. Price played a minor role in explaining product evaluations. In addition, a brand or COO’s favorable perception was affected by providing additional Product-related information. That is, the evaluations of products from prestigious developed countries (like Italy and Japan) went down considerably when other product 15 information was provided, and evaluations of products from newly industrialized countries (such as Mexico) went up. Similarly, after additional product information was provided, the evaluation of brands from developed countries dropped slightly, and evaluations of brands from developing countries increased slightly. Finally, the authors suggested that a brand’s quality image might decrease if it was designed or assembled in a less prestigious country. The evaluations of brands from newly industrialized countries, however, were less affected even if these brands were manufactured in prestigious countries. Iyer and Kalita (1997) also distinguished between COO -- as COM and country of brand origin (COB) —- and examined the effect of these two dimensions of C00 and price cues on consumer perceptions of quality and value. They used within-subjects designs with 2 price (high and low) x 2 COB (US, and an appropriate European country) X 3 COM (U .S., one newly industrialized country [South Korea], and one developing country [China]). They found significant effects of COM and COB cues. Price cues were found to have less impact on brand choice when COM and COB information was given. This study, however, did not include information on specific brands. C00 and Multi-Attributes or Mum-Dimensions of Quality Some COO researchers conceive of product quality as having several attributes or as having a set of dimensions from which quality is inferred rather than as being a SunlInary construct at the abstract level. Such multi-attributes or the multi-dimensional aSPCCt of quality have been reflected in the concept of country image. Some COO r eseaI‘Chers examined the effect of C00 in terms of the fit between countries and product categories (Nagashima, 1970, 1977; Roth & Romeo, 1992). These researchers 16 considered the country image as multi-dimensional. While some researchers used product specific attributes to examine the effect of C00 (e.g., Erickson et al., 1984), others used quality dimensions, which were extracted from many product attributes using principle components factor analysis or confirmatory factor analysis (e.g., Roth & Romeo, 1992). In this section, first, a discussion is provided of how country images using multiple dimensions were defined in previous studies. Then, the relationships between COO and country images in the previous studies of uninational and hybrid products is reviewed. The studies of C00 and multi-attribute or multi-dimensional quality also are classified according to whether a single cue (the “made-in” label) was provided, or multiple cues, including the country cue, were provided. C00 and Multi-Attributes or Multi-Dimensions of Quality: Country-1M Dimensions and Quality Dimensions COO researchers are interested in the dimensions of country image, since they believe the effect of C00 could be specific to the quality dimension as well as to overall quality, Several studies discuss the dimensions of country image/quality? (See Table 2.) \ SinCe country image IS Viewed as the overall perception of quality for products made in a given country t(fllkey & Nes, 1982; Han, 1989; Roth & Romeo, 1992), dimensions of country image are considered to be e Same constructs as the dimensions of quality. 17 .8553 :8: 2: 5 82: 82 9 $5885 2&3 853$ 05 5 85:5... =m 3:: 35553. 22:» 5:6 05 no 085 e508 2: E .5385 0.53 $25 98 3205—03 8523 £88m 52er 93. .mEu: =m 5 >558 o5 Eat 30:55 55 $5558.. 05:3 5.5 05 .5 365 “Eu 2: E 5555 ii 5:85 8:95 83 been onto $58 2F a 6me 5:50 83 5:3. 05 co £58 25. 29$ $0555.55 5:832: 28 awe—0:58 81m 8583 £88.“ 025 imbued 58$ 583 .533 weave—.82 A33 3 85: m.aEEmmwaZ 80¢ won: 203 25: 58:5. 32558 I 858m 30:55 1550 Lnauconoz a. chem ooggotom 65665 552533. wean—ESQ 3555 E 890$ 58595 on 9 3555—53 258 a; 5:? seem 83: $5252 5 8% $523 $2553 cm 05 no So 9523. one? @583 £53586 3E 56F. 32605 3350 d Rafi—2. .550 A595 b55385 A553 bin—35m 63 305350qu 208350 80:88. See 3535 055 83 ES”— cozoseoi 89a gauge 208350 m2 Soc.“ 3585 230$ Aswan—$3 owsmoi :5 £93.28 56$ an 53.5 fiancee 3E .53,“ Agaatme eoEewoooz 85$ m.mEEmwme Eo¢ e8: 20>» £5: bcozfi. Aceqatma >530 $0255 .2050 £5: «55552 £58586 o>c 805 .8 owns 83 535583 02 .3385 025 Soc 55:53 203 $5: 0:55 FEE—550 555 .23: 552 E 8208582 83:3 2?. s. :55 e5 55: 355mg main 3 @8852 203 5:852 a. wEmEo>e< 33: 80me co 2:2 3.3-3 be 35.8 2E 35055 e 8:3 52 25> a. 025 30:55 1550 65: “~822me 35.550 5:52 p.535—22— 85530 835.5 :55 53283.5 3:50 .N «EFF 18 >080 08 .«0 0:08 0E. p 80 05 .80 0:08 05. a l9 0558088 0005 m0 Coo: MARE—F» 05 0000 8 050 8000 00: 0:03 803080 02 005%“..U games— 28 50—05. 02 ._§§§ a. 32:0 $2 .5380 2322: 8:0: x0 9 . $505: 5 00000 00555 0.530086 008E. £0.80m 058— 000803 000080850 .3005 90:9 0:088, 00:05. moc0fim0< 020 805.2 .E .0008 m0 _0>0_ 03800000 0000088 b2508D £03000 00000.8 80888500 0568086 :30 0005 mum—380% 5a 080: 005.51.: 000.300 805:0 0E. .530 b__5000_>._0m Ea: 0.8800 808 00.50800 828086 Emma 00828800 385 0005 C00: 9800 a. 5 0.05003 00—055 @6305 00000.0 was: 02808053 3900 BEA—0:0 0003 800 00:3 008050 0.03 56808:. 05 030005 000% 80500»— >._=O .000: 028086 8000 m0 0.080.008 80: 030% 5609 00.308082 050.30. 008800: 808 003000 828086 80m E .00m 300: «00808 a. 80m .3550.» bmznaoomtom 00200 003 560088 00_>._0m aux—0:0 080005 0050833. .0000 80¢ 003000 80008 80m .581 9 00080103 I . 55$ mWEEmawaz 80¢ 0000200 080: 800%MMM IE m=0_m_>0_0._. $55 «085008. “WNW 220:0 00008 800 08080 80000 03 «.805 $080 855 0.0820052 88.00% 0000 080: SWEFM E 00—30823» «3080002 00. 500883 .1 380885 00802 nacégtousau 00:00.:— 7 .805 .8 N .8 Most studies identified the dimensions of country image by Using ex p 1013 tozy factor analysis with multiple quality items (Nagashima, 1970, 1977,. Nmyana’ 198 1,- Cattin, Jolibert, & Lohnes, 1982; Jaffe & Nebenzahl, 1984; Johansson & Nebenzahl, 1986; Han & Terpstra, 1988). A few studies attempted to determine quality dimensions based on systematic or theoretical approaches (Roth & Romeo, 1992; Li & Dant, 1997; Thakor & Katsanis, 1997). Although these studies are discussed in the literature review section, critical reviews of these studies are included in the following pages to focus on defining quality dimensions and the validity of these dimensions. Interestingly, as Table 2 indicates, all the studies performing factor analysis to identify quality dimensions used Nagashima’s (1970) measures; some used all twenty items and others selected a subset. In spite of using the same twenty items, the studies indicated that the factors extracted from them differed according to the subjects employed and the products used. In particular, Narayana (1981) found five factors (quality, recognition, prestige, production form and expensiveness) for US. consumers3, but tWO of these five factors were different for Japanese consumers (popularity and functionality as oPposed to production form and expensiveness). In the study of J affe and Nebenzahl (1984), different factors were obtained when different questionnaire formats were used With the same respondents, the same questions, and the same products. (See the“ literatllre review for a detailed description of this study.) Thus, we see that 1‘ is d. . ° to Imeult to obtain a set of widely accepted general quality dimensmns usmg “9‘0"“ ‘7 fact . or analyms. 31 .. 111 - onsumers "giggl‘Y. Narayana (1931) found four factors for US. consumers and five factors for Japanese c ‘n factor— ‘abel ~ factor analysis, bl“ he eXpanded the four factors to five factors for the (1.5. group to 0btal ‘h g consistency between these two groups. 20 T em t . . . . . . here have been afew att P s to identify quality dimensro 1'28 Usmga COHmeal or theoretical basis. Due to a lack 0f comparable data, Roth and Romeo (1992) used a qualitative, rather than a quantitative, approach to review eight studies that assessed dimensions of country image. That is, they used three criteria to determine country- image dimensions: The dimensions (1) were consistently found in previous research, (2) were related to perceptions of a country’s production and marketing strengths and weaknesses, and (3) were applicable to a broad range of product categories (p. 480). These criteria yielded four dimensions of country image, including innovativeness, design, Prestige, and workmanship. The definitions of these dimensions are as follows (p. 480): lnnovativeness: Use of new technology and engineering advances Design: Appearance, style, colors, variety Prestige: Exclusivity, status, brand name reputation Workmanship: Reliability, durability, craftsmanship, manufacturing quality Roth and Romeo (1992) mentioned that these dimensions were all production and marketing oriented and their use could be justified because consumers formed their Perceptions by relating a product to what they knew about a country’s ability to produce goods and services. Since the researchers used a single item for each dimension, they Could not obtain discriminant validity among these four dimensions and suggested a um.‘iil‘itlensional country-image construct. In addition, they omitted serviceability and price/Value dimensions, which were frequently mentioned in previous studies, WNW“ an - Y J hgtifications. (See Table 2.) . . - . 9 984 Li and Dant (1997) borrowed d1mensrons of product quallty from Gan/111 S (1 ) Garvin identified eight dimensions (performance, serviceability, reliablllt)’, 21 durability, aesthetics, conformance, features, and image) based on a . . ual' . , Syatficflsofgg/We approaches to VICWII'lg product Cl lty, including 1) the transcend . 6’71 approach, “411011 proposes that quality can only be recognized through experience and Gama, be precisely defi [16d or measured; 2) the produCt-based approach, which takes the View that the quality of a product can be measured based on the quantity of some desirable attributes; 3) the manufacturing-based approach, which assesses product quality based on the level of conformity to prespecified manufacturing standards; 4) the user-based approach, which determines product quality based on consum - d ers’ perceptionS; and 5) the value base approach, which d - The rice. efines quality based on performance at an acceptable 13 definitions of the eight quality dimensions are as follows: . 5 oi 3 Qtodugj‘ot . . ' 0 . Performance! Superiority or excellence of primary operating character‘s? ”paged 3“ Serviceability: The extent to which a pl'OdUCt can be easily and adequate Y . 06 OK maintained . ed gefl Reliability: Probability of the failure of a product within a pre-determ\“ time Durability: Physical life of a PTOdUCt Aesthetics: The look, feel, sound, taste, smell, style, or fashion ofa product Conformance: The extent to which a product’s design and operating charaCteristics meet F fife-determined standards eatures: vailability and excellence of “bells and whistles” . r Image.4 The extent to which a product conveys an upscale 0:31am]?- t9 a product (L1 & Dant, 1997, pp. 99-100) 0p ”neat . Image Garvin’s (1984, 1987) eight quality dimensions, however create 3 ’ O C onceptual e and operational problems. First, GarVin (1984) noted that in many cages it . . . Was diffic to distinguish Performance (primary product characteristics) from feature ( ult S SCCOnd ary characteristics), since the distinction between the two depends on the “331” s percepti 0“ of V\ 4 G . . . . arvm (1984) originally named this dimension “perceived quality.” 22 the degree of importance of the attribute in question. Second, the lis’l‘lbcfl'oa hem/5"” reliability and durability was not clear. None of the previous studi <19 ofcvo identified these two dimensions as separate Constructs, as indicated in Table 2 Roth and Romeo (1992) attempted to combine these two constructs in the dimension of worManship, which included craftsmanship and manufacturing quality. Third, the concept of aesthetics is confusing in that it contains the “look” and “feel” ofthe product. These attributes may overlap with image. For example, if the . . . I . . - m fi‘Ofn the product’s “look” and “feel” are luxurious, it is hard to distmglnsh the ° for the - a roprlate “upscale” image. Finally, the dimension of conformance might not be PP for diffie‘m . . - d. It is consumers’ perception of quality, since it IS managerially determine , d duo‘s 5 dCS‘gn 3“ ° to the consumer to find the source of pre-determmed standards for a p operating characteristics . \aiilOIlS . - colic The conceptual ambiguity of GarVin’s eight dimenSIOns leads to hlgh among these constructs in the empirical studies. Li and Dant (1 997) devised multlple measures for each dimension for men’s dress shirts and examined validity Criteria and their direct relevance to the C00 research. Although they rePorted a . . . - . t Garvin ’S taxonomy met key validity criteria, and these dimenswns were usefiu in C0 , . . . - . . research there was a problem with lack of discriminant validity. First, there was a . 2 8h leve multicollinearity among the constructs. The construct correlations table indie 1 Of a ted that th highest correlation (0.93) was between aesthetics and image, followed by rel e iability and performance (0.92), durability and reliability C33), and durability and perforrrmnce (0.77). Conformance was highly correlated with reliability (0.90), With performance (087), with aesthetics (0.87) and with durability (0.80). This lack of discrimination 23 among Garvin’s (1984) dimensions Was found in another study d 0115' 57511106 Md . ‘ne th ' ' ' ' , 251tharnl (1991), who also examr e empirical estimation of the 9012.9 6% [caved quality dimension for automobiles. They reported that there was no practical way to discriminate between the reliability and durability dimensions of quality. Further, they found that aesthetics was within the PreStige dimension, and that the conformance dimension did not belong with the others. F inally, Thakor and Katsanis (1997) proposed quality dimensions by following Karni (1973): the categorizing scheme suggested by Nelson (1970) and Darby and . Cram/in . with the . ' this scheme search, experience, and credence. They 21150 Incorporated emcnts’ . melt Stat - ordlng ‘0 (1984, i987) and Brucks and Zeitharni (199i) approaches. Ace me“ are {owed w ' of use, , the search dimension includes perceImonS 0f features and case (at 0i me Q0 serviceabXW . . . ed at based primarily on search, since these dimenSions can eaSflY be assess purchase. The experience dimension consists of perceptions of reliability, and performance, Which are likely to be formed based largely 0n experience, because consumers assess these aspects of quality only afier the product has been Purchased and used. The credence dimension5 includes aesthetics and prestige. These dimensions, however, are confusing, since the anthors do Hot n O ‘ . . r . preCISe definitions for them. For examPle, they identify features as a subd. P 0V1de Imensi o . 0 search, and performance as a subdivismn of experience. As noted above he 11 0f ’ weVer i . o a o . o . , t IS difficult to distinguish performance from features, Since the distinction between th e two 5D?“ by and Kami (1973) defined the credence dimension as a quality perception that the consumer cannot verify even afier use such as surgeries or car repairs. Thus, Thakor and Katsanisi (1997) credence dimension is QUite different from Darby and Kami’s (1973) original credence dimension. They, however, dd “0‘ eXplain these differences. 24 depends on the user’s perception 0f the importance of an attribute ( Q - 3’74}; L014 ). 7W" it is not cl 6 ar how features can be Included in the search dimension. Further, they mention that Products in the search dimension included clothing; the prodUCtS in the experience dimension were pocket pagers, cameras, and VCRS; and Rolex and Chanel were examples of credence brands. Thus, their classifications within the quality dimension are also applicable to categorizing product types. They did not explain, however, how their categorization scheme could be used in both cases. . ' ' of this Moreover, no empirical attempt has been made to examme the valldlty be to ° 3 . . ensions seem classification. Thus, Thakor and Katsams (1997) quallty dim \ to explain . . - mat he 9 inadequate for use in C00 research, although they proVide mslghts “3mg. - . . oduc‘v q the relative importance of C00 and brand Image in the perceptlon of pi Products: Singw Studies treating product-country image as a muui'dimensional constflICl are 1'00“?d in a StUdy by Nagashima (1970), as previously discussed. He asked . . ‘ “116mm and Japanese busrness executives to evaluate products from five different co un . . . es twenty items “Slug semantic-differential scales. He then grouped his twe based on n . f d' ' - - - - . thaws int “’8 lmenSIOIlS. 1) pnce and value, 2) serv1ce and engineer-mg 3) ad 0 ’ v el'tisin E reputatlon, 4) deSIgn and style, and 5) consumer S profile. Although no justif d lCatiOn made for these groupings, these twenty items have been used, in some COmbinat' 10m, as the basis of many product—country-image studies (Brodowsky, 1 996). Nagashima found that the products made in Germany received the highest ratings from the two groups of r eSpondents. Japanese respondents, however, perceived products made in France as 25 w having higher prestige value than did American respondents. PI'Od II 01:: . . made/)1 we “5’ were rated hlgher by American reSpondents than by Japanese I‘e Spa/meats- 771656 findingS, however, were not statistically tested. Later, Nagashima (1977) replicated the study using only Japanese subjects, The results indicated that the rating Of products made in the US. had decreased for the reliability and Workmanship dimensions. On the other hand, products made in Japan, that had been perceived as inexpensive and unreliable in the previous study, were now viewed as expensive as US. products and as reliable as German products. Although ‘10 StatiSfical tests were performed, these results indicate that the product—country unage over time . anese ewdudis Narayana (1981) examined the “made-in” image for US. and 33? Wet“? . , 00 as Pefeewed by US and Japanese consumers. Image data were obtamed W semantic-differential scales used in Nagashima’s (1970) study. 111 order to be in the factor labeling, five factors were obtained from each of the U.S. and the Japanese samples. These factors, however, differed between the two groups (See Tab 1 . . ' e 2. The results indicated that US. consumers perceived American-made produCt t ) S O . . e of hlgher quallty than Japanese-made products. In addition, U.S co generally U s ”Sinners . . products to be more expensive and more prestigious than Japanese p Perceived roduct other hand, the Japanese consumers perceived the Japanese products to S. 0n the be of ~ . . h quality than American-made products and Japanese products as Slightly leg lgher . S prestigious but more functlonal, than US. products. , Cattin et al. (1932) examined American and French purchasing managers’ view S about the effect of (:00 on the evaluation of industrial products made in France 26 Gennany, England, Japan, and the US. The researchers employed t}, e . . . “Vang/s M”? differential scales used by Nagashlma (1970, 1977). They selected 17,, d 6:10“ e [Mansion-5' of the twenty that previouSIy were “found to be important factors in 1‘00,me purchasing decisions” (p.135), including pricing, reliability, workmanship, technicality, and performance- For American and French respondents, the mean scores for these five dimenSiOIlS were obtained for each of five “made-in” countries, and then these mean scores were summed to indicate the degree of “favorableness” toward each “made-in” . , rably by country. The results showed that “made 1n Germany ’ was evaluated most favo -group (American both American and French respondents. They further performed two t e diffefe“ . e '1“ fiv versus French) discriminant analyses for each of the products mad cefi me 311665 ffer the ' countries using twenty items. The results indicated significant (11 the?“ ’ evaluations of American and French purchasing managers - 596° discriminating powers ((1)2) between the American and French groups were higher for products made in the US. (.638) and in France (.652) than for those madc in Germany (405), England (370) or Japan (353) Thus the . , - - ’ authors sug gested that American and French respondents might have home-country biases ba (1 se pride or patriotism, 0” national Roth and Romeo (1992) examined the effect of C00 in terms ofth e fit be countries and product categories. They proposed a framework of (lln tWeen )faVo fab] e 6 These five dimensions are five items of Nagashima’s (1970) twen - ty. These semann - - . :35: :11. 2:302:51)! priced/unreasonably priced, reliable/unreliable, careful and meticulou: figfinfiiifim with outward metlculous workmanship, tecmlcally advanced/technically backward, and more concerned analysis but f appearance/more concerned Wlth performance. They performed Principal components . Ound no major set of factors that explains the 20 items. 27 (mis)match between countries and product categories, which could proylae saategies for managing the products’ COO. They posited that a favorable product.counay match would occur when a country’s perceived strength is an important product feature for the particular product category (for example, French shoe). In this case, promotion of C00 is advised. Conversely, an unfavorable pI'Oduct-country match would occur when the important product features are not associated with the strength of the country (for example, Hungarian shoe). In this case, managers are advised to emphasize benefits other than C00 and to seek joint-ventures with favorable match partners. On the other hand, a favorable mismatch would occur when a country has a positive image related to a certain product feature, but that feature is not important for the particular product category (for example, French beer). In this case, managers are advised to downplay the importance of product-category image dimensions and to promote COO as a secondary benefit if it involves a compensatory choice process . Finally, an unfavorable mismatCh would occur when an image dimension is an unimportant product feature and negatively associated with the country (for example, Hungarian beer). In this case, the authors recommend that managers ignore COO. Using a cross-cultural sample consisting of graduate students in Ireland, Mexico, and the 11.3., Roth and Romeo (1992) measured country images fol. ten Coumn’ es regarding four image dimensions (innovativeness, design, prestige, and workmanship) and also measured image importance for six PTOducts in relation to these dimenSions. Each dimension had a single item. They performed principal -components analysis and confirmatory-factor analysis to ensure the multidimensionality of the country-image construct. The results, however, indicated that the country image Was a uni d i menSiona] 28 construct". The researchers suggested that this finding might be due to the use ofonly four items to capture the country image. They also found that favorable product-counfly matches on the image dimensions seemed to predict consumers ’ Willingness to buy foreign products. C00 and Multi-Attributes or Wensions of Quality of Uninational Products: Multiple-Cue Situations Erickson, et al. (1 984) examined the influence of an image variable, C00, on the formation of beliefs and attitudes in a multi-attribute model framework. Image was defined as “some aspect of the product that is distinct from its physical characteristics but that is nevertheless identified with the product” (p. 694). They modeled the possible bi— directional effects of beliefs about specific attributes on attitudes (Overall evaluation), and the effects of attitudes on specific product beliefs (halo effeCts). They measured individuals’ beliefs about specific attributes including reliability, durability, workmanship, mileage, and price for ten automobile models. These five attributes were reduced to two factors, quality and economy, and then these two factors were summed to yield a single score, which was used in the equations instead of the five attributes. Two equations were estimated. One had attitudes as a dependent variable, and its independent variables were beliefs about attributes, familiarity, COO (an image variable), and true value for each attribute obtained from Consumer Reports and Car & Driver. The other equation had beliefs about attributes as a dependent variable, 311d attitudes C00( ’ an 7 In the Case of principal-components factor analysis, unidimensionality of the coun -ima e found for reSpondents from all three nations. In the case of confinnatory-factor analysis Ugs ”Timid .was respondents showed unidimensionality Of this construct, but Irish respondents inditzatieditvvanf Mexican innovativeness and design, and 2) prestige and workmanship. These two factors, h °Wever 0 actors, 1) correlated (0.993). This indicates questionable discriminant validity of the two conga-u , were h‘gllly authors decided to use the unidimensional solution. ets. ThUS, the 29 image variable), and true value for each attribute as independent valiables. 7716 researchers found strong bidirectional effects of beliefs on attitudes and effects of attitudes on beliefs. Familiarity had a direct effect on attitudes. The image variable, COO, directly affected beliefs, but not attitudes. In a study similar to that of Erickson et al. (1984), Johansson et al. (1985) examined the effect of C00 on beliefs and attitudes in the multi-attribute model framework. In their study , however, thirteen attributes for automobiles were measured by Japanese and American students. These attributes were reduced to three factors: reliability, horsepower, and driving comfort. Three other attributes, gas mileage, handling, and styling, were included. Each of these was independent of the first three. Each of the five attributes8 was treated as a dependent variable or as an independent variable rather than summing these scores, as in the Erickson Ct 81. study (i984)- Car ownership and the demographic variables (age, gender, income, and nationality) were incorporated into the equations- The results of the simultaneous-equations estimation indicate that each of these attributes had a significant impalet 0“ the Overall rating (“Md“) for each car. In addition, a significant halo effect (a positive impact of the overall rating [attitudes] on the individual attribute ratingS) Was found. COO had no effeCt on attitudes, bUt it did have a modest impact on attribute ratings (that is, COO had significant effects for Some cars on some attributes, but not for all cars on all attributes). This study did net find a POSitive bias among Japanese respondents towards Japanese cars 01' among Arnerican respond ents toward American cars. 3 Styling was excluded in the equation, since it has no “true” scores from published sources 30 Jaffe and Nebenzahl (1984) report an interesting experiment Using Nagasju'nza’s (1970) framework. They used two formats; in the first, respondents rated products fiom one country on all thirteen items before rating products of other countries. In the second format, respondents rated all five countries on one attribute before evaluating the next attribute. The student sample was divided into two groups. One group received the first questionnaire format, and the other group received the second. Two weeks later, the students were asked to fill out the same information, but half the students had the same questionnaire fGenet they received the first time and the other half were given the alternative format. The researchers performed principal-components factor analyses for each questionnaire format. From the first format, they identified tWo factors -- product technology and marketing characteristics -- that explained 36 percent of the variance in the original thirteen items. Two different factors -- product technology and price __ were extracted from the second format, which explained 31 percent of the Variance in the original items. From these results, the researchers concluded that the dimensions Suggested by Nagashima did not sufficiently explain product-country image thermore, they suggested that the identification of the dimension might depend on the , questionnaire format. C_OO, Brand, and Multi-Attglggtes or Multi-Dimensions of uali , H . Products W Attempts were made to identify the dimensions of the quality Construct based on multi-attributes in early hybrid-products literature. Johansson and Nebenzahl (1986) 31 examined how consumers’ perception of the brand quality and the image ofpassgnger cars could be changed by moving the location of manufacturing fiom one place to another. Respondents were asked to evaluate six different COO concepts as well as two Japanese brand names and two American brand names on the basis of thirteen attributes. Using principal components analysis on the scores for each brand and each “made-in” concept based on the thirteen original rating items, they found that two factors, “economy” and “status” represented 58.1 percent of the variation in the original data. Respondents were also asked to rate twenty-four combinations of brands and “made-in” labels based on the original thirteen attributes. That is, respondents rated a hypothetical Honda made in Japan, a Honda made in the U.S., and Hondas made in four other countries: Germany, Mexico, South Korea, and the Philippines, The authors mapped the average scores for each of the twenty-four cars on the two dimensional spaces (status and economy). From this mapping, they found that manufacturing cars in a developed country contributed to a more favorable brand image than manufacturing Cars in an underdeveloped or newly industrialized country. That is, West Germany was rated as the best country in which to manufacture a car, and the images of all brands in this Stud were improved by moving production to West Germany. On the other hand, movingy pr0duction to a low-image country would result in deterioration of brand image_ The authors further quantified the potential benefit or losses from mOVing production to a different country by measuring dollar preference effects. That is, respondents were asked how much they would pay for a Honda if it Was manufactured . in Japan, or how much they would pay for a Honda manufactured in the U S Th 0 o e 32 regression analysis indicated that 17 percent of dollar preference effects could be attributed to changes in the dimensions of status and economy resulting fi-om moving production to another country. Han and Terpstra (1988) examined the effects of C00 and brand name cues on US. consumers’ evaluations of uninational and binational produCts. RespondentS were asked to rate ten automobiles and ten televisions on six dimensions: technical advancement, prestige, workmanship, serviceability, economy and overall qualitY- These ten automobiles and ten televisions represented different brand and COM combinations: 1) foreign made and foreign branded (e.g., Honda Accord made in Japan); 2) foreign made and domestically branded (e.g., Ford Tracer made in Mexico); 3) domestically made and foreign branded (e.g., Honda Accord made in the US.); and 4) domestically made and domestically branded (e.g-, Ford Mustang made in the US.) The researchers found that source country (COM) and brand name had significant main effects on the evaluations of five dependent variables (technical advancement, serviceability, etc.) for most TVs and automobiles. In additiong'gteracfion between source country and brand name was found to exist in most cases. 3:" The effect of source country was found to be stronger than that of bradd name for the U-S,-Japanese and the U.S.-German designs of both cars and TVs. In the case Of the U.S.-Korean designs in both product categorieS, however, the brand-name efibc t Was stronger than the source-country effect. In addition, the relative magnitude of brand- name effect and source-country effect varied across product dimensions, Such dimensions as serviceability and workmanship were found to be more Sensitive to SOUrce country than to brand name. 33 Tse and Gom 0993) examined the effectiveness of coo and global brand name in Amaican consumers’ evaluations of stereo systems at the attribute level and at the overall product level. The attribute evaluations included workmanship, performance, sound quality, likelihood of breaking down after warranty expires, and size of whole system. The overall product evaluations included proud to give as a gift, liking, product quality and product worth. The C00 and global brand name were manipulated in a 2 COO (a developed country [Japan] versus a developing country [Indonesia]) x 2 brand (well-known/unknown) design in WhiCh subjects’ evaluations were obtained both before and after they tried a product. The results of this study indicated significant main effects of the C00 and brand name at all the attribute levels, except that there was no brand effect on the size of the stereo system. Significant main effects of the COO and brand were 21150 found for the overall evaluation levels, except that them were no brand and country effects on product worth. The interaction between C00 and brand name was not significant at either attribute or overall evaluation levels. In addition, the C00 effect and brand effect declined afier experience, but experience did not seem to remove the C00 effect. Therefore, the authors concluded that COO was still a salient factor in consum e r product evaluation, and it appeared to be a more enduring cue than the glob a1 brand name- Tse and Lee (1993) decomposed country image into component and assembly origins. They examined the effects of country of component (COC) (Japan versus South Korea) and country of assembly (COA) (Japan versus South Korea) as well as the effects of global branding (Sony versus Gold Star) and product experience on product evaluation by American consumers. Product evaluation consisted of six measures includ' 9 lng 34 performance attributes (performance and sound quality), 10ng-tenn attributes (deterioration of sound quality and likelihood of breaking down after warranty expires}, social attributes (proud to give as a gift and proud to show fi'iends), purchase value (good/bad buy) overall evaluations (liking, product quality, and workmanship), and confidence in the above evaluations. In their first study, the effects of COC and CO A on product evaluations were examined using stereo systems as the stimulus. They found that COC was significant in both long-term attributes and overall evaluations and that COA had a significant effect on long-term attributes and only a marginally significant effect on overall evaluations. :Thus, the authors suggested that when the C00 was decomposed into component and assembly origins, its effects were less than those found in earlier COO Studies. 1 In their second study, the brand effect as well as the effects of COC and COA on Product evaluations were examined Unlike Ahmed and d’Astous (1996), they found that brand name effect was much stronger than component and assembly originxe’} In fact, while brand name had a significant effect on all types of evaluations, compo;ent origin had a significant main effect only on the performance attributes and overall evaluati on s and assembly origin had no Significant main effect across all measures. The Significant, interaction between brand name and component origin suggests that welldmewn brand names seem to overcome any negative effects due to component origin Finally if American consumers had eXperienee With the Pr 0‘1““, Dr 0‘1““ evaluation of a well- known global brand was not affected by Changing assembly origins, but a less-known brand assembled in South Korea was found to have a higher product evaluation than th e one assembled in Japan. 35 Chao (1993) also partitioned COO into COD and COA. Tl'lis study consisted of an experimental design of 2 price (low and high) x 3 COD (US, Japan, and Taiwan) x 3 COA (Taiwan, Thailand and Mexico). The price and design location were specified as between-subjects factors and the assembly location was specified as a within-subjects factor. For American consumers quality evaluations were measured at two levels including design quality and PIOdUCt quality. Design quality items were imitative- innovative, common-exclusive, and conventional-stylish. Product quality items included workmanship, reliability, durability, and product quality. His study, using televisions, found significant price and COD main effects on design-quality perception, but no significant COA main effects or any interaction effects between COD and COA. On the other hand, all three of these factors (COA, COD, and price) had significant main effects on the product quality perception. No significant interaction betWeen CO A and COD was found, That is, a low evaluation of a product assembled in a country associated With negatively perceived quality could “Qt be compensated for by having the prOdUCt designed in a country with a POSitlve design stereotype. AlthOUgh the authors provided multiple-attribute information about the stimulus to respondents, brand name was not included, even though another study found that it had a stronger effect than COO (Tse and Lee, 1993). Thus, the results of this study are questionable. Later, Chao (1993) decomposed COO into three components, Country of parts (COP), COD, and CO A, and examined the impact of these COO dimensions on Product quality and design quality perceptions. American respondents were provided a picture of the product (a stereo television) as well as a list of product features_ Each of the aSSCmbly, design, and parts locations Were manipulated using two Country levels (U S 36 d Mexico while the brand nam - . . an ) e and price information were Gonna/led The study found that the COA and COP have a Significant impact on product Cl Uality perception, but - f - . D has no Impact, nor do any 0 the mteraCtIOn effects That is, a negatively perceived CO th 21 n ' egatwely perceived COO-assembly-location quality associated Wi product ted for by a - - . . mpensa POSltlve (:00-des1gn—locatron stereotype 01‘ p6 could not be co stereoty 6. On th p e other hand, the COD had a significant 6 COO-part-location stereoty a positiv effect on design quality perception, but COA and C0? had no effect. The significant threeeway interaction (COA, COD, and COP) implied that Whereas a strong COD stereotype reinforced b y using parts from the UI'lited States could not be used to overcome a weak COA stereotype, U’S' parts could be Used to improve a negative COD ct was shown to be assembled in Mexico \it‘l “any ° '\ stereotype when the produ F‘mfiw’ (4" and Dani (1997) selected a number of GarVin’s (19843 0t dimensions that com be used MO 88 Pm‘h‘“t “ate gal“ and ‘6 sted them em They dew doped. quesixcmaires concerning office white dress Shirts for men based on Garvin:- s deftmt'ions of eight quality dimensions: performance, Service ability 1. b'l'ty , re 13 11 , durability, aesthetics, conformance, features, and image. USing a 4 COO (U '8', Italy Hong Kong, and Indonesia) x 2 brand (name brand [Van Huesen] versus fiCt' ItiOuS b1. and [W1nsl 0WD between-subj ect factorial design, they distributed questionn aires t 0 a c ' ' onvenienoe saInple of white-collar profess1onals (men and women) in the metro . polltan fir . ea Of a major Northeastern US. city. The data were analyzed to eXamine key validity criteria (unidime - HSIOnah' ty, te“ - o g C O lablli y, Convergent validlty, discnrninant validity, and 110m010g' al v t 1c alidity) Camimatory factor analysis showed that the revised mOdel had an acCeptabl f 6 It to the 37 data, and this indicates the unidimensionality of quality. The composite reliability coefficients indicated acceptable reliability for the revised dimension Scales. To evaluate convergent validity of the revised model, each indicator’s factor loading on its aSSigned construct was examined. All is were positive and significant. Thus, convergent validity was achieved for each of the eight quality scales. Discriminant validity was examined using X2 difference test procedures. (See Anderson and Gerbing 1988 for complete procedures.) The results indicated acceptable discriminant validity. However, the high correlation between image and aesthetics and between performance and reliability requires further investigation of the discriminant validity Of these dimensions. To examine nomological validity, the researchers ran MANOV A and ANOVA as analyses with the eight quality dimensions as dependent variables, and brand and C00 two independent variables. The individual items were averaged under comsponding latent constructs. The results Show significant effects of C00 on seven of the eight quality dimensions. (Serviceability was the exception.) On the other hand, significant main effects of brand were found only for performance, reliability, durability, and conformance. Significant interaction effects between C00 and brand were found f0 1‘ reliability, aesthetics, featureS, and Perceived quality. C00 and Purchase Intention The effect of C00 on purchase intention has been examined less frequently than its effect on quality perception. As indicated in Table 1, very few StUdies of hYbrid products have examined the effect of C00 on willingness to purchase. Thus, in this section. the focus will be on the studies of uninational products. 38 Schooler and Wildt (l 968) were the first to employ put'Chase intention as a dependent variable in the C00 literature. The results of the study indicated that in the multicue situation, American respondents were less likely to buy products made in Japan, but the unfavorable image 0f foreign PTOducts could be overcome by lowering their prices. Ettenson, Wagner, and Gaeth (1988) examined the effect of C00 on Willingness ‘0 Pm'ChaSC apparel before and after the airing of a “made in the USA” TV campaign. American students were asked to rate their purchase intention for women’s blouses and men’s dress shirts based on the information about sty 1e (trendy or classic for blouses and content (100% full or tapered cut for shirts), quality (average 01' above average), fiber and 01.111th Of origin (made in UHS 01' made in China)’ . 'te (designer or private label). They were asked to fill out the questionna‘ cotton or cotton-polyester blend), c brand name before the first airing of the “made in the USA” campaign (pretest) and thxee m0“ 5 later, during which time the campaign ran on television (posttest). Using con}Dim taskS, the authors found that fiber and price had the greatest effects. The COO main effect explained 4 percent of the variance in the female respondents’ purchase intention and 3 percent in the variation of the males’ purchase decisions in the pretest. In the POSttest, the C00 effect explained 6 percent of variance for the entire sample. Thus, the aUthors mentioned that respondents used other extrinsic cues more heavily than C00 and that the “made in the USA” campaign had no effect. Wall, Liefeld, and Heslop (1991) used multicues also to examine the effect of C00, as well as quality, risk and value perceptions, on the likelihood of purchase. For three products (shirts, telephones, and wallets), Canadian respondents were provided 39 information about price (high or low), country (Canada for all three products, a developed country [the US. for the shirt and telephone, and Italy for the wallet], and a less developed country [SOUth Korea for the Shirt, Taiwan for the wallet, and Hong Kong for the telephone]), and brand (Well-known or unknown). The authors employed an analysis-of-covariance model using age, gender, education, and perceptions of ability to judge products9 as covariates. {The results showed that COO had a significant effect on the quality perception for all three Products. but it had a significant effect on the \l‘ . likelihood of purchase only for shirts. Brand had less impact than COO dld. It had a . . . . n significant effect on the likelihood of purchase only for Shlrts and its stgmficant effect 0 phones. Although the analysis-of-covafiance models were S exp1ained only the quality was found only for tele significant for all three products, the R2 values indicated that the model . f 17.3 percent of the variance in likelihood of purchase for shirts, and 5.3 percem 0 variance in likelihood of purchase for wallets and telephones. Product Familiarity Some COO researchers found that COO effects are strongest for consumers with little or no product familiarity resulting from brand familiarity (Cordell, 1992; Han & Qualls, 1985; Samiee, 1994). That is, if brand familiarity is low or nonexistent COM will have a greater effect in evaluating PTOdUCtS- On the other 1‘21an brand familiarity is high, brand image will be more important in the purchasing decision than COM, /, For /I example, Han and Qualls (1935) found that consumer perceptions of product quality for ___— If 9This variable was measured by the item indicating the consumer’s perception of how good a judge he /she felt he/she Was for each of the three products. 40 Grundig televisions (an unfamiliar brand) were favorably influenced by their COM when they were made in Germany. Conversely, they found that COM had less impact on consumer perceptions for Honda (a familiar brand). Similarly, Cordell (1992) found that preference for Timex (a familiar brand name) decreased only slightly when the product was made in Pakistan as opposed to West Germany. In the case of a watch with an unfamiliar name, Tempomax, preference declined sharply when the product was manufactured in Pakistan rather than in West Germany. Some COO researchers have suggested, however, that consumers are more likely to use “made-in” labels when they were more fathiliar with C00. 1011213550n at al. (1985) found that familiarity with automobile models with different national origins (that is, familiarity with Japanese and American cars) had variable influence 0“ product evaluation, reflecting more positive or more negative perceptions of different ambutes- Johansson and Nebenzahl (1986) also found that self-assessed “knowledge aboei produa class” was positively correlated with the “importance of country of origin.” Heimbach, Johansson, and MacLachlan (1988) examined the relationslnip between familiarity and country-cue usage based on a cue-utilization framework, This framework suggests that the usage of a cue (such as COO) depends on its predictive value and its confidence value. The predictive value Of a cue is the extent to which a con SumCr associates a cue with product quality. The confidence value of the one is the extent to which a consumer has confidence in her/his ability to use that cue (Olson & Jacob y, 1972). The authors suggest that the predictive value 0f the “made-in” label increases with the degree of perceived differences in the quality of the goods each country produces. That is, if a consumer perceives that there are differences between the qualiti es 4l of goods produced in different countries, then he/she will use the C00 label in judging quality since the country cue has predictive value for quality. In addition, they mention that the predictive tendency of the country cue depends on the consistency of product quality among brands produced in the same country. That is, if a consumer perceives that all brands from the same country have invariable quality, the predictive value of the “made-in” cue will be high and this one will serve as a good summary construct. In terms of confidence value, the authors explain that consumers having more familiarity with the product category are more likely to be aware of the quality differences between countries and between brands within a country, and that they tend to be more capable of evaluating the quality of alternative products using the country cue. ' . - 'tude The researchers hypotheSIZed that the extent of C00 cue utilization m at“ . . 1th a judgement about brands is positively correlated With the cue’s Predictive value and W Person’s familiarity with the product class, that is, the cue’s confidence value - They also proposed that the extent of C00 cue utilization is a function of the interaction between the cue’s predictive value and its confidence value. The results of their experiment confirmed these hypotheses. Iyer and Kalita (1997) hypothesized that use of all available information on price, brand name origin, and COM will increase with increasing levels of prior knowledge. In their study, prior knowledge was operationalized as the number of brands of each Product identified from free recall by each respondent. Then they determined three levels of prior knowledge based on these responses. The results showed that while COM cues Were important in evaluating quality at all knowledge levels for all products (Stereos, watches 42 sneakers, and jeans), COB cues were used at all knowledge levels only for stereos and watches. Only jeans indicated increasing cue utilization with increasing levels of knowledge. Price was found to be SCldom used for most products. /7l/he findings reviewed above indicate that the effects of product familiarity on COO are mixed. This might be due, in part, to the fact that in some studies, product familiarity is defined as brand familiarity (Iyer & Kalita, 1997; Cordell, 1992; Han & Qualls, 1985), while in other studies it is defined as country familiarities (Johansson et al., 1985; Heimbach, et al., 1988; Roth & Romeo, 1992). Thus, product familiarity should be redefined in terms of brand familiarity and COM familiarity. Such an attempt will be especially Critical in explaining the differential effect of brand image and COM 0“ consumer purchasing behavior for hybrid products, where the COB and the COM are different. ’/ 43 CHAPTER III CONCEPTUAL FRAMEWORK First, this chapter will propose a new framework of quality dimensions that can be used across product categories based on Nelson’s (1970, 1974) quality dimensions and a critical review of the existing country image Studies provided in Chapter 11. Using the quality dimensions identified, several models are proposed for Korean consumers to examine the unresolved issues in the area of country of origin (COO) for well-known branded products: I) to identify the quality dimensions of product evaluation; 2) to examine the effects of country of manufacture (C OM) and brand image on Product evaluation (cognitive component), product-specific purchase attitudes (affeeuve component), and purchase intention for hybrid products; and 3) the effects of brand and COM familiarity on the relationships among COM, brand image, and product evaluation of hybrid products. The models in the current study are based on the Theory Of Reasoned Action”) (Fishbein & Ajzen, 1975) since this theory explains consumers’ purchasing behavior through the cognitive component (belief-evaluation composite), the affective component (attitudes toward an action), and behavioral intention. In the proposed models, the ‘0 Although Fishbein & Ajzen (1975) originally included normative components I , I ' C motivation to comply, and subjective norm) in the consumers decrsnon-makmg proces (normative belief, excludes these constructs. ' S, the current Stud Normative components in C00 studies are often found in the context of a domestic versus a foreign product purchase situation (Han ’19393 Ettenson eta]., 1933; Heslop et al., 1987- Kim & Pysarchik, 2001; Wall, Liefeld, & HeSIOP, 1989)- These studies employ ethnocentrism oi nationalism as correlates of country effects and attempt to explain the effect or buy-domestic-produets campaigns on consumers’ purchasing behavior. On the other hand, the context of the current study is the situation of hybrid product purchasing, where the effects of normative components stimulated by ethnocentrism or nationalism are expected to be minimal. 44 behavior of interest is a consumer’s intention to buy hybrid precincts including TVs, cameras, and sweaters. Discussion is organized according to the major constructs in the models. After reviewing 861(30th quality dimensions for C00 and brand image effect, the current study proposes a series 0f theory -based hypotheses regarding the causal relationships within the models. Preposed Quality Dimensions across Product Categories As discussed in Chapter II, the quality dimensions proposed in previous studies were found to be highly correlated (Roth and Romeo, 1992; Li & Dant, 1997)- The . . . . . ' to current study pr0poses that the multicollmeanty among the quality dunensrons 15 due . . ' used the fact that the same evaluation mechanism (for example, search or experience) 15 . ’ 'lity to assess related quality dimensions. For example, high correlations between rehab‘ 1997) and performance, and durability and performance as observed by Li and Dal“ ( dimensions might be due to the fact that consumers perceive and assess these quality through the same evaluation mechanism (experience with products). Thus, first, quality dimensions relevant to the C00 literature review are reviewed. Then, these quality dimensions are categorized based on the quality evaluation mechanisms associated with the products’ functional and symbolic aspects. These will be discussed in greater detail below. Quality_Dimensions The quality dimensions were identified in the C00 literature based on the Criteria below, which were adapted from Roth and Romeo (1992, P- 430) The dimensions ( I) Were consistently found in previous research; 45 Table 3. Defini (2) related to Pe‘cepfions Of a country’s production and marketing stre d llgt/Is an weaknesses; . tuall)’ and operationally distinctive; and (3) were concep a d (4) were applicable to a bro range of DI'Oduct categories. Based on the above criteria, the current smd)’ PrOposes six dimensions of quality, design, Perfomance’ Serviceability, brand prestige, and teehnieal including ease of use, prestige- The definitions of these dimemlons and equivalent dimensions in C00 studies are indicated in Table 3. tions of the Quality Dimensions and Equivalent Dimensions in C00 Equivalent Dimensions in C00 Studies Studies Quality Dimensions Definitions . —Ease oi U se Ease of use and operatIOn Design Style, fashion, colors, or varieties Design of a product Performance Excellence and dependability of workmanship or Performance ; the operating features of a product Reliability or Durabilit d Serviceability Accessibility of service center and SCWiCeability y and competence rvice rapidity, courtesy, of repair/maintenance se Brand Prestige Prestigious image stimulated by Reputation; Prestige 1 brand name Statush; Imagei us image stlmulated by Innovativeness or Techn- Icalit Technical Prestigio . use of advanced/hlgh-technology Prestige a Thakor & Katsanis, 1997 b Nagashima, 1970, 1977; c Cartin et al., 1982; Han & Terpstra, d Cattin et al., 1982-, Li &. Dant, 1997 6 Han & Terpstra, 1988 fNagaShima, 1970, 1977 g Ha" & Terpstra, 1988', Roth & Romeo, 1992 i‘ Johansson & Nebenzahl, 1986 {Li & Dam, 1997 J Catfin et al., 1932; Jaffe & Nebenzahl, 198 Roth 8L Rome, 1992 1988; Roth & Romeo, 1992; Li & Dant, 1997 4; Han & Terpstra, 1988; Roth & Romeo, 1992 46 There are several poi ”‘3 that ShOuld be noted. Fir-St, the dime-1mm of serviceability has a different meaning depending upon the product type. That is, serviceability for electronics refers to the acceSSibility and competence of a Service center (Han and Terpstra, 1988), whereas it means the ease of management/care for apparel goods (Li and Dant, 1997). Second, previous studies Of produq innovation or technicality did not analyZe the symbolic (prestigious) image aSSOCiated With advanced/high-teChnOk’gY (Cattin et al., 1982; Jaffe & Nebenzahl, 1984; Han & Terpstra, 1988; Roth & Romeo, 1992). The current Study proposes that the previously used innovation or technicality needs to be interpreted as technical prestige since the use of advanced/high-technology improves the periormance of the product, and accordingly, innovation or technicality is highly related to performance. it should be also noted that although COO researchers consistently identify the dimension of price (economy or expensiveness) or value, as one of the quality dimensions (Nagashima, 1970, 1977; Narayana, 198 1; Cattin et al., 1982; J0hansson & Nebenzahl, 1986; Han & Terpstra, 1988), this dimension is not included in the current study. Li and Dant (1997) are critical of the above studies, because they do not clear] distinguish quality from value and price, although these are three different construct: That is, price is what an individual has to sacrifice to obtain a product (Ahtola, 1984); perceived value is a consumer’s belief in the utility gained relative to What he/she has given up (Zeithaml, 1988), and quality is an individual’s judgment of the excellence of a product (Garvin, 1984). Therefore, price and value are hardly subdimensions of quality 47 according to these definitions. Thus, this Paper excludes price Or value as a quality dimension. Finally, the relative impor‘tance 0f each quality dimension differs based at?“1 the type of product. For example, the design dimension may be less important for electronics than fashion products, since there is 1683 Variation in designs for electronics than for apparel. The ease of use dimension may be less important for products such as earneras having Similar operating features or I V 3 having standardized operating functions. Classification of QWensions Based on VQuality Evaluation Mechanisms Consumers may select a product based upon its symbolic (prestige) or functional aspects (Mittal, 1990; Sirgy, 1 982)- Accordingly, quality dimensions can be dichotomized on the basis of their symbolic and fiInCtional aspects. As indicated in Table 3 COO researchers frequently identify these two types of quality dimensions, although they do not explicitly distinguish between them. Thus, the Present Paper proposeS that consumers perceive and evaluate the quality of a product in two Ways, fimCtionally and symbolically. As indicated in Table 4, the functional aspects of quality are evaluated through the search and experience mechanisms based on the logic of whether the attribute . “1 qUestion ' ' d t or afi - , can be evaluated pnor to purchasmg the pro uc er purchasing and usrng it (Nelson 1970, 1 974). In the search mechanism, consumers evaluate quality dimensions that can 48 Table 4. Definitions of the Quality Evaluation Mechanis Quality Dimensions ms and Co’mSPOn ding / Quality . Evaluation Mechanisms Defin'fions 0W Consumers. quality evaluation Ease of Use Search 9“” 685 acuvated Prior to / Design purchase- Functional , x. Mechanisms Consumers. qUaiity evaluation performance Experience process a(”Waited only after Serviceability purchase and use. WW evaluation Brand Prestige Symbolic process activated to assess the Technical Prestige Mechanisms Image prestige of the image stimulated by brand name or use of Wogy. be evaluated through information seeking and processing, Without purchasing or using the item (Nelson, 1970, 1974). The quality dimenSion assessed through this mechanism is design, since style, fashion, colors, or variety of a product can be examined prior to purchase. In addition to design, ease of use is also judged through the search mechanism as Thakor and Katsanis (1997) propose, although COO studies have not examined this , In the experience mechanism, consumers evaluate quality dimensions that cannot be evaluated unless they Pmcbase and use the item (Nelson, 1970, 1974) Th - at is, through the experience mechanism, consumers make inferences about quality based on intrinsic and extrinsic cues at the time they purchase a product, whereas in the search mechanism they can make evaluations at the point of purchase. The quality dimensions eVall—lated through the eXperience meChanism are product performance and serviceability, since the excellence and dependability of a product’s operating features as well as the 49 aCCeSSlbmW and competence Of a SCI’Vlce center 01‘ the ease of pro duct care can only be evaluated after experiencing the PrOduCt 01' Service. Finally, the Symbolic aSPeCtS of quality are evaluated through the image mechanism. Consumers attach symbolic meaning to products, since product consumption is social behavior. Such SymbOIic meaning is Ofien referred to as “13“)th image” or “brand image” (Assael, 1992; MOWen’ 1990} Many COO researchers have focused on the prestigious or upscale image of the PI‘Oduct, which is stimulated by the brand name and timber enhances the reputation of the product (Nagashima, 1970, 1977; Han & Terpstra, 1988; Roth 8; Romeo, 1992; Johansson & Nebenzahl, 1986', Li & Dant, 1997). Further, as mentioned above, the current study proposes that the prestigious image of a is also stimulated by using advanced/high-technology. Thus, the quality product dimensions evaluated through the image mechanism are brand prestige and technical prestige. in the following section, a comprehensive model of the differential effects of brand and COM on consumers’ purchasing behavior Wi” be discussed based on the quality dimensions identified above. Comprehensive Model of the Differential Effects on Consumer Purchasing Behavior of Brand Image and COM Comprehensive models based on the Theory of Reasoned Action (F‘ hb IS ein & Ajzen, 1975) are proposed to explain the relative importance of COM and brand ‘ Image on COnsumers’ perceptions 0f quality, P‘mhase attitudes, and intehtion to buy thr ee products, a TV, a camera, and a sweater. (See Figure l .) As described below FiShbein 50 and Aizen (197311 propose that behavtoral intention 1 d ea 5 to behavior ) and that behavioral intention is determined by the Consumer’s tt' (1 a nu es (in fav 01' 0f 01- against) toward PUIChasing or using a bran 6- The attitudes toward purchasing of “Sing a brand . iii are, In turn, a notion of the cognitive belief n i ,- . The c - - structure (2 B E ) OgmtIVe belief structure is a belief-evaluation i=1 compOSIte where B,- is the belief that Performance of the behavior Will lead t O a SpeCific outcome ° ' ° , 1, and E,- IS the evaluation of eaCh consequence and n is the number of sal‘ 3 lent outcomes. ponent in the model 0f Fishbein and Ajzen (1975) gives The cognitive com ers’ overall quality perceptions since it is the total set of their information about consum the evaluations associated With those beliefs A . 1 y . s prevrous salient beliefs and searchers have found that the ef fects of CO rand are qualltl' mentioned, COO re (Han & lerpstra, 1988' Li & Dant 1997 ’ 9 ; Th akor & Katsanis, 1997). dimension specific Therefore, the cognitive component in the Fishbein and Ajz en (1975) - model 18 decomposed into two constructs: quality dimensions and 0Verall qual'ty 1 . (See F’ Igure 1.) ___________________...—— mentioned, this study excludes the normative value or sub' t' 160 we Dunn k I F'shbein and H AS previously AJzen (1975) propose that the subjective norm is a function of a normative Sm tur Z c e( NB .M ' on that referent j thinks the ind' - H J C1 )’ Where er’s motivation to com 1 'th Indua‘ 8110“” 0‘" shOuld P Y WI referent j, and k is the milletb perform the m er 0f salie nt N5? is the perceived expectati of behavioral intention a on w' h att m l g It i de toward the a ct. bethavion MC]- is the consum re erents. Subjective norm is an antecedent 51 35 £83 All 3:52: 323:..— _ m u m .{ 3250i .20.... can 22m *0 38 855:3. gag—0.59 . . . . . . - . _ c . . . . . . . . . . _ . . . . . . . L. 25»: .3 55:2:— EE 02 i ....... . £283 5 Eggs .6: .2 . esomooeom a 2 mm ememoQa u. -e s s \ \ £25 =Eo>O a 25:28 :83: 5 .o .232 2355358 < . p 8.5.5 mEmmp—uE 05:8 «5 um 35: 2 S In this model, the quality eValuation mechanisms (search, 3"{Perience and . ) 2 Image are used as explanatory variables due to a statistical limitation o f Sanctum] eq ti ua on modeling (SEM)12. In the following section, each relationship in the pr0posed model . 3 IS discussed. The Relationships among COL/12ml Image, and Quality Dimensions \ COM, Brand Image and Search DimenSAions (Design and Ease W Perceptions of product design or ease of use are formed through information seeking and processing and can be easily assessed at the point of Pquhase. If consumers can assess quality first-hand, then they will rely less on other eXtrinsic cues, especially brand name, COO or COM (ThakOI' & Katsanis, 1997)° Therefore, COM would not be expected to impact the perception 0f the deSIgn dlmenslon Since consumers can assess style. fashion, or colors visually or manually or try on the Product. For example, consumers can assess the style, fashion, or colors of Ralph Lauren sweaters manufactured in developing countries like Mexico by examining these qualities before purchasing and wearing them. Brand image, however, is expected to have some influence on COns e rs’ quality perception (Thakor & Katsanis, 1997). The logic is based on the fact that through strict quality control, packaging and advertising, a company intensively deV§ lopS its brand 12 In SEM, the quality evaluation mechanisms are second-order factors, since multichlinean'ty among the quality dimensions suggests the existence of other latent variables, Wthh. are the three quality evaluation mechanisms in the present study (BOllen, 1989)- The PTOW§?d models "1 th is paper, however, are constructed in a way that these higher—order factors are hard to empirically test using S EM. That is, in order to be identified in the SEM, the six quality dimensions should be influenced only by the second-order Quality dimensions factors ("dilation mechanisms). However, the models are designed in a way that the are influenced by COM and brand image as Well as the second order factors. Thus, these models cannot be identifiedm the SEM' 53 image (Negley, 1999) and thus, brand names become a signal of I-~1110bservable product quality (Rao, Qu, & Ruekert, 1 999). Therefore, even though cc- lisumgrs can directly search some product dimensions, the brand image may influence their judgments of thes e dimensions. As previously mentioned, the ease of use dimension may be more important for products having sophisticated features- The current study employs Gaineras with an automatic focus, TV 3 having a standardized operating system, and cotton SWeaters with no Operating system. Therefore, the case Of use dimension iS less important lwhen evaluating product quality Since there is little variation in the way they operate across the brands Within the product category - Thus, the case Of use dimension is excluded in the present study. Further, the design dimension is examined only in the case of sweaters, since this dimension is assumed to be less impOrtant in the quality evaluation than other dimensions for TVs and cameras. Focus group respondents did not mention it when they were asked to identify what the salient attributes were when they purchased TVS and cameras. Based on the above arguments, the following hypotheses are sed for propO sweaters: Hl a: There Will be no difference between the design evaluation of a product manufactured in a deveIOped country and that in a developing counhfy. Hlb: Brand image and design evaluation will be positively related. COM, Brand Image and ExperieMnsions (Performance and 86W Since the evaluations of experience dimensions, including Performance and serviceability, are fonned after the product has been purchased and used (Nelson, 1970; 1974; Thakor & Katsanis, 1997), consumers may infer the qualities Of these dimensions 54 based on extrinsic cues rather the”jul intrinsic cues at the point of purchase Thus th fl. ' 9 e C ect of extrinsic cues such 35 CON and brand image may becorn more important ' e in consumers’ perceptions of the experience dimensions rather than the search dimensio ns. Many previous COO studies have found that a product manufactured ' In a developed country is rated higher in quality than one made in a developing country (S . ee Liefeld, 1993; Samiee, 1994; Verlegh & Steenkamp, 1999 for a Comprehensive review ) It is also well known that brand image and PrOdUCt quality are p OSitiVely related (Assael 1992). The current study, however, further proposes that brand image has a stronger effect on the functional dimension than does COM. AS mentioned above, this logic is based on the fact that a company invests in imprOVing brand name recognition and in establishing favorable brand image rather than in boosting the image of the country where the product is manufactured (Tse & Gom, 1993). Country images associated with COM are formed in the consumers’ mind through personal experience (for example, study and travel); knowledge of the country’s political Status and econom.lc developments, and such; or experience with a product from the country in question (Samiee, 1994). Thus, brand name provides a customer with more readily recognizable information about a product than does COM (Ahmed & d’Astous, 1996). 'The following hypotheses are Pmposed for the experience dimensions. For TVs, cameras, and sweaters: The performance evaluation of a product manufactured in a dCVCIOped country will H2a: . be higher than that in a developing country. . . H21): Brand image and performance evaluation will be posmvely related. H20: Brand image Will have a stronger effect on performance evaluation than will COM. H3a: The serviceability evaluation of a product manufactured in a developed country Wt" be higher than that in a deVeIOping country. . . Brand image and serviceability evaluation will be posrtlvely related. H3 b: 55 H3c: Brand image will have a Stronger effect on serviceability COM. evaluation than will COM, Brand Image, and Image Dimensions (Brand Prestigeand Te We) Consumers buy many products because of their prestigious image bol sym s (Eastman, Goldsmith, & Flynn, 1999; Kirmani, 800d & Bridges 1999) The ' a . Image dimensions represent the symbolic quality of the product, that is, the Prestigious i f mage o the product resulting from the brand name or the use of advanced/high techno] . ogy 1n manufacturing the product. Regarding the brand prestige dimension, if a consumer has a positive impression of a brand name, the product is assumed to have a prestigious brand image, since this image may deliver the promise of prestige as well as quality and value to consumers (Anonymous, 1998) Thus, it iS prOPOSGd that a prOdUCt’S PreStigious brand image is positively related to its overall brand image. On the other hand, the effect of COM on brand prestige is also proposed to be Significant, as Ahmed and d’Astous (1996) found that a brand’s Quality image decreases if it is assembled in a less prestigious country. The current study, however, propose;S that the effect of brand image on the brand-prestige dimension is greater than that of COM on brand prestige since brand name contributes more to the prestige of the brand than does COM. The present paper also proposes that COM will have a significant influence on the perception of the technical prestige. That is, if a product is manufactured in an 56 advance d/high-tech comm”, it is assumed that this product has a Prestigious image stimulated by using advanced/hi gh technology. Johansson and Nebenzah] (1986) found that a Product’s status image (aSSOCiated With pride 0f OwnerShip and Style) for Passenger cars (Buick, Chevy, Honda, and Mazda) was diminished by moving PI'Oduction to low- wage countries (S. Korea, Mexico or the Philippines). Thus, it can be inferred that developed countries will be perceived as having more advanced/high tech products than developing countries, and products manufactured in developed Countries will have more prestigious images than those made in developing countries. Further, brand image is proposed to have a POSitiVe influence on a consumer’s perception of technical prestige. This proposition is inferred fi‘Om findings of studies (Ulgado & Lee, 1993; Tse & Lee, 1 993) that uncovered a significant interaction between brand and COM. That is, if a PTOduCt has a favorable image bOosted by a well-known brand name, the unfavorable technical images aSSOCiated With less prestigious countries (such as developing countries) could be overcome. The current Study, however, Pmposes that the effect of COM on the technical-prestige dimension is greater than that of the brand image effect since technical PreStige is determined by the use of advanced/mg” technology. Based on the above arguments, the following hypotheses are proposed. For TVs, Cameras, and sweaters: H4a The brand-presti e evaluation of a product manufactured in a devfil oped country will be higher than that in a developing country. . . H4131 Brand image and brand-prestige evaluation Will be posmvely related - ”The current Study limits the scope of the prestige dimension to that associated with an advanced/high tech country, since mOSt previous COO studieS’ main interest is in nonfood products. .If a product is food lated with the place (either P need or processed), the prestigiOUS image of the product will be assoc famous for good food products (for example, the champagne produced m France W1" have a prestigious {triage}. 57 H4c: H5 a: H5b: H5c: Brand image will have a SHOnger effect on brand prestige than Will COM The technical prestige 6Valuation of a product manufactlJred in cOuntry will be higher than that in a deveIOping country. a deveIOped Brand image and technical prestige evaluation will be positively related COM will have a stronger effect on technical prestige than Wi 11 brand image The Relationships among COMfld Image, Quality—Dimensio Quality Perceptions ns and Overall WSions and Overall QualitLPerception The relationships between each quality dimenSion and overall quality are proposed to be positive since consumers use multiple quality attributes when evaluating overall QUality (Fishbein & Ajzen, 1975; Steenkamp, 1990), Although there is no empirical research examining these relationships, several C00 researchers (Han & Terpstra, 1933-, Li &, Dant, 1997; Roth & Romeo, 1992; Thakor & Katsanis, 1997) consider overall quality to be composed Of the subdimensions of quality. (See Table 3_) Therefore, a direct relationship between each quality dimenSiOnM and overall quality is proposed as follows: H6a: H6b: H6c: H6d: H66: Design evaluation and overall quality will be positively related, Performance evaluation and overall quality Will be POSitiVely related. Serviceability evaluation and overall quality Will be POSitiVely related - Brand-prestige evaluation and overall quality will be positively related. Technical prestige evaluation and overall quallty Will be positively rel ated- COM, Brand lWrall Quality Perception Many COO researchers have reported a significant impact of COM on overall quality. (See Samiee [1994] and Verlegh & Steenkamp [1999] for a comprehensive review.) Especially in the context of hybrid products, both COM and brand have been 58 reported to have influence on the Overall ‘quality perception when only brand name and country cues are provided (Han 8c Terpstra, 1988', Tse & Gorn , 199 3), That is, products from developed countries are perceived to have higher qual ity than those fro In developing countries, and products with wen-known brand names are perceived to haVe higher quality than those with unknown brands. The effect of COM, however, w as found to have no impact on the overall quality perception when other intrinsic information was provided, While the effect of brand remained strong (Ulgado 8‘ Lee, 1993). Similarly, using tangible products, Nes and Bilkey (1993) found that while both COM and brand significantly affected product quality perceptions, brand had a StroIlger effect. Thus, it is clear that when intrinsic information is provided in addition to brand name and COM cues, the effect of COM lessens, while that Of brand image remains influential. Therefore, when consumers are provided with intrinsic information about the product, the following hypotheses are proposed: H7 a: COM will have no relationship to overall quality. . H7b: Brand image and overall quality perception W111 be pOSltiVely related. The RelationsthM, Brand Imager Overall uali Perce tion and Purchase Attitucfi§ Overall QualitxPechWe Attitudes Personal attitudes toward the behavior (affective process) refers to whether the person is in favor of or against performmg the behavior in question (Fishbein & Ajzen, 1975). Thus, attitude formation is the affective process in consumers’ décision making- W W MAS previously mentioned, for TVs, cameras with automatic focus, and cotton sweaters, there is little variation Hi the way they Operate across the brands within the product category. Thus, the ease of use dimenSiO" '5 e)(oluded in the present study. 59 The Theory of Reasoned Action (FiShbein & Ajzen, 1975) suggests that attitudes toward an action are distinct from attitudes toward an object. That ' IS, the ficus is on measuring the attitudes a consMer has toward buying a PYOdUCt and “Qt her/hi S attitudes to d th war e Product itself. Shimp and Sharma (1987) emphasize that the attitudes measured must directly match the action being considered. Various studies examining the efficacy Of this theon ham: Confirmed the positive relationship between cognitive belief structure and attitudes (Knox & Chernatony,1989; Lee & Green, 1991, Mitchell & Olson, 1981; Netemeyel' & Bearden 1992; Shepfiafi", Hartwick, & WarshaW, 1988). Lee and Green (1991) also found that cogmtrve structure was a significant predictor of attitudes for bOth the US and Korean samples Therefore this study proposes that the higher an individual’s overall quality perception of a product the more favorable his/her attitudes will be toward the act of purchasing HSa'. Overall quality perception and purchase attitudes Will be POSitively related COM, Brand image, 3“de5: Attitudes Obermiller and spangenberg (1989) propose that C00 is not likely to trigger an affective process, that is, attitudes toward the behavior, although some Studies found that such an emotional reaction occurs when C00 is the only cue provided (C Q0 stereotypes in the studies of Bannister & Saunders [1978], and Reierson [1967]). In fact, in the situation where multiple cues are provided, two studies, Erickson et €11. (1934) and Johansson et al. (1985) report no effect of C00 (where the brand and manufacturing country were the same) on attitudes, although they found a significant effect of C00 on belief (quality perception). In both Studies, attitudes were significantly related to belief 60 Sin (:6 these researchers studied only uninational products, COO was equivalent to C 0M in these studies. Thus, thC COM 0f hybrid PTOducts seems to hav ' act on 8 I10 imp consumers’ attitudes. . ’ - - rce tion of , . Brand Image is a total p6 P tl'le brand that is obtained by processmg .1 ormation from various sources over time (Assael, 1992), Thus, it is expected that if a consumer has a favorable brand image, then he/She Will have a positive attitude tOWai-d buying the brand, while if s/he has an unfavorable brand image, a negative attitude toward buying the brand will be formed. Therefore, the following hypotheses are proposed: H8b'. COM will have no relationship to purchase attitudes. H8c: Brand image will have a pOSitlve relationship to purchase attitudes. The Relationships amonfg COM: Brand Irna e Purchase Attitudes and “denim“ to Buy Purchase Attitudes and Intention to Buy A consumer’s intention to buy is the ultimate dependent Vari' able 1' . . “ . n the current study, Fishbein and Ajzen (1975) define intention as a special case of belief . . . S in Which the object is always the person himself and the attribute is always a behavior» (p 1 . . ' 2). A consumer’s intention is formed by her/his evaluatmg the behavior and a” Coating subjective probability that he/she Will perform that behavmr. Fishbein and AjZen (19 75) , - ' to urchase serves ‘ - $t0‘90se that consumers intention P as a link between their attitudes toward buying products and their actual 1313011333 or use Of the Products. Intention to bu y has been frequently used as an alternative measure to purchasing behavior in PYCVious Smdies (Wilson, Mathews, & Harvey, 1975). 61 F ishbein and Ajzen (1975) posit that attitudes toward an action directly detenmhe par Chase intention. That is, the more favorably the consumer views engaging in a certain action, the more likely he/She is to have an intention to engage in that action. Several researchers have confirmed the positive association between attitudes and intention in the Theory Of Reasoned Action (Knox & Chematony, 1989; Lee & Green, 1991', Mitchell & Olson, 1981; Neterneyer & Bearden, 1992; Sheppard etal., 1988). This relationship was also found to be significant among Korean consumers in a cross-cultural study Of the ' 't've relationship Theory of Reasoned Action (Lee & Green, 1991). Therefore, a poSii between attitudes toward buying a product and intention to buy the product is proposed. H9a: Attitudes toward purchasing a product will be positively related to the intention to buy the product. COM, Brand Image, and Intention to By The effects of C00 and brand on purchase intention have been ex . ed 1688 ' ' act on uali erception. Es . frequently than me“ “up q W p peclally, few studies have examined the effect of COM on purchase intention in the conteXt of . , , hybnd Products, as indicated in Table 1. Past studies indicate that the effect of C00 on Purchase [me t. . n 1011 is limited (Ettenson, et al., 1988; Schooler & Wildt, 1968). in the study of w . . all et a1. (1991), the effects of C00 and brand on intention to buy are also found to be . weaker than their effects on quality perception Further, the effects of C00 and brand on ' ' ' . B th COO - . purchase intention are product SpeCifiC 0 and brand had SlgmfiCant e ffects on purchase intention for shirts, but no Influence was found for telephones and Wallets, These studies, however, did not include attitudes as another antecedem of Pm‘Chase intention as proposed by Fishbein and Ajzen (1975)- “1118, the finding of a significant 62 effeCt 0f C00 and brand image on purchase intention is doubtfiil. Moreover, willingness to buy is found to be closely related to the value of the product. That is, the trade-off between perceived quality and the monetary sacrifice may be another important factor in determinining purchase intention (Dodds, Monroe & Grewal, 1991; Monroe & Krishnan, 1985; Szybillo & Jacoby, 1974; Iyer & Kalita, 1997). Therefore, the following hypotheses are proposed: H9b: COM will have no relationship to intention to buy. H9c: Brand image will have no relationship to intention to buy. Models of the Differential Effects of COM and Brand Image on Quality Perceptions of Hybrid Products Using Brand and COM Familiarities as Moderating Variables COO studies found product familiarity to be one of the customei'ie‘ated influences that may have impact on consumers’ use of C00 as a purchaSing cue (Heimbachet et al., 1988', Johansson et al., 1985; Johansson, 1989- Sam' 1 Th ’ ‘33, 994). US, the current study proposes that product familiarity, later defined as brand f amili'ari and COM fafniliarity, iS a moderating variable that identifies the diflrerential eff ty . . eCtS Of bf image and COM on quality perception for hybrid products. In the follow,- and n secti brand and COM familiarities are defined, and then the causal relationships Ons, aInong these variables are discussed. Definitions of 131-Wand COM Familiarity In previous studies, product familiarity is defined as the “number of prod 110t- related experiences that have been accumulated by the consumer” (Alba & Hutchinso 11, 63 1987, p. 411) and is measured using items such as “previous ownership” (Bettman & Park, 1980; Johnson & Russo, 1984). This approaoh emphasizes actual experience with the product. Knowledge of the product, however, could be obtained without direct experience, through information search, word‘Of-mou’th from people close to the consumer, or media advertising exposure (J ohansson, 1989). Product familiarity is also Viewea as “the indiVidual’S prior knowledge level with respect to the brands in a product class” (p. 50) and is measured by using SCH-reported rating scales (Johansson, 1989) or by items such as number of brands he/she recalls in a product class (Iyer & Kalita, 1997). This approach implies that product familiarity is determined by brand familiarity. Such an approach, however, is insufficient to explain the effect of product familiarity on C00 usage for hybrid products. For uninational products, brand familiarity implies COO familiarity, since brand nMe is strongly associated with C00. For example, if a consumer knows SONY well, he/she knows that it is a Japanese brand. On the other hand, for hybrid products, COM is not the same as country of brand (COB), where the company’s headquarters are located. Thus, brand familiarity does not necessarily mean familiarity with the country where the product is manufactured. The current study, therefore, will specify conventional product familiarity as “brand familiarity” and “COM familiarity.” /érand familiarity is defined as the Consumer’s prior knowledge of the brand in a product class. Thus, brand familiarity is specific to a particular product. On the other hand, CO familiarity is defined as the individual’s prior knowledge of products made in a particular country. 64 Based on brand familiarity and COM familiarity, this study proposes four segments of consumers: high brand familiarity/high COM familiarity, high brand familiarity/low COM familiarity, 10‘” brand imi‘iafity’high COM familiarity. low brand fannliarity/low COM familiarity. The differential effects of brand image and COM on quality perception are proposed for each 0f the four segments. Quality Perception is decomposed into two levels, including quality dimensions (design, performance, serviceability, brand PieSiige’ and teChnical pr eStige) and overall quality perception. Quality dimensions (See Table 3) are employed to examine the effect of brand image and COM on these four consumer segments. Discussion of the causal relationships between each quality dimension and overall quality perception is not included in this section, Since this has already been examined in the previous section. Moderating Effects of Brand Familiarity and COM Familiarity on the Relationshig among Brand Image, COM, Qualifi Dimensions and Overall Quality Perception Johansson (1989) proposes that if a consumer has knowledge of a country’s products, he/she will use the C00 to evaluate products and brands. He explains the relationship between product knowledge and use of C00 as similar to that of product knowledge and use of brand name. That is, when someone buys a product on the basis of brand name alone, it seems reasonable to assume that he/she thinks he/she knows something about the brand and is familiar with it. Similarly, a consumer can obtain knowledge of a country’s products in various ways, such as product experience (trial), information search, word-of-mouth from people with whom he/she has close relationships, or media advertising exposure. If knowledge of a country’s products (for 65 example, whether or not the 001111th is a high-quality manufacturer Of a specific product) is relevant to the task at hand, he/ she Will use the information to purchase the pI'Oduct, Thus, it is logical to assume that the more a consumer knows about the country’s products, the more likely he/she will be to use the C00 cue to evaluate a product. Likewise, it is expected that the more a consumer knows about a brand, the more likely he/she will be to use the brand name when purchasing a product Johansson’s (1989) proposition is consistent with the confidence Value prOposed by Heimbach, et al. (1988). That is, if consumers are aware of the quality differences between countries and between brands within a country, they are more Capable of evaluating the quality of alternative products using the country cue, Therefore, for consumers with high brand familiarity and high COM familiarity, brand image and COM will be used to make judgments on the five dimensions of quality (design, performance, serviceability, technical prestige, and brand prestige) and the overall quality perception. For consumers with high brand familiarity and low COM familiarity, only brand image will have a significant impact on the evaluation of the quality dimensions and overall quality. Conversely, for consumers with low brand familiarity and high COM familiarity, only COM will have a significant impact on the evaluation of the quality dimensions and overall quality. Finally, for consumers with low brand familiarity and low COM familiarity, brand image and COM will have no impact on the evaluation of the quality dimensions and overall quality. For these consumers, other intrinsic or extrinsic cues, such as price, will be more important. Based on the above arguments, the following hypotheses are proposed: (See Figure 2-a to 2-d for each segment.) 66 H10a: HlOb: HlOc: H10d: Hlla: Hllb: Hllc: Hlld: For consumers with high brand familiarity and high COM familiari . . . . ’ b image and COM Will have a Significant Impact on the eValuation 03'1“: rauigi dimensions. q ty For consumers with high brand familiarity and low COM famili ' . . . an image will have a Significant impact on the evaluation of the qualityy 21323132? 3. For consumers with low brand familiarity and high COM famil' . will have a significant impact on the evaluation of the quality direfiizglsy COM For consumers with low brand familiarity and low COM falniliarity and COM will have no significant impact on the evaluation of the bfmd image dimensions. \ quality For consumers with high brand familiarity and high COM fantiliaiity b d , ran image and COM will have a significant impact on the overall e . quality. “"3"” of For consumers with high brand familiarity and low COM familiarity on] b image will have a significant impact on the overall evaluation of quality Y rand For consumers with low brand familiarity and high COM familiarity only COM will have a significant impact on the overall evaluation of quality. ’ For consumers with low brand familiarity and low COM familiarity, brand image and COM will have no significant impact on the overall evaluation of quality. 67 .mcoumoam ecu Bu: 2: ~ . 8: mm bEnmooEom e .m>._. ecu £558 .8 338:. no: a 56566 :wmmoa a lJ cacao; 131—soak $58.5 i omen: 6:95 200 > cm. as o :9: can >~_._m___Em.._ =2 :9: 5“? “Hawk“? .2 82.233. 830““; cm :8 es .255 .3 335 353% 2: e 332 < .3 2.5: 68 .meoumo a com BEBE 8: M. b:5moo§.~om . e i 0 cm éees 5.. 832: s: a 5.22% and . f. $58.5 552m owe—E 652m KOO-Nao 11111 11 1 =205 3532?; .. ‘1 11 11 ... 11111 111H/ _ l/x, . 1111‘ M . m m u. a_e___s£ 5.8 :3 2e a_.e___s£ 2.2m =9: 5:5 «35:88 .8 mcoemsgm 6:32“. .5 .28 Ba 2.2m .0 seem Eases 2: no is: < .3 23.... 69 £53026 .5.“ 86205 “on mm bzfimoomtom e .mxfi. can 8.588 how noes—05 Ho: mm comma—0E6 cmmmoQ e l'I‘---‘ Id'-"' ' 'I'll---'--""--'I-"'-Iu'-'ll"ll-"||"'-Ilnl'-"'. ' ' ' |“'|I 'l-‘ | -- -"'||I‘ '--- " ' guano...“ $38.5 win—m ----------------------------fi ‘. £55 .1330 ESE—ooh. e_._e___§”_ :8 3: use 255 26.. 55, 222388 .8 mac seam =0 .28 25 235 no 38% .mzceota 2: l I one: 2.2m \ \ \Q €2.23 a. 523:. e .282 < .o.~ 2:9”. 70 £55 .1230 + r----------‘------------------ $38.5 18:585. $68.5 33.5 I \\ I \\ \ I \ \ \4 11 \ \ \\ 1 1 s x \ 1 x x k 1 > . . 1 s 1 _ . 1 s. NQO—EQ . , a o I 5 x . 4 . . 1 \x A 1 I s 1 I \ s I I ’\ I I V1 s 1 \ I \ I I \ I\A I I \ I 1 1 \ \ II I 1 \ s I 1 1 \ \ 1 1 1 \ \\ ,5 :m. Ea 26 use 5:2 E5 :0 22m 26.. 53> «weargmojwwmcommagm .0:—25“.. .200 can 32m .3 1.88m Engage 2: .0 Enos. < .u.~ 2:9... 1 1 11 1 111 1 1 11 ilk. uuuuuuuuuu / ”My“ ,. E5 6525 1 1 \\ \\ I \\ \\ 1 \ s I 1 \ \ \ I 1 \\ \ s I I\\ \ \ 71 CHAPTER IV RESEARCH METHODS The models described in the previous section have been proposed 1) to identify the quality dimensions 0f PTOdUCt evaluation; 2) to examine the effects of country of manufacture (COM) and brand image on product evaluation (cognitive component), product-specific purchase attitudes (affective component), and purchase intention for hybrid products (behavioral component); and 3) the effects of brand and COM familiarity on the relationships among COM, brand image, and product evaluation of hybrid products. in this section, a discussion of research design is presented, followed by a discussion of sample selection and data collection. Then, the questionnaire items that measure each of the major constructs in the proposed models are discussed. Finally, techniques of data analysis to estimate the proposed models are provided. Research Design There were two phases to the study, the qualitative phase and the quantitative phase. In the qualitative phase, focus group interviews were conducted with Korean students in the US. and others in Korea to explore the conceptual meaning and cultural context of the target concepts of the study (Confucian values; ethnocentrism; and product beliefs, evaluations, purchase attitudes, and purchase intention). Salient product 72 features, brand names, and COMs ‘5 were also identified in this stage. Information gathered from the qualitative phase was incorporated into the development of the quantitative survey instrument. The resulting quantitative study was a cross-sectional Study using a between-subject-experimental design. The design used a self-report survey instrument to collect consumer behavior data from native Koreans. A cover letter explained the purpose of the study and directions for completion of the survey. Procedures of the quantitative study are explored in the following section. Sample Selection and Data Collection: Empirical Study Korean consumers living in Seoul, Korea and several satellite cities constituted the sample in this study. To assure the inclusion of a broad spectrum of geographic and socioeconomic groups, potential survey areas were thoroughly reviewed and finally selected. Greater Seoul could be subdivided into six different areas based on the socioeconomic status of the residents and described as: southeastern Seoul, which is mostly upper socioeconomic status; northern, eastern, central, and southwestern Seoul, each of which is mostly lower to middle socioeconomic status; and the western areas, which are mostly lower socioeconomic status. The satellite cities outside of Seoul included in this study are: Bun Dang, a middle to upper socioeconomic status city; Kwa Cheon, a middle socioeconomic status ”According to country-of-origin labeling regulations, indication of COM for imported products and the names of manufacturers and unporters are mandatory (The Office of Customs Administration, 1991). Thus, in the Korean market, only COM information is available to Korean consumers, although the various COOS mentioned above are available in the United States. Therefore, the effects of brand image and COM on Korean consumers’ purchasing behavior in regard to global products will be the foci of this study. 73 city; 11 San, a lower to middle socioeconomic status city; and two other mostly lower socioeconomic cities, An San and Kang Maung. Within each area, apartment complexes ranging from twenty to one thousand units were chosen to be included in the sample. Seven research assistants were selected from a major Korean university to collect data. Following extensive training to ensure consistency of administration and breadth of distribution of the questionnaires, each interviewer was assigned to a specific survey area. Once apartment complexes in each area were identified, the research assistant visited each unit and made contact with residents. Where necessary, the research assistant obtained permission from the apartment maintenance or security office to enter the complex. This procedure was required primarily in the upper socioecomic areas, such as Kang Nam gu. Every third unit was approached to participate in the study. When a resident answered the door, the research assistant introduced him/herself and briefly explained the purpose of the survey while showing the resident the cover letter that explained the purpose of the research (see Appendix I). If the resident agreed to participate, the interviewer left the questionnaire along with a token gift of an oven mitt. Before leaving the unit, the assistant secured the participant’s phone number and address, and arranged to phone for a return visit within three to seven days to retrieve the completed survey. The assistants collected data over a five-week period during June and July 1997, prior to Korea’s economic crisis. Research assistants distributed seventy to eighty questionnaires each via this method. A total of 550 questionnaires was distributed with 456 returned, yielding an 82 percent response rate. 74 Instrument As previously indicated, three target products and corresponding brand names were selected based upon the outcome of focus group interviews. Korean undergraduate and graduate students studying in the US. and Koreans living in Korea were involved in 561331 ate focus groups and in-depth interviews. From these discussions, the product categories 0f clothing and personal consumer electronics were identified and further narrowed down to three specific products: a camera, a TV, and a cotton sweater. These products met the criteria established by prior research suggesting that items should be 1) bi-national, and 2) familiar to the average consumer (Han & Terpstra, 1988). In the focus group interviews, discussion also ensued about brand names, country of origin (C00) and country of manufacture (COM). Participants cautioned that the brand names, C005 and COMs used in the study must be recognizable and realistic to Korean consumers. Thus, the following three products and corresponding brand names were selected: a Canon camera (imported brand), LG (Gold Star) TV (domestic brand), and Ralph Lauren Polo sweater (imported brand). The three target products were integrated into a hypothetical point-of-purchase Shopping scenario format within the self-report written survey instrument. The Korean consumers were asked to imagine that they were visiting different stores in Korea to compare various brands, features and prices for each of three products (cameras, TVs, and sweaters). Thus, subjects were presented with information about each of the products in a realistic manner, similar to a point-of purchase sign next to a product in a store. No specific attention was drawn to COM, brand, and such. This information was 75 randomly presented to avoid bias. The subjects were then asked to respond to questions about each product. Once developed in English, the instrument was translated into Korean by Koreans not affiliated with the project using a double-blind translation procedure to achieve construct equivalence. The Korean version of the instrument was pretested on groups of undergraduate and graduate students in the U .S. and in Korea. In addition, researchers at two major Korea universities again reviewed the Korean version of the instrument and revisions were made to improve question clarity, comprehension and readability. Original questionnaires included all three products. Pretest respondents, however, indicated that the instrument was too long and recommended that only two products be included in each questionnaire. As a result, four parallel survey instruments (A1, A2, B 1 , and B2) were developed so that each subject only responded to questions about tWO products. As indicated in Table 5, Al and A2 included questions about a camera and a TV, and Bl and B2 included questions about a camera and a sweater. Al and A2 contained identical hypothetical Korean market scenarios about the products specified, except the COM information was manipulated’._,—-’I‘hat is, Al examined a Canon camera manufactured in China (binational match) and a LG TV manufactured in Korea (uninational match), while A2 examined a Canon camera manufactured in Japan (uninational match) and a LG TV manufactured in Malaysia (binational match). Similarly, Bl examined a Canon camera manufactured in China (binational match) and a Ralph Lauren Polo sweater manufactured in the US. (uninational match), and 32 examined a Canon camera manufactured in Japan (uninational match) and a Ralph 76 Lauren Polo sweater manufactured in Mexico (binational match). Thus, in this experimental design, COM was the only cue manipulated. Table 5. Descriptions of Questionnaire Formats Instrument Al Canon camera (Japanese brand) manufactured in China LG (Korean brand) TV manufactured in Korea Instrument A2 Canon camera (Japanese brand) manufactured in Japan LG (Korean brand) TV manufactured in Malaysia Estrument Bl Canon camera (Japanese brand) manufactured in China Ralph Lauren P010 sweater (American brand) manufactured in the US. Instrument BZ Canon camera (Japanese brand) manufactured in Japan Ralph Lauren Polo sweater (American brand) manufactured in Mexico i j All of the measures used in this study, including the measures of beliefs, evaluations, purchase attitudes, and behavioral intention (Fishbein & Ajzen, 1975), were previously established. The measures of familiarity with brand and COM were adapted from previous studies (Han & Terpstra, 1988). These constructs were measured using seven-point Likert-type scales. When responding to Likert-type surveys, Asian pOpulations are thought to have central tendency, which refers to a prediSposition to answer suffey questions using the midpoint or midsection of a response scale (Si/& '- "L Cullen, 19%),- Thus, Korean researchers recommended that, for Korean respondentsra seven-point scale might provide finer gradations of responses than a five-point scale (Lee, 1 997 — personal interview; Hwang, 1997 — personal interview) Evaluations (E): Prior to exposure to the hypothetical buying scenarios, subjects’ evaluations of product attributes were measured by asking respondents about the 77 goodness/badness of the seleCted attributes when they shopped for each of the target products. Subjects were asked: “When purchasing any product X (camera/TV/sweater), how good or bad is it that the camera has each of the following features...” The attributes for each product are provided in Table 6. The question was measured on a seven-point ‘Likert-type scale (“1 = extremely bad to 7 = extremely good”). Brand Image: The brand image was measured by the item, “what is your general impression 0f the brand X IDl'Oduet X (Canon camera/LG TV/Ralph Lauren Polo sweater)?” on a seven-point Likert-type scale (“1 = extremely bad to 7 = extremely good”). Beliefs (B)--The respondents were asked to rate the likelihood that each of the products (camera, TV or sweater) would have the salient attributes described in Table 6- Subjects answered the following question: “How likely is it that the brand X product X (Canon camera/LG "TV/Ralph Lauren Polo sweater) described above would have the following characteristics. . .”, measured on a seven-point Likert-type scale (“1 = extremely unlikely to 7 = extremely likely”). Attitudes toward the act (Am): Respondents were asked to rate four separate items on a seven-point Likert-type scale (“1 = disagree extremely to 7 = agree extremely”). The statement used for these measures was “When you need a new product X (camera/ TV/ sweater), do you think that buying the brand X product X (Canon camera/LG TV/ Ralph Lauren Polo sweater) described above would be: 1) beneficial, 2) worthwhile, 3) wise, and 4) good?” 78 Table 6. Salient Attributes and Relevant Quality Dimensions for Each Product Product Dimensions Attributes Canon Performance It takes a sharp (clear) picture. _ Serviceability There are easily accessible authorized serv1ce centers. Camera Brand Prestige It has a prestigious/famous brand name. Technical Prestige It is manufactured in an advanced/hi-tech country. “CG Performance It has a high-clear picture. TV Performance It has a quality hi-fi stereo sound system. Serviceability There are easily accessible authorized service centers. Brand Prestige It has a prestigious/famous brand name. Technical Prestige It is manufactured in an advanced/hi-tech country. Ralph Design It has a fashionable design. There are a variety of colors. Lauren Design It is easy to manage/care for. P 010 Performance Sweater Performance It is comfortable. Performance It has a good fiber contenta. Brand Prestige It has a prestigious/famous brand name. Technical Prestige It is manufactured in an advanced/hi-tech country. / a This item was deleted due to significant cross-loadings with other constructs. Behavioral Intention (BI): Behavioral intention was measured by asking respondents to answer three questions: “I would consider buying the brand X product X described above. I would recommend the brand X product X described above to people who are close to me. Next time I intend to buy the brand X product X described above.” These items were also measured on a seven-point Likert-type scale (“1 = disagree extremely to 7 = agree extremely”). Familiarity with Brand: Familiarity with the brand was measured as product specific. The item was “How familiar are you with brand X product X (Canon camera] LG TV/Ralph Lauren Polo sweater)?” It was also measured on a seven-point Likert-type scale (“1 = extremely unfamiliar to 7 = extremely familiar”). 79 Familiarity with COM : Familiarity with COM was measured by the item, “How familiar are you with products made in country X?” on a seven-point Likert—type scale (“ 1 = extremely unfamiliar to 7 = extremely familiar”). Data Analysis For each product, COM was a dummy variable, coded as l=Japan, 2=China for the camera; 1=K0rea, 2=Malaysia for the TV; 1=U.S., 2=Mexico for the sweater. For eaCh quality dimension, the belief (Bi) that performance of the behavior will lead to a specific outcome, 1', was employed as an indicator of the latent constructs, that is, the five quality dimensions’ 6 (design, performance, serviceability, brand prestige, and technical prestige). Further, the overall quality perception was calculated based on the mean 0f zBiEi , where B; is the belief that performance of the behavior will lead to a specific i=1 outcome, i,, Er is the evaluation of each consequence, and n is the number of salient outcomes. The proposed structural relations were tested using EQS 5.1 (Bentler, 1993). In order to examine the moderating effects of brand and COM familiarities, the respondents were divided into four groups according to the level of brand and COM familiarities for a camera and a sweater. A TV was excluded in this analysis, since the brand used in this study was a well-known domestic one (LG TVs). A high brand familiarity group was identified as those respondents whose familiarity scores with Canon brand cameras or Ralph Lauren Polo brand sweaters were greater than median (5,0 on a seven point Likert-type scale for each product). A low brand familiarity group was specified as those whose scores were lower than median. The cases having median 16AS previously mentioned, the ease of use dimension was excluded in this study. 80 scores were excluded. L'tch"i Se, a group with high country familiarity Was classified, when their country familiarity Scores were greater than the median (4.0 on a seven point Likert-type scale for each product), and a group with low brand familiarity was identified, when its scores were lower than the median. The cases having median scores for country familiarity also were excluded. 8] CHAPTER V RESULTS AND DISCUSSION The purpose of this chapter is to report the results of statistical analysis and hypothesis testing, and to discuss these findings. The demographic characteristics of the sample are described and the results of reliability tests and confirmatory factor analysis are reported. This chapter concludes with a discussion of the current study’s findings. Demographic Characteristics of Sample - Respondents were asked about the following demographic Characteristics: age, gender, marital status, education, and income. As previously mentioned, these data were collected using the four parallel survey questionnaires (A1, A2, B1 and 32). ANOVA was performed to examine whether or not these four samples have differences in age and income. The results showed these groups were homogeneous in terms of age and income. In addition, chi-square tests were performed to see group differences for categorical variables such as gender, marital status, and education. The results indicated that none of these tests was significant. Thus, these four samples were combined to examine demographic characteristics and the profile of the sample is shown in Table 7. The ages in the sample ranged from 18 to 65, with an average age of 36. Females constituted 90 percent of the sample and males 10 percent. Most respondents (85 percent) were married, 131“ 15 percent were single. Most respondents (75 percent) received at least a university education. According to the Korea Statistical Yearbook (2000*,200/0 of the total Korean adult population (Whose age is above 20) held a college education. Thus, 82 the study SEInple is more highly educated than the general Korean population, me levels of family income were reasonably balanced. Table 7. Sample Profile (N=456) Age Gender Marital % % °/o 1 8—20 2 Male 10 Single 1 5 21 -30 28 Female 90 Married 85 3 1-40 45 41 -50 20 5 1-60 3 61-65 2 Education Monthly Family Incomea ($) % % High school degree or below 20 Less than $1 ,000 x 4""?— College degree 5 $1,000 but less than $2,000 25 University degree 55 $2,000 but less than $3,000 31 Master’s degree or higher 20 $3,000 but less than $4,000 20 $4,000 but less than $5,000 11 $5,000 or more 8 ‘ iverage currency exchange rate in 1997: $1 = 1,000 Won Reliability Tests and Confirmatory Factor Analysis As presented in Table 8, the results of reliability tests indicated that most constructs with multiple items achieved good reliability. Cronbach’s alpha coefficients for most constructs were greater than .7, except in two cases. The alpha coefficients of design and performance constructs for sweaters were .65 and .67, respectively. 83 Table 8. Results of Reliability Tests r Products (# of items, coefficient alpha) Camera TV Sweater Attitudesa (4, a: .93 ) Attitudesa (4, a: .95) Attitudesa (4, a: .94) Beneficial Beneficial ‘ Beneficial Worthwile Worthwile Worthwile Wise Wise Wise Good Good Good I______ntenti0nb (3, OF .77) Intestine" (3. a= .81) Intentionb (3, a: .87) Consider Consider Consider Recommend Recommend Recommend Intend Intend Intend Performance g2, OF .85) Design (2, or== .65) Hi gh-clear picture Fashionable design Stereo sound system Variety of colors W (2, a= .67) Manage/care Comfortable Fiber contentc a[Attitudes = Purchase Attitudes ;b1ntention = Purchase Intention °This item was deleted due to cross loading. Confirmatory factor analyses (CFA) were performed for the constructs with multiple indicators for each product using EQS 5.1. Covariance matrices for these analyses are presented in Appendix II. In the case of sweaters, results of Multivariate Lagrange Multiplier (LM) tests indicated that one of the attributes (fiber) for the performance dimension was cross-loaded with other constructs. Thus, this attribute was excluded in this analysis. The results showed a good fit of the model to the data for each product (for cameraS, x2=45.358, df=13, p <.001; CF1=0,98, RMSEA :5 0,07; for TVS, X2=34.513, df=24, p> 0.05, CFI = 0.99, RMSEA = 0.04; for sweaters, K2‘74'074’ (if-=38. p883. e m_ 200 n .nofioam .1. m m>._. u H “8250 n U .om_>cofio .9333 En 8e 8:8E:wfi 3.865 condemned 638d 02 e ecu—.8:— omaaoua. + ------—--—--——--—--————-----1 + emf—Em 6:95 83.53.. + own—3...:— . Fe . .3200 e o - h----—------------------‘--- $2695 .3 mamamzetfiom 335:».5. .3 Eugen—m .m oaswwm 92 F3F1 = -.150, p<.05; for sweaters, F4F1 = -.056, n.s.). (See Tables 11 a-c.) Thus, H2a was accepted for cameras and TVS, but not for sweaters. H2b proposed a positive effect of brand image on performance, and all three products supported this hypothesis (for cameras, F3F2= .328, p<.01; for TVs, F3F2= .542, p<.01; for sweaters, F4F2= .268, p<.01). Therefore, H2b was confirmed. H2c: Brand image will have a stronger effect on performance evaluation than will COM. The results of chi-square difference tests in Table 10 indicated that chi-square was improved Significantly at p < .01 when the equality constraint was released between F3F1 and F3F2 for cameras and TVS. In the case of sweaters, chi-square change improved significantly to p= .055 when the equality constraint between F4F1 and F4F 2 was removed. Thus, all three products had a stronger effect of brand image on performance. than of COM, as expected (for cameras, Ax2=64.47, df =1, p<.01; for TVs, Ax2=23.18, df =1, p<.01; for sweaters, Ax2=3.67, df =1, p=.055). Thus, H2c was supported for all three Products. gerviceabilitLDimension H3a: The serviceability evaluation of a product manufactured in a developed country will be higher than that in a developing country. H3b: Brand image and serviceability evaluation will be positively related. As mentioned earlier, serviceability was examined only for cameras and TVs. Results of t-tests indicated that the serviceability evaluation of cameras and TVS manufactured in develOped countries was not higher than that in developing countries 93 (for cameras, F4F1= -.028, n.s.; for TVS, F4F1= -.OO9, n.s.). (See Tables 11 a-b.) Thus, H3a was not supported. Regarding H3b, only TVs had a positive effect of brand image on serviceability (for cameras, F 4F2= .062, n.s.; for TVS, F4F2= .296, p<.01). Therefore, H3b was confirmed for TVS, but not for cameras. H3c: Brand image will have a stronger effect on serviceability evaluation than will COM. H3c was tested only for cameras and TVS, as previously discussed. The results of chi-square difference tests showed that chi-square change improved significantly when the equality constraint was released between F4F 1 and F4F2 only for TVs. (See Table 10.) Thus, brand image had a stronger effect on serviceability than COM did for TVS, but not for cameras (for cameras, Ax2=1.29, df =1, n.s.; for TVs, sz=4.42, df =1, p<.05). Accordingly, H30 was confirmed for TVS, but not for cameras. Biand Prestige Dimension H4a: The brand-prestige evaluation of a product manufactured in a developed country will be higher than that in a developing country. “4b: Brand image and brand-prestige evaluation will be positively related. Regarding brand prestige, H4a proposes that the brand-prestige evaluation of a product manufactured in a developed country will be higher than that in a developing country. The results for cameras and TVS supported this hypothesis, but not for sweaters (for cameras, F5F1= -.234, p<.01; for TVs, F5F1= -.119, p<.05; for sweaters, F5F1= - .052, n.s.). (See Tables 11 a-c.) Therefore, H4a was supported for cameras and TVS, but not for SWeaters. On the other hand, all three products indicated a positive effect of brand 94 image on brand prestige (for Cameras, F5F2= .364, p<.01; for TVs, F 5F2= .468, p<.01; for sweaters, F5F2= .402, p<. 01). Therefore, H4b was confirmed for all three products. H4c: Brand image will have a stronger effect on brand prestige than will COM. The results of chi-square difference tests Showed that chi-square change improved significantly when releasing the equality constraint between F5F1 and F5F2 for all three products. (See Table 10.) That is, brand image had a stronger effect on brand prestige than did COM (for cameras, Ax2=10.23, df =1, p<.01; for TVs, Afr—14.76, df =1, p<.01; for sweaters, Ax2=12.46, df =1 , p<.01). Thus, H4c was confirmed for all three products. Technical Prestige Dimension H5a: The technical prestige evaluation of a product manufactured in a developed country will be higher than that in a developing country. H5b: Brand image and technical prestige evaluation will be positively related. Regarding technical prestige, H5a proposes the technical prestige evaluation of a Product manufactured in a developed country will be higher than that in a developing country. The results for all three products supported this hypothesis (for cameras, F6F1= 4.483, p< .01; for TVs, F6F1= -.372, p<.01; for sweaters, F6F1= -.485, p<.01). (See Tables 11 a-c.) Therefore, H5a was supported. In addition, all three products indicated a positive effect of brand image on technical prestige (for cameras, F6F2= .139, p<.01; for TVS, F 6F 2: .244, p<.01; for sweaters, F6F2= .245, p<.01). Therefore, H5b was confirmed. 95 HSc: COM will have a stronger effect on technical prestige than will brand image. The results of chi-square difference tests showed that chi-square change improved significantly when the equality constraint was released between F 6F] and F 6F2 for all three products. (See Table 10.) That is, COM had a stronger effect on technical prestige than brand image did (for cameras, Ax2=12.63, df =1, p<.01; for TVs, Ax2=20.29, df =1, p<.01; for sweaters, Ax2=55.95, df =1 , p<.0 1). Thus, H5c was confirmed. Overall Quality H6a: Design evaluation and overall quality will be positively related. H6b: Performance evaluation and overall quality will be positively related. H6c: Serviceability evaluation and overall quality will be positively related. H6d: Brand-prestige evaluation and overall quality will be positively related. H6c: Technical prestige evaluation and overall quality will be positively related. Note that the design dimension is not examined for cameras and TVs and the serviceability is not proposed for sweaters. As represented in Tables 11 a-c, results of t- tests indicate that overall quality (F7) was positively related to performance (F3 for cameras and TVs, and F4 for sweaters), brand prestige (F5), and technical prestige (F6) for all three products. Overall quality also was positively related to design (F3) for sweaters, and serviceability (F4) for cameras and TVs. (For cameras, F7F3=.320, p<.01; F7F4=.323, p<.01; F7F5=.285, p<.01; F7F6=.278, p<.01; For TVS, F7F3=.598, p<.01; F7F4=.299, p<.01; F7F5=.226, p<.05; F7F6=.181, p<.05; For sweaters, F7F3=.550, P<-01; F 7F4=.478, p<.01; F7F5=.219, p<.05; F7F6=.229, p<.05) Therefore, H63 to H66 were all confirmed. 96 H7a: COM will have no relationship to overall quality. H7b: Brand image and overall quality perception will be positively related. H7a proposes no effect of COM (F 1) on the perception of overall quality (F7) and all three products indicated that COM was not related to overall quality (for cameras, F7F1= .005, n.s.; for TVs, F7F1= .050, n.s.; for sweaters, F7F1= .062, n.s.). (See Tables 11 a-c.) In addition, brand image (F2) had no influence on the perception of overall quality (for cameras, F7F2= .059, n.s.; for TVS, F 7F2= -.077, n.s.; for sweaters, F7F2= .012, n.s.). Therefore, H7a was supported, but H7b was not supported. Purchase Attitudes H8a: Overall quality perception and purchase attitudes will be positively related. As Fishbein and Ajzen (1975) proposed, the consumers’ perception of overall quality (F7) had a positive effect on consumers’ attitudes toward buying the product (F8) for all three products (for cameras, F 8F7= .316, p<.01; for TVS, F8F7= .416, p<.01; for SWeaters, F8F7= .271, p<.01). (See Tables 11 a-c.) Therefore, H8a was confirmed. “8b: COM has no relationship to purchase attitudes. H8c: Brand image has a positive relationship to purchase attitudes. H8b pr0poses no effect of COM (F 1) on consumers’ attitudes toward purchasing a product (F8), but only sweaters supported this hypothesis (F8F1= -.047, n.s.). (See Tables 11 a-c.) Surprisingly, cameras and TVs indicated a negative effect of COM on consumers’ attitudes toward purchasing these products (for cameras, F8F 1= -.184, p<.01; for TVs, F8F 1= -.225, p<.01). That is, Korean consumers had more positive attitudes toward purchasing these products manufactured in developed countries than those in 97 developing countries. Thus, H8b was rejected for cameras and TVs, but not for sweaters. On the other hand, a positive effect of brand image on purchase attitudes was found for all three products (for cameras, F8F2= .183, p<.01; for TVS, F8F2= .316, p<.01; for sweaters, F8F2= .252, p<.01). Thus, H8c was confirmed. Intention to Buy H9a: Attitudes toward purchasing a product are positively related to the intention to buy the product. As F ishbein and Ajzen (1975) proposed, consumers’ attitudes toward purchasing a product (F 8) had a positive effect on their intention to buy the product (F9) for all three products (for cameras, F9F8= .611, p<.01; for TVs, F9F8= .686, p<.01; for sweaters, F9F8= .709, p<.01). (See Tables 11 a—c.) Therefore, 11% was supported. H9b: COM will have no relationship to intention to buy. H9c: Brand image will have no relationship to intention to buy. Regarding H9b, no effect of COM (F1) on the intention to buy the product (F 9) was found for the three products as expected (for cameras, F9F1= -.034, n.s.; for TVs, F9F1= -.050, n.s.; for sweaters, F9F 1: -.013, n.s.). (See Tables 11 a-c.) Therefore, H9b was not rejected. On the other hand, although no effect of brand image (F2) on the intention to buy (F9) was found for cameras and TVs, a positive effect of brand image on intention to buy was found for sweaters (for cameras, F9F2= .049, n.s.; for TVs, F9F2= 906, n.s.; for sweaters, F9F2= .173, p< .01). Therefore, H9c was rejected for sweaters, but not for cameras and TVs. 98 Testing the Moderating Effects of Brand and C(ZM Famih‘arifies on the Relationships among COM, Brand Image, and Quality Perception of Hybrid Products fissification of Four Groups based on Brand and COM Familiarities In order to examine the moderating effects of brand and COM familiarities on the relationships among COM, brand image, and quality perceptions of hybrid products, the respondents were divided into four groups according to the level of brand and COM familiarities for a camera and a sweater. A TV was excluded in this analysis, since the brand used in this study was a well-known domestic (Korean) one (LG TV)- The high brand familiarity group was identified as those respondents whose familiarity scores With Canon brand cameras or Ralph Lauren Polo brand sweaters were greater than median (5.0 on a seven point Likert-type scale for each product). The low brand familiarity group was Specified as those whose scores were lower than the median. The cases having median Scores were excluded. Likewise, a group with high country familiarity was classified, when their country familiarity scores were greater than the median (40 on a seven point Likert-type scale for each product), and a group with low brand familiarity was identified, when its scores were lower than the median. The cases having median scores for country familiarity also were excluded. Based on these classifications, four groups were identified including high brand familiarity/high country familiarity, high brand familiarity/low country familiarity, low brand familiarity/high country familiarity, and low brand familiarity/low country familiarity. As indicated in Table 12, due to the need to delete cases with median scores, 203 cases Were excluded for the catnera and 73 cases were eliminated for the sweater. This table also indicated that each segment did not have a large enough sample Size to run 99 SEM analysis. For the camera, almost half of the total respondents were classified into the high brand/high country familiarity group. Therefore, the data in this study may not be Useful for examining the moderating effect of brand and country familiarities in the relationships among COM, brand image and quality perceptions of hybrid products. No further analyses were performed. Table 1 2. Frequencies of Four Groups Based on Brand and Country Familiarities Camera Sweater Group Frequency Percent Frequency Percent HB lPIC 123 48.6 43 27.9 HlB/LC 46 18.2 30 19.5 LB/HC 37 14.6 26 16.9 #3ch 47 18.6 55 35.7 Total 253 100 1 54 100 Missing 203 - 73 - HB/HC = high brand/high country familiarity; HB/LC = high brand/low country familiarity; LB’HC: = low brand/high country familiarity; LB/LC = low brand/low country familiarity. 100 DISCUSSION The proposed quality dimensions were found to be useful in explaining the relative importance of COM and brand image on Korean consumers’ quality perception, and these five dimensions were not independent of each other. That is, the performance and serviceability dimensions in the experience mechanism were correlated and the brand Prestige and technical prestige dimensions in the image mechanism were correlated. Overall, the differential effects of COM and brand image on the five quality dimensions were well explained based on the logic of the evaluation mechanisms. (See Figure 3.) In the case of the search dimension (i.e., design for sweaters), COM had no impact, but brand image had a positive influence, as expected. Regarding the image evaluation mechanism, the brand and technical prestige evaluations of cameras, TVs, and SWGEIters manufactured in developed countries were higher than those in developing Coul'ltl‘ies as expected, except in the case of sweaters, where COM had no effect on brand preSti ge. These results may reflect the fact that manufacturing locations are more imPCDI’Iaant in a consumer’s perception of electronics’ brand prestige than in those of fashi on goods. Brand image also positively influenced these evaluations for the three m“finctts, as expected. For all three products, brand image had a stronger effect on brand Q‘K‘fgfi than COM did, while COM had a stronger effect on technical prestige than brand \ age did, as proposed. fl . . . Experience dimensions (performance and serwceability), on the other hand, had so"jewhat mixed results. (See Figure 3.) The performance evaluations for cameras and TVS manufactured in developed countries were higher than those in developing countries, but r: or in the case of sweaters. Brand image influenced performance significantly and 101 exerted a greater influence on performance than COM did for all three products as eXpected. Therefore, these results indicate that consumers rely more on brand image than on COM when evaluating products’ performance, and they are more sensitive to manufacturing locations for cameras and TVs than for sweaters. On the other hand, COM had no effect on serviceability for either cameras or TVs, and even brand image had no influence on serviceability for cameras. This may reflect the fact that because cameras are portable and consumers can bring them anywhere to be fixed, Korean consum(amined the effects of COM and brand. Due to the power of SEM, the present Stud" Can examine the effects of the quality dimensions (design, performance, N&\Qeability, brand prestige, and technical prestige) on overall quality, as well as those Qf COM and brand image. Thus, this study found that COM and brand image have direct effigcts on the quality dimensions and they have effects on overall quality perception ind irectly through some quality dimensions. Indirect effects of COM (F1) and brand im3g¢ (F2) on overall quality (F 7) were additionally tested and the results indicated that 102 COM and brand image had significant indirect effects on overall quality for all three Products (for cameras, F 7F1 = -.279, p<.01; F 7F 2 = .268, p<.01; for TVs, F 7F 1: -.186, P<-01, F7F2 = .551, p < .01; for sweaters, F7F1=-.196, p<.01, F7F2 = .376, p < .01). Another intriguing result of this study is the finding of a significant effect of COM on purchase attitudes toward electronics (camera and TV), even in the presence of multiple cues. (See Figure 3.) That is, Korean consumers had more positive attitudes toward purchasing electronic products manufactured in developed countries than those in developing countries. Previous studies (Erickson et al., 1984; Johansson et al., 1985) found no effect of C00 on purchase attitudes, when multiple cues were provided. Such a different result could be due to different samples. That is, due to their limited experience With Other COMs, Korean consumers could be more sensitive to COM than American or JaPa-Ilese consumers when they form attitudes about purchasing electronics, as in the Studies of Erickson et al. (1984) and Johansson et a1. (1985). Unfortunately, no studies e""‘fi‘d-Igllined the effect of COM on purchase attitudes of Korean consumers. However, Ulgado and Lee (1998) reported that Korean consumers consider COM to be more imPC>rtant in evaluating product quality than the American sample does. Thus, further study is required to verify this finding. l’. " if p This study also found the “product effect” in the relationships among COM, brand .\\$B%e" purchase attitudes, and intention to buy.\.’ (See Figure 3.) That is, COM had no foect on the purchase attitudes for sweaters, while it had a significant impact for cameras d TV 5, as previously mentioned. COM also had a greater influence on the brand prestige of cameras and TVs than of sweaters. That is, the brand prestige evaluations of came Tas and TVs produced in developed countries were higher than those in developing 103 countries, but there Was no difference between them in the case of sweaters. In addition, for sweaters, the brand image had a significant positive effect on intention to buy, while for cameras and TVs, it had no impact. Thus, brand image is more important for SWeaters than for TVs and cameras in purchase attitude formation and purchase intention, While COM exerts more influence on the purchase attitude formation for TVs and CMeras than for sweaters. Using meta-analysis of effect sizes, Liefeld (1993) reported that the magnitude of C00 effects is related to product type. That is, COO effects were larger for technically complex and expensive products than for inexpensive products low in teChnical complexity. Cameras and TVs are considered more technically complex and exPctlsive than sweaters. Thus, the current study’s findings are consistent with those of Liefeld (1993). In sum, this study reveals the interesting fact that regarding electronics (TVs and Canleras), Korean consumers are somewhat sensitive to COM for their performance eval nations, image evaluations (brand and technical prestige), and affective evaluations (Purchase attitudes). Brand image is more important than COM in most consumer decision-making. In the case of fashion goods (sweaters), Korean consumers show less sensitivity to COM than with electronic productsnf’B’fand imagef exerts great influence on I 2. o o o it o e } CV97 decrsion-makmg process, and even on purchase intention. : 104 CHAPTER VI SUMMARY AND CONCLUSIONS The purpose of this chapter is to summarize the research study that was discussed in detail in Chapters I through V. The objectives of the study, past literature, conceptual model , research methods, and empirical results are summarized in the next section. Managerial implications of our findings are addressed in the following section. It condudes with a discussion of the study’s limitations and directions for future research. Summary of Research Objectives and Past Literature Marketing researchers have identified studies of products’ country of origin (COO) as “the most researched” domain of international buyer behavior (Tan & Farley 1987; HeSIOp, Papadopoulos, & Bourk 1998). In spite of its quantitative achievement, exi Sting COO research still has several unresolved issues. This paper focuses on three pr Dblfitms. First, COO research has been criticized as being atheoretical. This is partially due to a lack of well-defined quality dimensions. A determination of quality dimensions that could be used across product classes is necessary in order to construct a theory, but PrWiOuS studies lacked a set of widely accepted general quality dimensions because these \g‘fi‘?’ “0t well defined. Second, COO researchers have failed to provide a comprehensive Vi 3w of the effect of C00 on consumer decision making. That is, most studies have 65mm“ the effect of C00 on product evaluations, which is the cognitive part of consumer decision making. The effect of C00 on affective or behavioral components such as attitudes or behavioral intention is rarely studied. Third, country-of-origin 105 deSlgnations have pIOliferated due to firms’ expanded global sourcing. Research on the hybrid product phenomenon, however, is lacking. The relative importance of brand Image and the various components of C00 in consumers’ decision making is still in CI UeSti on. Finally, previous literature suggested that product familiarity is an important influence on the relationship between C00 and product evaluation. The definitions of product familiarity in these studies, however, are too vague to differentiate product familiarity from brand familiarity. Especially in the case of hybrid products, COM is not the SaJne as country of brand (COB), where the company’s headquarters are located. ThUS, brand familiarity does not necessarily mean familiarity with the country where the PTOdl—lct is manufactured. Therefore, the concept of product familiarity should be redefined and reexamined in the context of hybrid products. The objectives of this study are to examine Korean consumers’ behavior regarding 1) the quality dimensions of product evaluation; 2) the effects of country of manufacture (CO M) and brand image on product evaluation, purchase attitudes, and purchase intention for hybrid products; and 3) the effects of brand and COM familiarity on the re“\m‘iorlships among COM, brand image, and product evaluation of hybrid products. Summary of Conceptual Model The current study identifies six quality dimensions including ease of use, design, efformance, serviceability, brand prestige, and technical prestige. These quality dimensions are dichotomized on the basis of their symbolic and fimctional aspects. The functional aspects of quality are evaluated through search and experience mechanisms. In the search mechanism, consumers assess quality dimensions that can be evaluated 106 through information Seeking and processing, without purchasing or using the item. The quality dimensions assessed through this mechanism are design and ease of use. The experience mechanism is a consumers’ quality evaluation process that is activated only after purchase and use. Thus, consumers make inferences about this type of quality based on intrinsic and extrinsic cues at the time they purchase a product. The quality diluensions evaluated through the experience mechanism are product performance and serviceability. Finally, the symbolic aspects of quality are evaluated through the image meChaIIism, in which consumers attach images to the product. The quality dimensions evaluated through the image mechanism are brand prestige and technical prestige. Based on the Theory of Reasoned Action (F ishbein & Ajzen 1975), a comprehensive model is proposed to explain the relative importance of COM and brand image on consumers’ purchasing behaviors for hybrid products. The cognitive co’1'113-onent in Fishbein & Ajzen’s model (1975) is dichotomized into two components including quality dimensions and overall quality. Regarding the search mechanism, COM is hypothesized to have no impact on the perception of the design dimension, because perceptions of the design or ease of use are formed at the point of purchase, and ““3 therefore, rely less on extrinsic cues. Brand image, however, is expected to have So’le influence on consumers’ quality perceptions because a company intensively \L‘QQPS its brand image through various marketing tools. Since the evaluations of the experience dimensions (performance and 5&fliceability) are formed after the product has been purchased and used, COM and brand image are hypothesized to have an influence on the qualities of these dimensions. The 107 Clay-en . t StUdy further proposes that brand image has a stronger effect 011 the perception of the experience dimensions than does COM. Regarding the evaluations of image dimension (brand prestige and technical prestige), both COM and brand image are hypothesized to affect the qualities of these dimensions, The current study further proposes that the effect of brand image on the brand-prestige dimension is greater than that of the effect of COM on brand prestige, while the effect of COM on the teChnical-prestige dimension is greater than that of the effect of brand image- The relationships between each quality dimension and overall quality are P roposed to be positive since consumers use multiple quality attributes that are integrated into an overall quality perception. PTeViOUS Smdies found no effect of COM when other . . . . l: ° . . intnnSIC lnforrna ion was prov1ded, while the effect of brand remained strong. Thus, in the provision 0i intrinsic cues, no effect 0f COM, but a Significant effect of brand image on Overall quality are proposed. The COM of hybrid products 15 hypothesized to have no impact on consumers’ aSe attitudes, While brand image is expected to have an influence on it since brand im ge is a total perception of the brand that is obtained by processing information from arious Sources over time. Regarding consumers’ intention to buy hybrid products both v 9 and brand image are hypothesized to have no impact, CO j The current study also proposes models of the differential effects of COM and brand image on the quality perception of hybrid products using brand and COM farni/iarities as moderating variables. Conventional product familiarity is specified as “brand familiarity” and “COM familiarity.” Brand familiarity is defined as the 108 consumer’s prior knowledge regarding the brand in a product Class, 0,, the other hand, COM familiarity is defined as the individual’s prior knowledge regarding products made in a Particular COUNTY- Based on brand familiarity and COM familiarity, this study proposes four segments of consumers: high brand familiarity/high COM familiarity, high brand familiarity/low COM familiarity, 10W brand familiarity/high COM familiarity, low brand familiarity/low COM familiarity. The differential effects of brand image and COM on quality perceptior1 are proposed for each 0f the four segments. The logic is that the more a consumer knows about the country’s product, the more likely he/she will be to use the C00 cue to evaluate the product. Likewise, it is expected that the more a consumer knows about the brand, the more likely he/ she Will be to use the brand name to evaluate the product. Summary 01' Research Methods There Were two phases to the study: the qualitative phase and the quantitative In me qualitative phase. focus group interviews were conducted with Korean P1l ase. s dents in the [1,3 and others in Korea to explore the conceptual meaning and cultural text of the target concepts, salient product features, brand names and COMS Cori ’ ' {ma ti on gathered from the qualitative phase was incorporated into the development inf?) oi the quantitative survey instrument. The resulting quantitative study was cross- actional' The design used a self-report survey instrument to collect consumer behavior d a ta from native Koreans. Based upon the outcome of the focus group interviews, the folloWing target products and correSponding brand names were selected: Canon cameras, LG TVS, and 109 P010 Sweaters. Once developed in English, the instrument was translated into Korean by Koreans not affiliated with the prOject using a double-blind translation procedure to achieve construct equivalence. All of the measures used in the study were previously establiShed and seven-point Likert-type scales were used. Korean consumers living in Seoul, Korea and several satellite cities constituted the sarnple in the StudY- To assure the inclusion Of a broad Spectrum of geographic and socioeconomic groups, potential survey areas were thoroughly reviewed and selected. Seven research assistants were selected from a major Korean university to collect data. The assistants collected data over a five-week period during June and July 1997, prior to the economic crisis- Of the 550 questionnaires that were distributed, 456 were returned, yielding an 82 percent response rate. For eaeh Product COM was °°ded as a dummy variable to differentiate the developing and developed countries. For each quality dimension, the belief (Bi) that the mget pmduet W1“ have a specific attribute (i) was employed as an indicator of the five quality dimensions (design, performance, serviceability, brand prestige, and technical Pr‘éstige} Further, the Overall quality perception is CfllCUIated based on the mean of the fomula indicated in a Fishbein and Ajzen (1975) study, The proposed structural . s were tested using EQS 5.1 (Bentler, 1993). -, re\ 9‘10” Summary of Empirical Results A confirmatory factor analysis was performed for the constructs with multiple £11611: Gators for each product. Results showed a good fit of the model to the data for each product and achievements of convergent validity and discriminant validity. Since the q (.131 ity evaluation mechanisms cannot be included as second-order factors, the structural 110 en- . 01‘s were correlated 1n SEM programs. That is, for TVs and earneras, stnzctura! errors 0 g o o . f Performance and servrceability, and brand prestige and technical prestige were correlated, and for sweaterS, those of brand prestige and technical prestige were correlated, The results of each hYPOtheSiS test are summarized in Table 13 based on the jesults of t-testS and chi-square difference tests. Some hypotheses of the current study include a comparison of the relationShip between COM and brand image. Thus, chi- square difference tests were performed by imposing four equality constraints for TVs, cameras and sweaters. For all three products, releasing most equality constraints improved the model fit significantly, which indicates that the effects of COM and brand image are different in each proposed relationship, except in one case. For cameras, both COM and brand image had no influence on serviceability, The proposed model has an acceptable fit to the data for all three products. For “he. searei‘ dimension (design), no effect 0f COM, but a positive effect of brand image for sweaters, as expected. For the experience dimensions (performance and WQTC found se: hiceabflm’)’ COM had a significant effect on the Performance dimension for cameras TVS, but not for sweaters. It had no effect on serv1ceability for either cameras or 1" 25. Brand image had a stronger effect on these dlmensmns than did COM, except in case of cameras where no effects of COM and brand image on serviceability were ma {oyd . Regarding the image dimensions (brand prestige and technical prestige), COM 3135 brand image had significant effects on these dimensions as expected, except in the case of the effect 0f COM on brand prestige for sweaters. lll Table 13. Results of Hypotheses Tests Hypotheses Standard“ Run“; of Coefficients Hypothesis Tests‘ Hla: No effect of COM on design S: -.l33 Supported J H1 b: Positive effect of B14033 Adesign S: .308" Supported J H2a: Negative effect of COM on performance C: -.213** Supported T: -.150* Supported I x S: -.056 Not Supported H2b; Positive effect Of Bl on performance C: .328'” Supported T: 542'” Supported S: .268'Ml Supported H20: Stronggr’éffect of B1 on performance than C: sz=64.47** Supported COM T2AX2= 2318’.“ Supported S: A12: 3,67b Supported H3a: Negatimect of COM on serviceability C: -.028 Not Supported . . . T1 --009 Not Supported H3b: Positive effect of Bl on serv1ceability C: .062 Not Supported . . . T: 296‘” Supported H3c: Stronger effect of B1 on servrceabillty than C: sz=1.29 Not Supported COM . T: Ax2=4,42* Supported H4 a: Negative effect of COM on brand prestlge C: -.234u Supported T: -.119* Supported . S: '-052 Not Supported [£443. Positive effect of B1 on brand prestlge C- .364“ Supported - T .468" Supported . S .402" Supported H 4 . Stronger effect of BI on brand prestlge than C: sz=10.23** Supported c. (30M T: Ax2=l4.76** Supported a1 S: Ax2= 12.46* Supported ative effect of COM on technic prestige C: -ET** Su cried H3 at Neg T: -.372** Suggorted I . . 33 ‘-435** Supported b‘ Positive effect of B1 on techmcal prestige C: .139“ Supported 3; ° T! .244" Supported . . S: .245" Supported If: stronger effect of COM on technical prestige C: Axf=12,63** Supported “'5 than BI T: Ax2=20.29** SUpported S: Ax2=55.95** Stipported : Twymil tests for H l a, H7a, H8b, H9b, anfi9c and one-tail tests for the rest hypotheses ”Significant at p = 055; "' p < -05; "p < .01 BI QBfand Image : Country-of-Manufacture (Dummy Variable coded as l=Japan, 2=China for cameras; 1=Korea, M - CO =Malay5ia for TVs; 1=U.S., 2=Mex1co for sweaters) C , Czamera; T = TV; S = Sweater 112 Table 13. Cont’d / Hypotheses ‘ snowman of - Coefficients Hypothesis Tests H6a: Positive effect of desrgi on overall quality S: .550“ I Supported J H6b: Positive effect of performance on overall C: .320“ Supported quality T: .593" Supported A. .K S: .478" Supported H6c:: Positive effect of servweability on overall C: .323" Supported 1 quality A Ti 299'” Supported H6d: Positive effect of brand prestige on overall C: .285" Supported quality T3 226" Supported j . , S: '219" Supported H6c: Positive effect of technical prestige on C: .278" Supported overall quality T: ~181“ Supported I . S: .229‘" Supported H7a: No effect of COM on overall quality C: .005 Supported T2 .050 Supported . . Ls: .062 Supported H7b! Positive effect of BI on overall quality C: .059 Not Supported T -.077 Not Supported . S .012 Not Supported H8a: Positive Effect of overall quality on purchase C .316" Supported attitUdeS T 416* * Supported . S .271** Supported H ‘8b' No effect of COM on purchase attitudes (3; -.18—4** Not Supported - T: -.225** Not Supported k , S: 'Oi Supported H8 . Positive effect of BI on purchase attitudes C: .183" Supported c“ T: .316‘* Supported ff f h tt'tud (S; 26512; Supported . ositiVe 8 ea 0 PW0 386 a 1 es on i ' 1'" Supported H9 8‘ Emotion to buy T3 585‘” Supported , t b (S; ”E; Supported ‘ tion 0 uy : -.O34 . No effect of COM on inten SUpported “gt' T: '.050 Supported . S: -.013 SUpported , No effect of BI on intention to buy C: .049 Slipported ug& T2 .006 Supported S: .173" Not Supported BI ‘ CON? C 5 Camera; T: , Twat?!“ tests for H 1 a, H7a, H8b, H9b, and H9c and one-tail tests for the rest hypotheses ’P <1.05; "P<-01 grand Image ___ Country-of-Manufacture (Dummy Variable coded as 1=Japan, 2=China for cameras; 1=Korea, =Malaysia for TVs; 1=U.S., 2=Mexico for sweaters) W; S = Sweater 113 N 0 effect of COM on overall quality was found for 2111 three Pro ducts. Brand image also had no effect on overall quality. For the purchase attitudes, COM was Proposed to have no effect, but only the sweater supported this hypothesis. A significant effect of COM on purchase attitudes was found for cameras and TVs. On the other hand, brand image had a significant effect on purchase attitudes for all three products. No effect of COM on intention to buy was found for the three products, as expected. Regarding cameras and TVS, no effect of brand image on intention to buy was found, but for sweaters, a positive effect of brand image on intention to buy was found. The current study failed to test the models of the differential effects of COM and brand image on quah'ty perceptions of hybrid products using brand and COM familiarities as moderating Variables. Because the brands employed in this study were well-known to Korean consumers, we were unable to identify a group of low brand familiarity. Managerial Implications The findings of the current study provide strategic implications for multinational and apparel companies that are targeting Korean consumers. International 61 Q etronics marketing managers in these companies ShOUId fiI'St aCknowledge that establishing a 11:. orig afld favorable brand image is important in Korean consumers’ Product evaluations S tows of performance, brand prestige, technical prestige, and purchase attitude in atiorr. The findings of this study, however, indicate that moving production facilities {0% z" o o l o 0 developing countries can damage prestlglous brand and technical images and that this 11133 negatively affect Korean consumers’ attitudes toward purchasing the product. ‘1 Thus, _/ electronics marketing managers who decide to manufacture their products in deve10ping countries should emphasize the excellence of product performance in their 114 advertisements. Prestigious brand and technical images shoujd be highlighted to compensate for the negative images associated with certain COMs. 1,; International apparel managers, on the other hand, should acknowledge that the establishment of a strong and favorable brand image is critical in Korean consumers’ deCiSion-making processes, while the effect of COM is limited. This study found that brand image strongly influences Consumers’ product evaluations of design, performance, brand and technical prestige, purchase attitudes, and behavioral intention. On the other hand, these consumers did not consider the COM of the hybrid sweaters when they evaluated design, performance, and brand Prestige of the products. Accordingly, country-sourcing considerations become less significant for apparel with strong and favorable brands. Limitations and Recommendations Manageria‘ implications 0f the present StUdYa however, should be considered in light 6 f I656211611 limitations. This study employed only three products from two product Categories: electronics and apparel goodS- Although they were selected because they w§ e bi,nati0nal and well known to the subject POPUlation, future research should 1' amine the generalizability of the current findings to other products with different C): ii c“gristics. Especially, this study revealed that the relative importance of COM and ch fl a (1 br/ ensitive to COM and more to brand image in their purchase decisions for apparel image is different based on the product categories. Korean consumers are less pro ducts than for electronics. Electronics are considered to be more technically complex aria expensive products than apparels. Thus, the moderating effects of price and level of technology on the consumers’ views of COM versus brand should be examined, 115 This StUdY employed the mean of product beliefs and 6Valuation Composites for the construct of overall quality, as prOposed by Fishbein and Ajzen (1975), and used product beliefs for the constructs of quality dimensions. Accordingly, the relationships between quality dimensions and overall quality could be too strong to detect the influence of COM and brand image on consumer evaluations of overall quality, although indirect effects of COM and brand image on overall quality were significant. Therefore, future study should incltflle a measure 0f overall quality to Verify the findings of the current study. The current Study also proposed brand and COM familiarities as moderating Variables to explain the differential effect of COM and brand image on the quality perceptions of hybrid products- However, since the brands employed in this study were 1 known to Korean consumers, we failed to test the model because we were unable to Wel jqentify a low brand familiarity group. Therefore, further study should examine the “h deratlng effects of brand and COM familiarities among different products. o 116 APPENDICES ll7 APPENDIX I QUESTIONNAIRE 118 Greetings! The Life Science Institution at Sung K , - is a research institution Cfmtl‘f'butlng to community development 3:“ Kwa“ Umverzlyo consumer problems: At this tune, this institution is conducting a ough ‘he Feseam of Korean and American consumers’ PUIChasing behaviors. Comparative 5m Because this surve ' ' nduc liter, . . y 15 being (:0 ted anon ' zed by comp. your individual responses will be confidentigLInOUSlY and W1“ be analy nous, . ur CfCCp and there are no right or wro 11g answers. Y3K: igggziiysnilfing: aged 2“” for our acadernicopurpose. This survey Will take about 15 minutes to complete “5631mm partiCIPatlon in this study is voluntary, you can stop answering the questions at y iii/011 want. Your cooperation Wth this survey i s greatly appreciated. 1991.6. Life Science Institution at an Stat Sung Kym Kwa“ Univers° I\Ilichig e Unive ‘ meessor sun fin Hwang lty I:‘mfessor Pysarclu'k Plisl) 119 MICHIGAN STATE UNIVERSITY 1 777' - . indica" your vo untary agr’:e’q,:le:!lonnalre will take approximately 15 ”my,“ to comp/“oe- "Zine. You may discontinue em to participate by completing and returning (his question No different products 1) a Pf’lci'aafl'on at any time as . camera and 2) ‘ ”stirrup: yourself in the follow“.8 mum“, you would like to bgzfiures and prices for each of - 0“ ‘v Isl! different stores to comparethc varimls sent View by Chukka the number (hag pt'OdUClS. Please a tuner each question With th e response that best descn you Section I. In this section we are 2516:: h res when E °“’ mu re the ioliowing feat“ buy a product, regardless of the speci fic brand. ch do 3'0“ like or how 30"de a 1. When purchasing any camera. how 300d or bad is it that the (miners has eat}! of - following features- the Neither m amend! o Ellrcmc.y 0““: sughdy 33d m Sll‘hlly 0:” Good Bad Bad Bad Good 6004 G > it has automatic focus l 3 3 4 5 6 7 . 7 > it takes a sharp (clear) pieture l 2 3 4 S 6 o I t 7 > It IS easy to load I 2 3 5 6 O 4 1 > it has a prestigious brand name 1 ., 3 5 6 7 its price is expensive 1 2 3 5 6 1 4 > it has awarranty l 2 3 5 6 1 . - 4 6 > they: are €38“)! accessible authorized service l 2 3' centch 4 5 > as produot quality is good 1 2 3 S 6 7 4 > it is manufactured in an advanced/hi-tech l 2 3 6 country 4 i 5 7 2. When purchasing any TV, how good or bad is it that the TV has each of the following rm. Bad Bad ’ Bad 00’ 3‘80”, - M > it has a high-clear picture l 2 3 G W 0004 51mm"), 4 5 > there is a hi-fi stereo sound system 1 2 3 5 7 4 > it has remote control 1 2 3 5 6 7 4 > it has a prestigious brand name l 2 3 5 6 4 7 > its price is expensive 1 2 3 4 5 6 7 3' it has a warranty 1 2 3 4 5 6 7 5 6 7 120 ’ Lfi‘ > there are gas” 7 v - 6 centers . 3CCe551blc authorized service 1 2 3 4 5 r its prOducr qUality is good 5 6 7 l 2 3 4 r II is manufactwed I 5 6 7 country n 3" adVanced/hi-tech 1 3. ' a . 0 Ho“’ ’mponnnt ‘8 If that you do what ‘he follo‘vino \V'ifl‘ respect ‘0 bIIYlBg a camera? . PeOplc think you should do Extremely Quific Ncnhgf Sommhfl 0““: Extremely Unimportant Unimportant 32?;32‘: n‘ 32:32:? ‘mwmnt lmpomM Important f I 5" Friends I 2 7 3 4 5 6 > P am in/rela ti ves 1 2 7 3 4 5 6 > ’ ' 7 I\ e r ghbors 1 2 3 4 5 6 Vex a 7 r S a lesperson 1 2 J 4 S 6 o o . buy'ing a Hoxv important 15 it that you do ‘Vhat the following p eople think you 3] id d with respea ‘0 mu 0 4' TV? A imam“; . . . «a ‘T‘T‘m" 0““ 50m evhat high" 0““: mm“ “W0 L-nanOru-nt unimportant Uni.“ pom." ”Martin”. no.- somcfihi“ “n9" 1 Pom". ‘mporun 7' Friends 1 2 3 4 6 u 5 1 *; Family/rcmwes 1 2 3 4 6 5 > Neighbors 1 2 3 4 6 7 5 > satesperson 1 2 3 4 5 6 7 5. HO\\' familiar are You “1th: Surf“?! Sure Somewhat :Qi ‘l‘ler n amt tar ' - - . n lmrltar Unfamiliar [1‘ ”I iliill' ”or Someh'bil 0”,! "e r gr r ”“1 ‘ I 9 > Canon brand products 1 2 3 “m m '3’ ”0,7,.” image) ‘1 5 "W”: > Canon brand cameras 1 2 3 5 7 <1 5 -> LG (Gold Star) brand products 1 2 3 4 6 7 5 > LG (Gold Star) brand TV sets I 2 3 4 6 7 5 6 7 121 6' \Vhat is y our general 5 mcwm‘ 0"": tummy ""Pression ofth ,. ‘. .m r m“: " Good Good prOdUCts: C {0’10“ng EILr‘mdy (33:2: :zdmcwmt gt: 520d Good / Canon brand Cameras 1 a a 4 1- a r L (001d Star) brand TV sets 1 2 3 4 acts idemified in sections “-111. . . 5 it with ihe Section II. In a department store y 0“ find 8! “mm-n N to the camera is a 9%“ that descqzequesfions 7' [ea turcs listed below. Please read thf description carer‘s? because we are going to 35“ you ‘ 11 to evaluate the camera described so the box below. y Sapposc that during your visits to the Stores, you Come the “v0 prod across ‘— {i ‘ C‘Mn One-touch Perso n 31 Compact Camera ‘ _One.year wan'an ‘7 ‘Manu (‘21de in cm“ __Automatic focus ’Zoom lens ‘ ' a _Ezuy to load _Easily accessible amhofized service 6"“ _Date on picture ____ _Takes sharp (clear) pictures . . \s How likely is It that the Canon brand 6‘“? 7 . camera described above “0““! have V . \ Q‘f‘f‘m 9‘“ ' fix e fouowing characteristics: 5"""M'F Quite Somewhat 1.52".“ "‘“v‘h’ u‘» 1 Unlikely Unlikely L'nlikel "'Y "of ~ W 31 tnlikely Like 6 ‘; 'u takes a Sharp (clear) pictUre l 2 3 4 5 7 > it is easy to use and operate 1 2 3 4 5 > it has a P'eStigious/famous brand 1 2 3 q 5 6 7 namc > its price is expensive 1 2 3 5 . 4 6 7 > there are easily accessible authorized 1 3 5 service centers <1 5 7 > it is manufactured in an advanced/ h" 1 2 3 , 4 .3 6 tech country 122 Extremely Quite Nam" somewh a r 6 ”Id nor Good Extremely 0"": Somewhat Good Good Bad 8:: d Bad °°d 5 6 7 4 8. How 8006 do you think ' t . the Canon brand came he quality of l 2 3 manufactu, ed in China would be? a Canon camera when you need a 9. ow like! is it t h " co le . buy :mem: Y but t e following 9 p “ould think that You shou‘d fiei‘hcr it: attend! . “0' Samaria! 0‘I d" :3 ‘Ttmfl’ Quice somwh" 9":‘3'!’ Likely UR“, U“ nlnkely Unlikcly “may be i 7 6 > your friends 1 2 3 4 5 7 6 >- your family/relatives 1 2 3 A 5 7 6 > your neighbors 1 2 3 A S 7 6 ' y the salesperson 1 2 3 4 5 .ch‘N‘ b¢' \ I {.‘bcd abo vi 0. \Vhen you need a new camera. do you think that buyi mg the Cano“ brand camera d c “f“c°g 1 0““: 9053“" Extremely Quite Somewhat V.“ ‘6“; 9055““ _ . . - her Som‘ 1 ‘\ "V V l. . ' ' fl- ' . x ‘ . . c.1u e \egaus c .\e.,.une rouu“ 6 > beneficla‘ ‘ 2 3 4 5 b 1 'h“ l ‘7 > \Vonh“ e "' 3 4 5 l o , - 6 7 > \\"\SC ' J 4 3 i . > good 2 J 4 5 6 7 123 \’ ‘7 ‘O The follow° "'8 he ' agreement “it Q “was a h . l l. the {Oilowmg “Moments: 3' My decision t . o b described above woz'i’d ":3: whether owning it would hurt re related to the next time .Canon camera influenced by my reputation, Disagru Us) My decision to buy the Canon camera 1 described above would be influel‘ced b svhcther I feel ashamed when 930p": who are important to me see me using ‘his product. My decision to buy the Canon brand camera would beaffected by whether I mink the brand and the price would improve my reputation. I \vould consider buying the Canon Camera described above. Most PCOPIC Who are important to me would mink 1 should buy the Canon brand camera My decision ‘0 buy the Canon camera described above would be influenced by knowing how many of the People who are close to me would also buy ““5 brand, 5105‘ people “'ho are important to me would l think tha! ii is good for me to buy the Canon brand camera described above. NI); decision to buy the Canon camera described 1 above would be influenced by whether owning it \vould make me [it in with pcOple who are close to me. I would recommend the Canon brand camera described above to people Who are close to me. > Next time I intend to buy the Canon camera described above. U U U l\J N 124 Disagree Enremc')’ Quite 0338’“ Sam” so hll 3 o) b) be you would purchase a camera- Neill“, MW“ at!" olsa$n¢ 4 Pleas A‘rcc Sam“ whit 5 Agree Quite e indicate your level 0! Agree Extre only 7 the {cam res listed ' 'th the TV set is ”Isz'grg‘z, evaluate the TV set u in Question, Section "1 Su V o . o . PP 5c that In a department store you find a [V 59" 1 ext ‘ below. Please rea _ . a th . « described in an. 1’01: bzlgscnption carefully because we are going to ask yo LG (Gold Star ) 20" Color TV 5“ i d 'm K0r¢3 _Multi-function remote coon-o1 _High-clear picture . _Manufactur° W - .. ' r - __Split-screen viewing Option _0ne-yeal' 'b\¢ “marked semce centers On-screen display 09“"! Easily aeeesS‘ t em . ; - . s — ' somd ”s - Hi-t't stereo . *‘ - ‘4 12 - Hour likely is it that the LG (Gold Sta r) TV set described above wo 0' '4 MM “may ha".e the r01.0“".“3 Characteristics ' Ev'W'fl)’ Quite Somewhat Likely nor Somewhat Quit: 1)de um"? """k *1 .v Unlikely L'nlikd)‘ Likel!’ we" 1 > it has a high-clear picture I 2 3 4 s 6 1 > it has a quality hi-fi stereo sound I 2 3 4 6 system 5 1 > it has a PreStigious/i‘amous brand ‘ 2 3 b 4 name 5 1 - ' is e ‘ b > its 9““ xPstnswe l 2 3 4 5 1 . . b > there are ca'5\\y accessrble l 3 3 authoriled service centers 4 5 y, it is mantlfactured in an ' 2 3 4 6 7 advanced’hi-tech country 5 Extremelv - - Bad . Eu": Somewhat ‘\ § ith¢f ad Bad ad nor 5mm.“ ’7 . Q . ' Cam] 0W: E", 13. Has" good do you think the qual"? ' 2 3 Good ‘ "My of the LG (Gold 5“") TV 5“ a 5 6 7 manufactured in Korea would be? 125 14’ "W ""°’-" ‘5 “ "'3‘ ”"3 {Mowing maple would think that you should buJ’a LG (Gold 5m) TVS“ “’h" m need a TV: €" tremely Quite Somewhat L""""-" Unlikely Unlikely > your friends I 2 3 > your family/relatives I 2 3 . l > your neighbors 2 3 l 3‘ 3 > the salesperson 15. ll’hcn you need a no“. TV, ho“. do you think that buying Disagree Extremely ; it is beneficial I 1 7, it is \Vorthwhi le 7, it is wise ‘> at is mt l The folloW'mg questions are related to the next time you would agreement with the {dimming statemen ts: 16. > M y decision to buy the LG(Gold SW) TV set described above would be influenced by whether owning it would hurt my reputation. NI)! decision to buy the LG(Gold Star) TV Set described above would be influenced by whether l feel ashamed when pCGP'.e who are important to me see me using ““5 Product. ‘1 Disagree _ Disagree Quttc Some- what 2 3 2 3 2 3 2 3 Disagree Disagree Extremely Quite 1 2 l 2 t 126 V itber ‘Ucltely nor Somewhat Quite Extremely Unlikely Likely Likely Likely 4 5 6 7 4 S 6 7 4 S 6 ‘ 7 4 S 6 7 a LG(Gold Star) brand TV described above would be: Neither AFC! Agree AV“ A3"? nor Some- Quite Extremdy Disagree what 4 5 6 7 4 5 6 7 4 s 6 7 ur p Chase a TV. Please indicale your level of . Neither Din Sonar-r" Act“ Acre: l "Or ’ I"! A 1. Son: (rte H at Disagree c. Quite blue Eur-may 3 4 5 6 7 3 4 5 6 7 ‘I My decision to bU)l the LG (Gold Star) 1 brand TV described above would be 2 affected by whether l think the brand and the price would improve my reputation. b) I?- U: Os \l ‘I 1 would consider buying the LG (Gold 1 5W) brand TV described above. > Most peeple who are important to me 1 2 would think i should buy the LG (Gold Star) brand TV described above. 0) b M > My decision to buy the LG(Gold Star) TV I 2 set described above would be influenced 3 4 by knowing how many of the peOple who are close to me would also buy this brand. > Most people who are important to me l 2 would think that it is good for me to buy 3 4 the LG (Gold Star) TV described above. > My decision to buy the LG(Gold Star) TV I set described above would be influenced 4 7 by whether owning it would make me fit in 5 6 with people who are close to me. > i would recommend the LG (Gold Star) 1 TV described above to peOple who are 4 close to me. > Next time 1 intend to buy the LG (Gold 1 Star) TV described above CONTINUED ON NEXT PAGE 127 ' "" —_ 3 e response that best describes your View b d in your ggnirgl beliefs. Please answer each question with the y ch ' 17. in general, how much do you wan, to 4:32;? (1-5), Never Don't Sometimes Neither sometimes AMI)! “in! Want Don‘t Want nor in... Wm w", Want Don't > your friends think that you should do 3' your relatives think that you should do > your neighbors think that you shou l d do > the salesperson thinks that you shoU‘d do Section V. grcerne . ' ht “1‘ O c e appropriate number. n. h each of the following statements by 0'0"” "19 Mgr" Disan gree D. ' Q“; . ”agree \ - (\‘TCC A‘r“ ‘ 501mm,," ‘ either Ag“: Quite Extremely Extremely Agree nor Suchlll‘ > Korean people should always buy Dinars: domestic products instead of imported l products. 'J Lu 0\ q > Only those products that are l unavailable in Korea should be 4 - imported. id to as \l I») ‘r Buy domestic products. Keep Korea 1 working. <4 5 6 7 > Domestic products are the best. 1 2 3 k) b) \I ‘3' Purchasing imported products is un- l Korean. ‘ 4 5 > It is not right to purchase imported l 2 3 4 products because it puts Koreans out of 5 jobs. 6 > A real Korean should always buy l 2 domestic products. ‘ b) & M > We Should purchase domestic prOdUCtS 1 instead of letting other countries get 6 rich off of us. 7 Id 9) A M I») u A "r To purchase domestic products is 1 always the best. 6 128 > There should be very little trading or purchasing of imported products unless out of necessity. ; l 2 3 4 ‘r" Koreans should not buy imported products because this hurts Korean 1 2 3 4 5 6 7 business and causes unemployment. > Curbs should be put on all im ported l 2 3 4 5 6 7 products. > It may cost me in the long-run, butl l 2 3 4 5 6 7 prefer to support domestic products. > Foreigners should not be allowed to l 2 3 4 5 6 7 put their products on our markets. 7 > Imported products should be taxed l 2 3 4 S 6 heavily to. reduce their entry into Korea. 6 7 > We should buy from foreign countries 1 2 3 4 5 only those products that we cannot ' obtain within our own country. 1 i 6 2 3 4 S > Korean consumers who purchase I imported products are responsible for putting their fellow Koreans out of work. 19. The following questions are related to your familiarity with a country. Extremely Quite Somewhat Neither Unfamiliar Unfamiliar Unfamiliar Familiar" S Cur . . 0’- omculhlf Q . art-uh". 5mm" m.e . Fa . > How familiar are you With I 2 J 4 Inuit.- products made in China? s ’ 6 > How familiar are you with l 2 3 4 products made in Korea? 5 s 129 Section VI. Demographic questions. Please fill in the blank or check the a ' PP r01) hare res ' ponsc [breach ques tron 20. Age: _________ years 21. Gender of person who is completing the questionnaire: Male Female 22. Your present marital status: Single Married 23. Where do you live? gu dong 24- Please check the following category that "105! accurately identifies your total monthly family income- ( 1) less than 1,000,000_ (5) 4,000,00 but less than 5,000,000. (2) 1,000,000 but less than 2,000,000_ (6) 5.000.000 but less than 6,000,000.. (3) 2,000,000 but less than 3,000,000_ (7) 6,000,000_ or more (4) 3,000,000 but less than 4,000,000_ 25. Education (level of degree) (1) some high school (6)__ university student (2) high school degree (7)__ university degree (3) some junior college/no degree (3)__ graduate Student (4) junior college degree (2 year) (9)__ master’s degree or higher (5)___ some University/no degree 6. Occupation:. (1) personal business/business owner, <9 employees ' (2) personal business/business owner, >9 employees (3) sales and service, saleSperson (4) skilled worker/tradesperson, manual worker (S) businessperson, technical worker (6) marketing and management (executive level) (7) specialist or freelancer (ex: professor, medical doctor, lawyer, artist) (8) student (9) unemployed or others Thank you for taking the time to help us with our sun'ey 130 APPENDIX 11 COVARIANCE MATRICES FOR CON F IRMATORYANAL YSIS 131 Covariance Matrices for Confirmatory Analysis Camera (n=456) \ V1 V2 V3 V4 V5 V6 V7 V1 1.460 V2 1.140 1.462 V3 1.112 1.152 1.556 V4 1.130 1.177 1.128 1.594 V5 0.571 0.529 0.602 0.716 1.676 V6 0.601 0.629 0.769 0.779 0.830 1.711 V7 0.701 0.739 0.752 0.786 0.921 1.158 2.039 TV(n=229) ' V1 V2 V3 V4 V5 V6 V7 V8 V9 V1 1.248 V2 0.855 1.132 V3 0.706 0.607 1.397 V4 0.669 0.639 1.207 1.447 V5 0.762 0.658 1.297 1.291 1.709 V6 0.760 0.682 1.241 1.317 1.438 1.731 V7 0.184 0.292 0.509 0.493 0.577 0.621 1.917 V8 0.461 0.395 0.796 0.856 0.899 1.005 0.905 1.735 V9 0.603 0.563 1.040 1.073 1.139 1.233 0.973 1.527 2.117 / Sweaters (n=227) V11 V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V1 1.295 V2 0.600 1.270 V3 0.444 0.478 1.357 V4 0.321 0.486 0.642 1.175 V5 0.400 0.364 0.255 0.339 1.403 V6 0.331 0.381 0.228 0.363 1.186 1.481 V7 0.345 0.268 0.154 0.280 1.163 1.135 1.389 V8 0.273 0.294 0.236 0.385 1.092 1.198 1.173 1.503 V9 0.245 0.287 0.166 0.323 0.766 0.862 0.802 0.828 1.554 V10 0.336 0.272 0.128 0.413 0.979 1.062 1.061 1.125 1.177 1.923 V11 0.362 0.364 0.255 0.475 1.067 1.127 1.088 1.123 1.304 1.707 2.25 Eamon: V1—V4=purchase attitudes l to 4; V5—V7=intention to buy 1 to 3 \4 TV: V1=picture; V2=sound; V3-V6=purchase attitudes 1 to 4; V7—V9=intention to Sweater:V1=color; V2=design; V3=care; V4=comfort; VS-V8=purchase attitude 113, I to 3 S’ intention to buy 1 to 3 I ’0 4 V9 V ; - 1 I § \ 132 APPENDIX III COVARIANCE‘ MATRICES FOR STRUCTURAL EQUATION MODEL 133 Covariance Matrices For Structural Equation Model Product: Camera (n=456) V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12 V13 V14 V1 0.25 V2 0.01 0.85 V3 -0.11 0.31 1.05 V4 -0.02 0.09 0.74 2.49 V5 -0.13 0.38 0.38 0.26 1.28 V6 -0.39 0.19 0.65 0.49 0.91 2.64 V7 -0.90 1.99 4.73 6.01 4.50 6.67 50.49 V8 -0.13 0.24 0.39 0.43 0.46 0.67 3.25 1.46 V9 -0.14 0.27 0.44 0.44 0.54 0.74 3.38 1.14 1.46 ) V10-0.13 0.32 0.49 0.41 0.49 0.68 3.16 1.11 1.15 1.56 V11-0.17 0.29 0.41 0.40 0.43 0.74 3.21 1.13 1.18 1.13 1.59 V12-0.08 0.21 0.24 0.32 0.31 0.30 2.73 0.57 0.53 0.60 0.72 1.68 V13 -0.08 0.19 0.24 0.48 0.32 0.58 2.71 0.60 0.63 0.77 0.78 0.83 1-71 V14-0.13 0.22 0.40 0.39 0.37 0.59 2.51 0.70 0.74 0.75 0.79 0.92 1.16 2-04 v6gtechnical . ‘ 6'» Camera: v1=COM; v2=brand image; v3=p1cture; v4=serviceability; v5=bfand 1:26:35 — '. 'on prestige; v7=overall quality; v8-v11=purchase attitudes 1-4; v12-v14=an// Product: TV (n=229) \N‘ \115 V1 V2 V3 V4 vs V6 V7 vs V9 v10 V11 v12 V13 V1 0.25 V2 0.01 1.10 V3 -009 0.56 1.25 V4 005 0.46 0.86 1.13 vs -0.00 0.45 0.64 0.67 2.06 V6 .007 0.56 0.69 0.62 0.64 1.32 V7 027 0.37 0.67 0.59 0.60 1.02 2.20 V8 007 0.54 0.93 0.92 1.11 0.92 0.98 1.41 V9 011 0.59 0.71 0.61 0.63 0.72 0.80 0.80 1.40 V10-0.15 0.56 0.67 0.64 0.63 0.72 0.88 0.82 1.21 1.45 V11-0.17 0.63 0.76 0.66 0.66 0.78 0.95 0.87 1.30 1.29 1.71 V12 -0.19 0.62 0.76 0.68 0.60 0.80 0.95 0.84 1.24 1.32 1.44 1,73 V13 -0.08 0.29 0.31 0.29 0.38 0.38 0.44 0.31 0.51 0.49 0.58 0.62 19 v14 -0.15 0.44 0.46 0.40 0.40 0.53 0.60 0.48 0.80 0.86 0.90 1.01 0:92 v15 -0.15 0.49 0.60 0.56 0.64 0.72 0.65 0.53 1.04 1.07 1.14 1.23 0.91 1.74 TV: v1=COM; v2=brand image; v3=p1cture; v4=sound; v5=serviceability; V6=brand 1.53 2.12 prestige; V3=overall quality; v9-v12=purchase attitudes 1-4; v13-v15=intention to buyrleitige; V7‘te hr) - “ c . lea] \ 134 Covariance Matrices for Structural Equation Model C o n t’d ___\ __‘_—_______/ Product: Sweater (n=227) A v1 V2 v3 V4 V5 V6 V7 V8 5 v16 V1 0.25 V2 -001 1.13 V3 -005 0.21 1.30 V4 .001 0.30 0.60 1.27 V5 -001 0.08 0.44 0.48 1.36 V6 003 0.30 0.32 0.49 0.64 1.18 V7 003 0.49 0.35 0.42 0.18 0.41 1.31 V8 037 0.41 0.35 0.34 0.04 0-13 0.82 2 V9 -0.48 2.79 4.09 4.21 3.48 4-08 . '35 ‘Vlo -0.04 0.49 0.40 0.36 0.26 334 0.45 0 11 -007 0.40 0.33 0.38 0.23 -36 0.50 ' V12 -005 0.39 0.35 0.27 0.15 0.28 0.42 (if; 5?: {'12 fig; 1.39 V13 -0.04 0.45 0.27 0.29 0.24 0.39 0.51 0.49 2'91 109 1120 1.17 1.50 55 V14 -005 0.45 0.25 0.29 0.17 0.32 0.45 0.59 2:49 0:77 0-86 0.80 0.83 1-18 1.92 225 V15 -005 0.55 0.34 0.27 0.21 0.41 0.55 0.73 2.84 0.98 1-06 1.06 1.13 ‘- .7 . V16 -0.05 0.66 0.36 0.36 0.26 0.48 0.59 0.85 3.53 1.07 1-13 1.09 1.12 ‘g Sweater: v1=COM; V2=brand image; v3=color; v4=f‘ashion; v5f'care; v6=com10m 7111609 v8=technical prestige; v9=overall quality; v10-v13=p urchase attitudes 1.4; “4511 410 V9 V10 v11 V12 V13 V14 V1 135 APPENDIX IV INPUT SEM MODELS FOR EQS 136 ITITLE Camera:SEM: Raw quality dimension /SPECIFICATIONS VARIABLES=14; CASES=456; METHODS=ML; MATRIX=CORRELATION;AN=COV; /Equations v1 = fl + e1; V2= f2+ e2; v3= f3+ e3; v4= f4+ e4; v5= f5+ e5; v6= f6+ e6; v7= t7+ e7; v8= f8+e8; v9= ‘f8 +e9; v10=*f8+e10; vll =‘f‘8+e11; v12= 19+elZ; vl3=*f9+e13; vl4=*F9+el4; f3 = ‘fl + *f2 +dl; f4 = *f] + *0 +d2; f5 = *fl + ‘12 +d3; f6 = *fl + *12 +d4; f7=*f1 +*f2+*f3+*f4+*f5 + *f6+d5; f8=*f1+*fz+*f7+d6; f9=*fl+*f2,+*f8+d7', N ARIANCES fl to f2 = 0.1“, e1 to e7 = 0.0; e8 to e14 = 0.1", (11 to d7 = 6““. Icovariances dl,d2=3*; d3,d4=3*; llabels V l=country; v2=brimage; v3=piCt; V4=SCW; v5=pres; V6=tech' v7= V9=att2; , quality v10=att3; vl l=att4; v12=intl; v13=iflt2; vl4=int3; ’ Vghattli f1=com; f2=brand; Bzworkman; f4=service; f5=prestige; f6=techn' f8=attitude; f9=intent; lea; ”toga . /constraints 81113,; !(f3,fl)=(f3,t2); -' (f4,f1)=(f4,12); ’(f5,fl)=(f5,f2); 137 !(f6,fl)=(f6,f2); /Matrix 1.000 0231.000 -.206 .324 1.000 -.027 .061 4591.000 -.226 .359 .325 .145 1.000 -.481 .128 .389 .191 4961.000 -.252 .303 .649 .536 .561 .577 1.000 -.216 .218 .318 .223 337.339 .379 10 -.227 .237 .356 .231 .393 .377 393- 00 -.211 .274 .381 .207 .346 .335 .356 7801.000 -.261 .247 .317 .200 .303 .360 .353 '738 .764 1.000 -.128 .171 .180 .156 .215 .142 .297 ‘3741 -771 film-00° 0 -.119 .159 .177 .231 217.274 .291 ‘365 .338 .373 .4381000‘000 -.181 .167 .275 .172 .227 .256 .247 :4 80 393 .471 .472 A90 - 010000 /Standard deviation 6 7 O6 '428 .422 .436 .498 52 ' .5005 .9243 1.0251. 771.1291.62 .1055 1.20 /LMTEST 5 831‘2091 1.24751.26251.29471.30811.428 /PRINT effect=yes; digit=3; linesize =80; fit=all; ITECHNICAL iteration= 500', IEND 138 [TITLE TV:SEM: Raw quality dimension /SPECIFICATIONS VARIABLES=15; CASES=229; METHODS=ML; MATRIX=CORRELATION;AN=COV; lEquations v1 = fl + e1; v2 = f2 + e2; V3 = B + 63; v4 = *3 + e4; v5 = f4 + e5; v6 = f5 + e6; v7 = f6 + e7; v8 = 17 + 68; v9 = f8 + e9; v10 = *f8 +610; v1] = *fB + ell; v12 = *t8 + e12; v13 = t9 + e13; v14 = ‘1‘9 + e14; v15 = *9 + e15; f3 = *fl + *12 + d1; f4 = ‘fl + *12 + d2; f5 = *fl + ‘12 + d3; f6=*f1 +*12+d4; f7=*f1 +*f2+*f3+*f4+*f5 + *f6+d5; f8=*f1+*f2+*f‘l +d6', f9=*f\ +*f2+*f8+d7', NARIANCES f1 to 12 = 1*; e1 to e2 = 0.0; e5 to e8 = 0.0; e3 to e4 = 1*; e9 to e15 = 1"; dl to d7 = 1*; /covariances; dl,d2 = 1"; d3,d4 = 1*; konstraints ’(B,fi)=(f3.0); !(f4,t‘1)=(f4,12); ’(f5,fl)=(f5,f2); 106.0 )=(f6,f2); /1abe1s V1=°°untry; V2=brimage; v3=picture; V4=50‘md; V5=Service; v6=pre t S 3' 139 v7=tech; v8=quality; v9=attl; v10=att2; vll=att3; v12=att4 f1=com;12=brand;13=workman; f4\s t7=0quality; f8=attitude; f9=intenton; erttljengivm‘mQ v151nt3; \Prestige; f6-techn1'ca; /Matrix 1.000 0101.000 -.152 .476 1.000 -.097 .414 7191.000 -.006 .296 .399 4371.000 -.114 .467 .535 .504 .390 1.000 -.370 .240 .403 .377 .283 5961000 -.121 .435 7.02 .730 6.53 676 558 1.000 -.187 .477 .535 .483 .369 .523 455 567 1. 0 -.251 .445 .498 .499 .365 .517 «494 .570 80" -.265 .461 .522 .473 .353 .517 .488 .557 .‘839 1.000 -.292 .451 .517 .487 .319 -530 .487 .537 .79: '2?" L000 - .118 .197 .199 .198 .193 .238 .216 .186 -3 11 '29: .836 L000 -.227 .319 .313 .282 .212 .351 .306 .309 ,5 11 340 .319 3411.000 -.211.319 .371 .364 .306 .431 .302 .308 -605 .513 '522 -580 .4961.000 1000 /Standard deviation ' ~599 644 .483 .797 - 4549 31' .501.05071.ll71.0641.4341.149 1.482 1.1891.1820 ILMTEST 1'2030 ”074 1 31561384413” IPRINT effect = yes; digit=3; linesize =80; fit=a11', ITECHNICAL iteration= 500', IEND 140 ITITLE Sweater:SEM : Raw quali dim ens' . /SPECIFICATIONS W m"’ delete fiber 67) VARIABLES=16; CASES=22 7 ; METHODS=ML; ‘ MATRIX=CORRELATION;AN=COV; /Equations V1 = fl + e]; V2 = f2 + 02; V3 = B + e3; V4 = ‘13 + e4; V5 = f4 + e5; v6 = ‘f4 + e6; v7 = f5 + e7; v8 = f6 + e8; v9 = 17 + e9; v10 = f8 + e10; VI] = *f8 + e11; v12 = *fB + e12; v13 = ’18 + e13; v14 = f9 + e14; v15 = ‘19 + e15; v16 = *f9 + e16; f3 = ‘fl + *12 + d1; f4 = ‘fl + ‘12 + d2; f5="‘f1 +‘f2+d3', f6=*f1 +*12+d4; f7=*f1 +*t2+*f3+*f4+*f5+ *f6+d5; f8=*f1+*f2+*f7 +d6; f9=*f1+*f2+*f8+d7; NARIANCES f1 t0 t2 = *; e1 t0 e2 = 0.0; e3 to e6 = 05*; e7 t0 e9 = 0.0; e10 t0 e16 = 05*; d1 t0 d7 = 2*; /covariances d3,d4=2*; /Constraints !(B’fl)=(f33f2); !(f4.f1)=(f4,12); !(f5,r1)=(fs,12); !(f6,fl )=(f6,f2); /LABELS v1=country; v2=brimage; v3=design; V4=00101’; V5=care; V6=Comf0rt- 141 v7=prest;v8=tech;v9= uali ; v1 = _ vl4=int1; v15=int2; vl6gint3;t-y O att], v11 fl=com; f2=brand;f3=design; f4==w , f8=attitude; t9=intent0n; Orkman, f5=Prestigef6=technjcak f7=oqualityi /Matrix 1.000 -.0151.000 -.083 .1761.000 -.107 .252 4681.000 -.012 .061 .335 3641.000 -.059 .260 .260 .368 5081.000 -.058 .403 .269 .324 .136 .327 1.000 -.488 .252 .201 .198 .020 .075 463 1.000 -.151 .412 .564 .586 .469 .591 .578 .4271000 -.071 .387 .297 .273 .185 .264 .335 .273 .368 1 -.120 .306 .239 .278 .161 .275 .361 .306 .380 £31310] -.083 .309 .257 .202 .112 .219 .312 .264 .336 .3 3 3 7'90100 -.064 .342 .196 .213 .165 .290 .363 .261 .372 -752‘ 8031000 -.082 .336 .173 .204 .114 .239 .315 .310 .313 ,5 ,9 .568 3‘2 1-000 -.075 .370 .213 .174 .129 .275 .347 .344 .321 -596 ‘629 -546 -5421000 0 -.067 .413 .212 .215 .146 .292 .346 .370 .369 -600 '61—, '349 662-681 1002 1.000 /Standard deviation . . 15 ‘610-697 '80 3867 .50 1.0634 1.1381.1271.165 1-0841.1441.5346.37()41.1843 12 31.2466" 1.5013 ' 171117361225 ILMTEST IPRINT effect=yes; digit=3; linesize =80; fit=al1; ITECHNICAL iteration= 500', IEND =att2; V12=att3; v13=att4 J 142 REFERENCES 143 REFERENCES Aguilar, L. M. & Singer, M. A. (1996). Big Emerging Markets and US. Trade. Competitiveness Review, 6(1), 14-20. Ahmed, S. A. & d’Astous, A. (1996). Country-of-Origin and Brand Effects: A Multi- Dimensional and Multi-Attribute Study. Journal of International Consumer Marketing, _9_(2), 93-1 15. Ahtola, O. T. (1984). Price as a Given Component in an Exchange Theoretical Multicomponent Model. In T. C. Kinnear (Ed), Advances in Consumer Research: Vol. 11, (pp. 623-640). Ann Arbor, MI: Association for Consumer Research. Alba, J. W. & Hutchinson, J. W. (1987). Dimensions of Consumer Expertise. Journal of Consumer Research, 13(4), 411-455. Anderson, J. C. & Gerbing, D. W. (1988). Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach, Psychological Bulletin, 1921(3), 41 1-23 . Anonymous. (1997, November, 29). The End of the Miracle. The Economist, 21'23- Anonymous. (1998). Keeping Promises: Today’s Currency — Tomorrow’s Future of Branding. Discount Store News, 3_7(17), Sl l-Sl 8. Assael, H. (1992). Consumer Behavior and Marketing Action (4th ed.). Boston: PWS- Kent Publishing. Bannister, J. P. & Saunders, J. A. (1978). UK. Consumers’ Attitudes Towards Imports: The Measurement of National Stereotype Image. European Journal of Marketing, 12(8), 562-570. Bentler, P. M. (1993). EQS: Structural Equations Program Manual. Los Angeles, CA: BMDP Statistical Software, Inc. ' Bettman, J. R. & Park, C. W. (1980). Effects of Prior Knowledge and Experience on Consumer Decision Processes: A Protocol Analysis, Journal of Consumer Research, 2, 234-248. J Bilkey, W. & Nes, E. (1982, Spring/Summer). Country-of-Origin Effects on Product Evaluations. Journal of International Business Studies, 89-99. 144 Bollen, K. A. (1989). Structural Equations with Latent Variables, New York: A Wiley- Interscience Publication. Brodowsky, G. H. (1996). The Role of Country of Origin in Consumer Purchase Decisions: Development and Testing of A Comprehensive Theoretical Model, Doctoral Dissertation, The State University of New York at Buffalo. Brucks, M. & Zeithaml, V. (1991). Price and Brand Name as Indicators of Quality Dimensions. Marketing Science Institute Working Paper, No. 91-130. Bureau of Economic and Business Affairs. (2002, February). 2001 Country Reports on Economic Policy and Trade Practices, US. Department of State. Retrieved March, 2002, from World Wide Web: http://www.state.gov/documents/organization/ 8 1 68 .pdf Cattin, P. Jolibert, A. & Lohnes, C. (1982). A Cross-Cultural Study of “Made in” Concepts, Journal of International Business Studies, 13(3), 131-141. Chao, P. (1989). The Impact of Country Affiliation on the Credibility of Product Attribute Claims. Journal of Advertising Research, @(2), 35-41. (1993). Partitioning Country of Origin Effects: Consumer Evaluations of a Hybrid Product. Journal of International Business Studies, 24(1), 291-306- (1998)~ 1111135119t Of Country Of Origin Dimensions on Product Quality and Design Quality Perceptions. Journal of Business Research, 4_2_, 1-6- Cordell, V. (1992). Effects of Consumer Preferences for Foreign sourced Products. Journal of International Business Studies, _2_3(2), 251-269, Compeau, L. D., Grewal, D., &Monr0e, K. B. (1998). Role of Prior Affect and Sensory Cues on Consumers’ Affective and Cognitive Responses and Overall Perceptions of Quality. Journal of Business Research, 4_2_, 295-308. Darby, M. & Karni, E. (1973, April). Free Competition and the Optimal Amount of Fraud Journal of Law and Economics, 16, 67-86. Dodds, W. B., Monroe, K. B. & Grewal, D. (1991, August). Effects of Price, Brand and Store Information on Buyers’ Product Evaluation. Journal of Marketing Research, 28, 307-319. Eastman, J. K., Goldsmith, R. E., & Flynn, R. (1999). Status Consummion in Consumer Behavior: Scale Development and Validation. Journal of Marketing Theory and Practice, 2(3), 41-52. 145 Erickson, G. M., Johansson, .1. & Chao, P. (1984, September). Image Variables in Multi- Attribute Product Evaluations: Country-of-Origin Effects. Journal of Consumer Research, 11, 694-699. Ettenson, R. Wagner, J. & Gaeth, G. (1988, Spring). Evaluating the Effect of Country of Origin and the “Made In USA.” Campaign: A Conjoint Approach. Journal of Retailing, 64, 85-100. Fishbein, M. & Ajzen, I. (1975). Belief Attitude, Intention, and Behavior: An Introduction to Theory and Researchz Reading, MA.: Addison Wesley Publishlng Company. Gaedeke, R. (1973, Summer). Consumer Attributes Towards Products ‘Made In’ Developing Countries. Journal of Retailing, 19, 13-24. Garten, J. (1997). The Big Ten: The Big Emerging Markets and How They Will Change Our Lives. New York: Basic Books. Garvin, D. A. (1984, Fall). What does Product Quality Really Mean? Sloan Management Review, 25—43. (1987, November-December). Competing on the Eight Dimensions of Quality. Harvard Business Review, 66, 101-109. v, Han, M. C. & Qualls, W. J. (1985). Country of Origin Effects and Their Impact upo'n Consumers’ Perception of Quality. IN A. A, Michell (13¢), Advances "1 Consumer Research: Vol. 9 (pp.162-167). Ann Arbor, MI: Assomation for Consumer Research. (1989, May). Country Image: Halo or Summary Construct? Journal of Marketing Research, 66, 222-229. & Terpstra, V. (1988, Summer). Country-of-Origin Effects for Uni-National and Bi-National Products, Journal of International Business Studies, 235-254. Heimbach, A. E., Johansson, J. K., & MacLachlan, D. L. (1988, October). Product Familiarity and Country of Origin Cues, Presented at the Association for Consumer Research, Annual Conference, Honolulu, Hawaii. Heslop, L., Liefeld, J., & Wall, M. (1987, June). An Experimental Study of the Impact of Country of Origin Information. In R. E. Turner (Ed.), Marketing: Vol. 8 (pp. 179-185). Toronto, Ont.: Proceedings, Administrative Sciences Association of Canada-Marketing Division. Holstein, W. J. & Nakarmi, L. (1995, July). Korea. Business Week, 56-64. 146 Hong, S. & Wyer, R. Jr. (1989, September). Effects of Country of Origin and Product Attribute Information on Product Evaluation: An Information Processing Perspective. Journal of Consumer Research, 16, 175-1 8 7. (1990, December). Determinants of Product Evaluation: Effects of the Time Interval between Knowledge of a Product’s Country of Origin and Information about Its Specific Attributes. Journal of Consumer Research, 1’_7, 277-288. Hwang, S. J. (1997, May). Personal Interview at Sung Kyun Kwan University in Seoul, Korea. Iyer, G. R. & Kalita, J. K. (1997). The Impact of Country-of-Origin and Country-0f- Manufacture Cues on Consumer Perceptions of Quality and Value, Journal of Global Marketing, 11(1), 7-28. Jaffe, E. D. & Nebenzahl, I. D. (1984, November). Alternative Questionnaire Formats for Country Image Studies. Journal of Marketing Research, 21, 463-471 . Johansson, J. K., Douglas, S. P. & Nonaka, I. (1985, November). Assessing the Impact of Country-Of-Origin on Product Evaluations: A New Methodological Perspective. Journal of Marketing Research, 22, 388-396. & Nebenzahl, 1- (1936, Fall). Multinational Production: Effect on Brand Value, Journal of International Business Studies, 1_7, 101-126, . (1989). Determinants and Effects of the Use of “Made-In” Labels, International Marketing Review, 6(6), 27-41. Johnson, E. J. & Russo, J. E. (1984, June). Product Familiarity and Learning New Information. Journal of Consumer Research, 111, 542-5 50. Kim, S. & Pysarchik, D. T. (2001). Predicting Purchase Intentions for Uni-National and Bi—National Products. International Journal of Retail & Distribution Management, 2_8 (6), 280-291. Kirmani, A., Sood, S., & Bridges, S. (1999). The Ownership Effect in Consumer Responses to Brand Line Stretches. Journal of Marketing, 63(1), 88-101. Knox, S. & Chematony, L. (1989). The Application of Multi-Attribute Modeling Techniques to the Mineral Water Market. Quarterly Review of Marketing, 15(4), 14-20. V Lee, C. & Green, R. T. (1991, Second Quarter). Cross-Cultural Examination of the F ishbein Behavioral Intentions Model. Journal of International Business Studies, 289-304. 147 Lee, C. (1997, May). Personal Interview at Hong-Ik University in Seoul, Korea. Li, A. & Dant, R. (1997). Dimensions of Product Quality and Country-Of-Origin Effects Research. Journal of International Consumer Marketing _l_O_(l/2), 93-114. V Liefeld J. (1993). Country-of-Origin Effects: Review and Meta-Analysis of Effect Size. In N. Papadopoulos & L. Heslop (Eds), Product and Country Images: Research and Strategy (pp. 1 17-156). New York: Haworth Press. Mindak, W. A. (1961, April). Fitting the Semantic Differential to the Marketing Problem, Journal of Marketing, 26, 28-33. '“ Mitchell, A. A. & Olson, J. C. (1981). Are Product Attribute Beliefs the only Mediator of Advertising Effects on Brand Attitude? Journal of Marketing Research, 13(3), 312-332. Mittal, B. (1990). The Relative Roles of Brand Beliefs and Attitude Toward the Ad as Mediators of Brand Attitude: A Second Look. Journal of Marketing Research, 21, 209—219. Monroe, K. B. & Krishnan, R. (1985). The Effect of Price on Subjective Product Evaluations. In J. Jacoby & J. Olson (Eds), Perceived Quality: How Consumers View Stores and Merchandise (pp. 209-232). Lexington, MA: Lexington Books- Mowen, J. C. (1990). Consumer Behavior (2“d ed.). New York: Macmillan Publishing Company. Nagashima, A. (1970, January). A Comparison of Japanese and US Attitudes toward Foreign Products. Journal of Marketing, 35, 68-74, (1977). A Comparative Product “Made In” Image Survey Among Japanese Businessmen. Journal of Marketing, fl, 95-100. Narayana, C. L. (1981, Summer). Aggregate Images of American and Japanese Products: Implications on International Marketing. Columbia Journal of World Business, 16, 31-35. Nebenzahl, I. D., Jaffe, E. D., & Lampert, S. I. (1997). Towards a Theory of Country Image Effect on Product Evaluation. Management International Review, 32(1), 27-49. Negley, J. (1999). The “Brand” Wagon. Discount Store News, 38(9), 11-12. Nelson, P. (1970). Information and Consumer Behavior. Journal of Political Economy, 7_8_(2), 31 1-329. 148 (1974). Advertising as Information, Journal of Political Economy 8_l_, 729- 754. Nes, E. & Bilkey, W. (1993). A Multi-Cue Test of Country-of-Origin Theory. In N. Papadopoulos & L. Heslop (Eds), Product-Country Images: Impact and Role in International Marketing (pp. 179-196). Binghamton, NY: International Business Press. ‘/ Netemeyer, R. G. & Bearden, W. O. (1992). A Comparative Analysis of Two Models of Behavioral Intention. Journal of the Academy of Marketing Science, 26(1), 49-59. '/ Obermiller, C. & Spangenberg, E. 1989. Exploring the Effects of Country-of-Origin Labels: an Information Processing Framework. Advances in Consumer Research, 16, 454-459. Olson, J. C. & Jacoby, J. (1972). Cue Utilization in the Quality Perception Process, In M. "“‘ Venkatesan (Ed.), Proceedings, Third Annual Conference of the Association for Consumer Research. Osgood, C. E. (1952, May). The Nature and Measurement of Meaning, PSyChOIOBICUI Bulletin, 19, 197-262. Park, 0. w. & Lessig, V. P. (1981). Familiarity and Its Impact on Consumer Biases and Heuristics. Journal of Consumer Research, 3(2), 223-230. Rao, A. R., Qu, L. & Ruekert, R. w. (1999). Signaling Unobservable Product Quality through a Brand Ally. Journal of Marketing Research, 36(2), 258-268- & Monroe, K. B. (1989, August). The Effect of Price, Brand Name, and Store Name on Buyer’s Perceptions of Product Quality: An Integrative Review. Journal of Marketing Research, 2_6, 351-357. / Reierson, C. November, (1967). Attitude Changes toward Foreign Products. Journal of Marketing Research, 5: 385-397. Roth, M. & Romeo, J. (1992). Matching Product Category and Country Image Perceptions: A Framework for Managing Country-of—Origin Effects. Journal of International Business Studies, 2_3(3), 477-497. V Samiee, S. (1994, Third Quarter). Customer Evaluation of Products in a Global Market, Journal of International Business Studies , 579-602. Schooler, R. (1965, November). Product Bias in the Central American Common Market, Journal of Marketing Research, 3, 394-397. 149 (1971, Spring). Bias Phenomena Attendant to the Marketing of Foreign Goods in the US. Journal of International Business Studies, 71-80. & Wildt, A. (1968, February). Elasticity of Product Bias. Journal of Marketing Research, 6: 78-81. Schuman, M. (1996, Septermber 11). US. Companies Crack South Korean Market: Ford, Whirlpool, Estee Lauder Target Affluent Consumers, Wall Street Journal, A14. Sheppard, L., Hartwick, P. & Warshaw, R. (1988). The Fishbein Extended Model and Consumer Behavior, Journal of Consumer Research, 3(3), 118-236. Shimp, T. A. & Shanna, S. (1987, August). Consumer Etlmocentrism: Construction and Validation of the CETSCALE. Journal of Marketing Research, 341, 280-289. V' 81, S. & Cullen, J. B. (1998). Response Categories and Potential Cultural Bias: Effects of an Explicit Middle Point in Cross-Cultural Surveys. International Journal of Organizational Analysis, 6(3), 218-230. Sirgy, M. J. (1982). Self-Concept in Consumer Behavior: A Critical Review. Journal Of Consumer Research, 2, 287-300. Steenkamp, J. (1990). Conceptual Model of the Quality Perception Process. Journal of Business Research, 2_1, 309-333. Szybillo G. J. & Jacoby, J. (1974, February). Intrinsic Versus ExtrinSiC Cues as Determinants of Perceived Quality. Journal of Applied Psychology, 39, 74-78. Thakor, M. & Katsanis, L. (1997). A Model of Brand and Country Effects on Quality Dimensions: Issues and Implications. Journal of International Consumer Marketing, 2(3), 79-100. The Office of Customs Administration (1991). Country-Of-Origin Labeling Regulations, Industrial Information Network. Retrieved April, 15, 2002, from World Wide Web: http://magic.iin.co.kr/db_total/muyok06.htm. \ . \// Tse, D. K. & Gorn, G. J. (1993). An Experiment on the Salience of Country-of-Origin in the Era of Global Brand. Journal of International Marketing, 1(1), 57-75. (1993). Removing Negative Country Images: Effects of ‘ __ & Lee. W. Journal of International Decomposition, Branding and Product Experience. Marketing, 1(4), 25-48. 150 V Ulgado, F. M. & Lee, M. (1993). Consumer Evaluations of Bi-National Products in the Global Market. Journal of International Marketing, 1(3), 5-22. (1998). The Korean versus American Marketplace: Consumer Reactions to Foreign Products. Psychology & Marketing, 16 (6), 595-614. US. Census Bureau (2001). Exports, Imports and Trade Balance by Country and Area, Not Seasonally Adjusted: 2001. Retrieved February, 2002, from World Wide Web: http://www.census.gov/foreign-trade Verlegh, P. W. J. & Steenkamp J. E. M. (1999). A Review and Meta-Analysis of Country-of-Origin Research. Journal of Economic Psychology, _2_Q, 521-546. 1/ Wall, M. & Heslop, L. (1986). Consumer Attitudes Toward Canadian-Made versus Imported Products, Journal of Academy of Marketing Science, 141(2), 27-3 6. Liefeld, J., & Heslop, L. (1989). Impact of Country of Origin Cues and In A. d’Astous Patriotic Appeals on Consumer judgments: Covariance Analysis. (Ed), Marketing: Vol. 10 (pp.306-315). Montreal, Canada: Administrative Sciences Association of Canada-Marketing Division. Origin Cues on (1991). Im act of C -of- p ountry Journal of Consumer Judgments in Multi-cue Situations: A Covariance Analysis, the Academy of Marketing Science, 1_9_ (2): 105-113. Wilson, D., Mathews, H. L., Harvey J. (1975). An Empirical Test of the Fishbein Behavioral Intention Model. Journal of Consumer Research, 1(4), 39'49- Zeitharnl, V. (1988, July). Consumer Perceptions of Price, Quality, and Value: A Means-End Model and Synthesis of Evidence, Journal of Marketing, 2, 2-2. 151