|
Article Excerpt
Although there are numerous studies related to country-of-origin (COO) effects, empirical findings are dispersed because of the limited coverage of the origins, brands and countries used for investigation. This paper uses an existing data set that consists of a survey conducted across 20 nations to evaluate 11 automobile origins with 53 brands. This data set facilitates the verification of COO effects previously addressed in the literature from a holistic viewpoint. It also provides insight into the circumstances under which, and the extent to which, the COO effects could differ. The results derived from correspondence analysis (CA) suggest there are brand and national variations in the magnitude of COO effects. At brand level, COO effects appear to be more influential on the purchase behaviour of consumers who have a positive attitude towards the brand being investigated or perceive it to be of high quality. At the national level, COO effects seem to be more significant among nations where the availability of international automobile brands is lower. Furthermore, the findings not only support the notion that consumers tend to have a stronger preference for products that originate from their own countries, but also concur with the proposition that consumers also tend to have a stronger preference for products from countries in the same geographic region.
**********
Introduction
Since the mid-1960s, country-of-origin (COO) effects have been the subject of many studies. While it is widely acknowledged that COO has an impact on product evaluations, there have been divergent empirical findings concerning the magnitude of the COO effect when different brands are used (Erickson & Johansson 1992; Tse & Gorn 1992; Ettenson 1993; Hulland 1999) for consumers from different nations (Sharma et al. 1995; Gurhan-Canli & Maheswaran 2000; Amine & Shin 2002) to evaluate products of different origins (Papadopoulos et al. 1990a, b). It is therefore desirable to conduct further empirical investigation to validate research findings in a multi-brand, multi-origin and multi-nation context. The objective of this research is to address these issues by employing empirical data from the international automobile industry. In particular, it attempts to:
1. verify the notion that COO has a positive impact on consumers' purchase behaviour across a wide range of nationalities, brands and origins;
2. explore COO effects in relation to brand associations;
3. explore COO effects in relation to market conditions;
4. measure the degree of ethnocentric tendency and verify the notion of consumers' ethnocentric tendencies across various nationalities.
While most existing research uses experimental designs to examine COO effects, this study employs the alternative of a survey method and adopts the well-established correspondence analysis (CA) technique to analyse the data in response to the above questions. The CA technique has frequently been utilised to generate perceptual maps that portray the spatial location of categories (e.g. brands and product attributes) based on joint-space comparisons (Carroll et al. 1986).
This study posits that the application of CA could facilitate analyses of the relationships among origins, brands and nationalities for the data collected from cross-national survey research. For instance, to examine the extent of ethnocentric tendency, CA is performed on 20 nationalities and 11 origins to depict the closeness between consumers' nationalities and the countries of origin they favour. The greater the extent of a nation's ethnocentrism, the closer (more distant) nationality will be related to the domestic (foreign) origin of a product. Similarly, to explore COO effects, another CA is performed on 53 purchased brands and 11 origins, with the aim of demonstrating the relationships between brand purchase behaviour and favourable COO effects. The greater the extent of the COO effect, the closer the purchased brand will be related to the favoured origin, and vice versa.
In addition to applications of closeness/distance among the categories mentioned above, this study exploits two other important statistics derived from CA: inertia and contributions of points to inertia. In this study, total inertia indicates the strength of the relationship between favourable COO and the purchased brands on a country-by-country basis, while contributions of points to inertia aid the assessment of the relative contributions of respective brands and nations to the identified relationship in each country. Categories (i.e. brands and origins) that contribute substantially to the inertia of a particular dimension are considered to be playing dominant roles in the relationship.
The study is organised as follows. First, the relevant literature regarding COO effects is reviewed. CA analysis is then introduced as an alternative means to examine COO effects. The data set employed in this study is described, and the sample composition is presented. The results and implications are then presented and discussed. Finally, a summary of the limitations of this study and suggestions for future research are presented.
Literature review
According to Heslop and her colleagues (Heslop et al. 1998), there were more than 300 COO-related studies as of mid-1997. In these studies, COO is defined in many ways, and the definitions can be classified into three groups: overall country image, aggregate product-country image, and specific product-country image. Given that COO effects are found to vary considerably according to the product group under consideration (Roth & Romeo 1992), this study follows the third definition, by which COO refers to information pertaining to where a particular product is 'made'.
The role of COO in information processes and its impact on product evaluation are acknowledged in many studies (Al-Sulaitu & Baker 1998; d'Astous & Ahmed 1999). Various authors suggest that COO affects product evaluation and subsequent purchase behaviour not only directly but also indirectly through beliefs (Erickson et al. 1984; Han 1989; Hong & Wyer 1989). For instance, Han (1989) identifies two major functions of COO information: halo function and summary function. The halo function affects a consumer's evaluation indirectly through beliefs, whereas the summary function affects the consumer's evaluation directly (Figure 1).
According to Han, consumers can use COO as a cue to generate inferential beliefs when other information is not available. An initial awareness of a product's COO often triggers a global evaluation by which consumers infer all manner of quality, performance or product attributes they are unable to evaluate directly (Strutton et al. 1994). These inferred evaluations reflect halo effects or stereotyping. The other function of COO effects is derived by the summary function, where COO helps consumers summarise product information as they become more familiar with a country's products. Consumers are particularly prone to use COO category--by which the information is summarised--to streamline their decision making when the amount of attribute information is large and otherwise difficult to integrate (Huber & McCann 1982; Bodenhausen & Lictenstein 1987).
Figure 1 summarises the total COO effect on consumers' purchase decisions, which consists of the direct effect, COO [right arrow] purchase behaviour ([[gamma].sub.pc]), and three indirect effects: (1) COO [right arrow] evaluation/attitude [right arrow] purchase behaviour ([[gamma].sub.oc] X [[beta].sub.po]); (2) COO [right arrow] beliefs [right arrow] purchase behaviour ([[gamma].sub.bc] X [[beta].sub.pb]); and (3) COO [right arrow] belief [right arrow] evaluation/attitude [right arrow] purchase behaviour ([[gamma].sub.bc] X [[beta].sub.ob] X [[beta].sub.po]), which are hierarchical effects. The direct effect (e.g. Hong & Wyer 1989) suggests that a consumer's attitude towards a brand's COO is positively related to his or her brand purchase behaviour, with other things being equal.
P1: Consumers tend to purchase the brand whose COO they have a positive attitude towards.
On the other hand, the indirect effects of COO are the functions of brand belief and attitude. It is evident from previous studies that consumers seek and use extrinsic cues such as COO and brand name during the product evaluation process (Han & Terpstra 1988; Cordell 1991; Leclerc et al. 1994; Ahmed & d'Astous 1995; Kim & Chung 1997; Hulland 1999), and rely on these extrinsic cues to draw inferences. Although it has been suggested that COO effects on consumers' product evaluations could be moderated by brand name (Han & Terpstra 1988), the results remain inconclusive on the kinds of brand, in terms of brand associations, on which the COO effect is greater.
Examining COO effects becomes complicated when brand associations are taken into consideration. As shown in Figure 1, indirect COO effects are determined by the interactions among COO, brand beliefs and evaluation/attitude, suggesting that positive brand evaluation/attitude and beliefs will ultimately enhance the total effect of COO on purchase behaviour. Based on brand associations' levels of abstraction, Keller (1993) indicated different types of brand associations such as attributes, benefit and attitude. Associations at lower levels of abstraction such as attributes must be integrated by corresponding evaluation to form a preference. On the other hand, benefit and attitude, which are the 'summarising' and 'evaluative' information, have a more immediate impact on subsequent purchase behaviour than do attributes (Wu et al. 1988). As such, this study focuses on examining only brand associations at higher levels of abstraction, including brand attitude and perceived quality. It posits that positive brand attitude and perceived quality will, ultimately, contribute to the total COO effect on purchase behaviour. The following research propositions summarise the anticipated impacts of these two factors.
P2.1: The COO effect on purchase behaviour is more likely to be greater when the consumer has a positive brand attitude towards the brand being investigated.
P2.2: The COO effect on purchase behaviour is more likely to be greater when the brand being investigated is perceived to be of high quality.
Next, as the...
|
