Sales Promotion Effectiveness: The Impact of Culture at an Ethnic-Group Level

Simon Kwok & Mark Uncles

School of Marketing University of New South Wales Sydney NSW 2052 Australia

School of Marketing Working Paper 02/4 2002

Address for correspondence: Professor Mark Uncles, School of Marketing, University of New South Wales, Sydney 2052, New South Wales, Australia; Telephone: +61-2-9385-3510; Fax: +61-2-9663-1985; Email: m.uncles@unsw.edu.au

Sales Promotion Effectiveness: The Impact of Culture at an Ethnic-Group Level Abstract There is considerable interest and debate over the effectiveness of sales promotion. Previous studies have shown that sales promotions are more effective when they provide benefits that are congruent with those of the promoted product. This study explores and extends the congruency framework by analysing the impact of culture at an ethnic group level. The purpose is to investigate the popular assumption that cultural differences exist at this level and to see whether these differences have an impact on sales promotion effectiveness. A quasi-experimental design is used to test a series of hypotheses based on a sample of AngloAustralians and Chinese-Australians. It is found that despite the existence of cultural differences at an ethnic level, culture does not appear to have a significant impact on consumer responses to sales promotion. It is also found that the congruency effects between product and promotion type are weak and may be non-existent in some cases. Finally, the study also provides evidence that further validates a scale used for the measurement of culture. Keywords: Culture, Ethnicity, Sales Promotions, Congruency Theory 1. Introduction The widespread use of consumer sales promotions has sparked considerable interest and debate over their effectiveness. Critics argue that sales promotions are ineffective as they make consumers more promotion prone, resulting in market share losses in the long run (Ehrenberg, Hammond and Goodhardt, 1994; Totten and Block, 1987). However, other researchers have shown that sales promotions lead to real increases in sales and profits (Dhar and Hoch, 1996; Hoch, Dreze and Purk, 1994). This discrepancy suggests that there are conditions and factors that can influence the effectiveness of sales promotions. For instance, it has been shown that sales promotions are more effective when they provide benefits that are congruent with those of the promoted product (Chandon, Wansink and Laurent, 2000). This paper explores and extends the congruency framework by incorporating the impact of culture. Culture has an impact on many aspects of consumer behavior, from service expectations to consumer innovativeness. It also has been suggested that an understanding of culture can assist in making marketing decisions, such as whether to pursue standardised or localised strategies something that has been discussed recently in the context of retailing strategies (Mooij and Hofstede, 2002). However, many cross-cultural studies in consumer marketing have only examined the impact of culture across nations. It has been argued that cultural differences may also exist at an ethnic level, for instance among Caucasian-Americans, Hispanic-Americans, African-Americans, North African-French, Italian-Australians, etc. (Tan and McCullough, 1985; Donthu and Yoo 1998). This is reflected in the growth of interest in ethnic marketing by both researchers (Lee, Fairhurst and Dillard, 2002) and practitioners (Jarvis, 2002). However, there is limited empirical research focusing on culture at an ethnic-group level and its impact on consumer behaviour. The purpose of this study is to investigate cultural differences at this level and to gauge the impact on the congruency framework of sales promotion effectiveness. The study was undertaken in Australia. As in many nations, increasing use is being made of consumer sales promotions by a diverse ethnic mix of consumers (Lyons, 2002; Millett, 2002).

This study makes several important contributions to both marketing theory and practice. Firstly, although Chandon, Wansink and Laurent (2000) attempted a cross-national replication, their analysis did not specifically explore the impact of culture. Thus, by incorporating culture, this study redresses one of the limitations of the earlier study and helps extend the congruency framework of sales promotion effectiveness. Secondly, this study is one of the few pieces of research in marketing that empirically measures culture at an ethnicgroup level. It provides evidence to examine the popular assumption that cultural differences exist at this level. Thirdly, the study also contributes to theory development by providing further validation of a new scale for measuring culture in a consumer context, namely the CVSCALE (Yoo, Donthu and Lenartowicz, 2001). Finally, the study provides insights to marketing practitioners in the design of sales promotion strategies. It addresses the issue of whether to standardise or localise sales promotions between targeted ethnic markets. It should be noted that the focus of this study is on consumer sales promotions for packaged goods The paper is organised as follows. The next section provides a review of the sales promotions literature and considers the potential impact of culture. Hypotheses are presented in section three. Key measures and stimuli are discussed next. In sections five and six, the methodology and results of the study are described. A discussion of the findings is presented in section seven. The paper concludes by highlighting the limitations and opportunities for future research.. 2. Sales Promotion and the Potential Impact of Culture 2.1 Types of Sales Promotion The majority of past studies on the effectiveness of consumer sales promotion have focused on monetary sales promotions (Dickson and Sawyer, 1990; Dhar and Hoch, 1996; Hoch, Dreze and Purk, 1994). However, in practice, a range of both monetary and non-monetary sales promotions are used (Campbell and Diamond, 1990; Tellis 1998), and there are important differences between them. Monetary promotions (e.g., shelf-price discounts, coupons, rebates and price packs) tend to provide fairly immediate rewards to the consumer and they are transactional in character; non-monetary promotions (e.g., sweepstakes, free gifts and loyalty programs) tend to involve delayed rewards and are more relationship-based. In assessing the effectiveness of sales promotions, it is necessary to examine both types. 2.2 Benefits of Sales Promotion Sales promotions can offer many consumer benefits. Past studies have concentrated on monetary saving as the primary consumer benefit (Blattberg and Neslin, 1993). However, there is evidence to suggest consumers are motivated by several other benefits, including the desire for: savings, quality, convenience, value expression, exploration and entertainment. These benefits are further classified as either utilitarian or hedonic (Babin, Darden and Griffin, 1994; Hirschman and Holbrook, 1982). Utilitarian benefits are primarily functional and relatively tangible. They enable consumers to maximise their shopping utility, efficiency and economy. In general, the benefits of savings, quality and convenience can be classified as utilitarian benefits. By contrast, hedonic benefits are more experiential and relatively intangible. They can provide consumers with intrinsic stimulation, fun and pleasure. Consistent with this definition, the benefits of value expression, exploration and entertainment can be classified as hedonic benefits. 2.3 Promotion Types and Promotion Benefits

Based on the distinction between the types of sales promotions and promotion benefits, Chandon Wansink and Laurent (2000) showed that monetary promotions provide more utilitarian benefits whilst non-monetary promotions provide more hedonic benefits. These relationships are a matter of degree rather than absolutes; for example, coupon promotions (i.e., a monetary promotion) may still provide some hedonic benefits such as the enjoyment in redemption, although its main benefit of saving is utilitarian (Mittal, 1994). 2.4 Congruency Theory and Sales Promotion The basic principle of congruity states that changes in evaluation are always in the direction that increases congruity with the existing frame of reference (Osgood and Tannenbaum, 1955). In other words, people have a natural preference for consistent information. The principle has been examined in many marketing contexts, including studies of brand extensions and advertising appeals. Applying the congruity principle to sales promotions, it is expected that sales promotions will be more effective when they provide benefits that are compatible with the benefits sought from the promoted product. The relevance of this principle is evident from some past studies of sales promotions. For example, Roehm, Pullins and Roehm Jr (2002) showed that loyalty programs are more successful if they provide incentives that are compatible, rather than incompatible, with the brand. Likewise, Dowling and Uncles (1997) suggest the effectiveness of loyalty programs is enhanced if program benefits directly support the target products value proposition. Congruency effects for sales promotions were directly tested and confirmed by Chandon, Wansink and Laurent (2000), who showed that: (a) monetary promotions are more effective for utilitarian products as they provide more utilitarian benefits, which are compatible to those sought from utilitarian products; and (b) non-monetary promotions are more effective for hedonic products as they provide more hedonic benefits, which are compatible to those sought from hedonic products. For example, price cuts are more effective than free gifts for influencing brand choice of laundry detergent (i.e., a utilitarian product), whereas sweepstakes are more effective than price cuts for influencing brand choice of chocolates (i.e., a hedonic product). However, it is noted that there are other factors that may impact on the congruency effects, including the product life cycle, purchases situations and consumer demographics. Another possible factor is culture, which is the focus of this study. 2.5 Culture and Ethnic Groups Culture is difficult to define, but typically it is seen as a set of norms and beliefs that are shared amongst a group of people and that provide the guiding principles of ones life (Goodenough, 1971; Kroeber and Kluckholn, 1952; Schwartz and Bilsky, 1987 and 1990). Here, culture is defined as the way of life of people grouped by ethnicity, including shared norms and beliefs that can impact on behavior. This definition is appropriate for several reasons. Firstly, it implies that culture encompasses all the norms and beliefs of a society it is the total way of life in a society. As Triandis (1989) suggests, these societal norms and beliefs will ultimately have an impact upon the dispositions and behaviours of society members. Thus, the definition allows for the possibility of culture to have an impact on consumer behaviour. Secondly, the definition is flexible in allowing for different levels of culture. This is evident by the notion of society within the definition, which means culture is not necessarily restricted to a country basis. This is important given the focus of this study is not on national culture. Furthermore, it has been suggested that equating culture with nations can be inappropriate (Lenartowicz and Roth 1999; Usunier, 2000). Instead, culture can be

conceptualised at different levels and in a variety of contexts (Dawar and Parker 1994; Hofstede, 1991) For instance, culture defined by age or music, as in youth or jazz culture. In this study, culture is examined at the ethnic-group level within the domestic Australian context. Ethnic groups can be considered as subcultures within a country. They preserve the main characteristics of the national culture from which they originate but also develop their own unique norms and beliefs (Steenkamp, Hofstede and Wedel, 1999; Usunier 2000). Indeed, central to any ethnic group is a set of cultural values, attitudes and norms (Tan and McCullough, 1985). Each ethnic group constitutes a unique community because of common culture (Lee, Fairhurst and Dillard 2002). Thus, the study of culture by ethnicity within a domestic context is feasible and appropriate since each ethnic group will have its own unique set of cultural values. In fact, it has been suggested that intra-country variations of culture can be as large as the variation across countries (Au, 1999). This is particularly the case in the Australian context given its increasingly diverse ethnic mix (Millett, 2002). In effect, a single national culture to represent Australia is likely to be inappropriate (Bochner and Hesketh, 1994). 2.6 Culture, Ethnic Groups and Sales Promotion As Nakata and Sivakumar (2001) noted, the impact of culture has been well documented in many areas of marketing, including consumer responses to sales promotion. For example, Bridges, Florsheim and Claudette (1996) argued that there is a need for research directed at understanding culturally-driven responses of consumers to promotional activities. They showed that in the service context, cultural values affect the effectiveness of promotion strategies. This is supported by McCort and Malhotra (1993) who claim that as cultures differ in their value systems, evaluations of marketing communications will differ. However, most of these assessments have been conducted at a national level, whereas there is also a need for research to examine the effects of promotional activities on cultural groups within countries (Albaum and Peterson, 1984). There has been some consideration of this in previous studies. For example, it has been argued that various cultural sub-groups should react differently to different promotion strategies (Laroche, Pons and Turnel, 2002) and Green (1995) found that coupons are relatively less effective for African-Americans than Anglo-Americans. Nevertheless, evidence at an ethnic-group level remains limited. 2.8 Cultural Dimensions Given the potential relevance of culture, a basis is required for assessing its impact. Here use is made of the five cultural dimensions popularised by Hofstede (1991): power distance, uncertainty avoidance, individualism-collectivism, masculinity-femininity and the Confucian dynamism. Alternative dimensions have been suggested by other researchers (Clark, 1990), but Hofstedes dimensions are by far the most widely accepted (Sondergaard, 1994) and have been applied in many cross-cultural studies (Steenkamp, Hofstede and Wedel, 1999; Lynn, Zinkhan and Harris, 1993; Roth, 1995). Admittedly, there are several poignant criticisms of Hofstedes dimensions. Firstly, his original study is relatively old and may be outdated. Secondly, the dimensions were developed from work-related values and thus, they may not fully apply to a consumer context. However, despite these limitations, Hofstedes dimensions remain conceptually valid for explaining cultural differences. The appropriateness of these dimensions for this study is supported by the suggestion that there are specific relationships between (Hofstedes) cultural dimensions and the appropriate promotional policy (Kale and McIntyre, 1991). Indeed, one of the aims of this study is to verify whether there are any relationships between the cultural dimensions and consumer

responses to sales promotion. Furthermore, although developed for cross-country comparisons, Hosfstedes dimensions are believed to be capable of explaining intra-country variations (Au, 1999), including at an ethnic-group level. 3. Hypotheses In general, it is hypothesised that differences based on Hofstedes (1991) five cultural dimensions can lead to relative differences between ethnic groups in their preference for promotion types. In effect, within the congruency relationships established between product and promotion types, ethnic groups may differ in their relative choices of monetary and nonmonetary promotions. For example, whilst monetary promotions might be more effective for utilitarian products, the choice share of monetary promotions may be higher for one ethnic group than another due to cultural differences. In the following, the hypotheses are detailed based around the five cultural dimensions. It is should be kept in mind that the theoretical strength of the hypotheses is not equal across the five dimensions. For example, hypotheses regarding collectivism have a stronger theoretical basis than hypotheses regarding power distance. Also, as is in the nature of any testing of this kind, it is possible to conceive of alternative arguments. However, all five dimensions have been included to ensure the study is comprehensive. 3.1 Power Distance Power distance deals with the acceptability of social inequalities, such as in power, wealth and status (Nakata and Sivakumar 2001). In high power distance cultures, inequality is prevalent and accepted. Indeed, privileges and status symbols are both expected and desired (Hofstede, 1991). Consumers in such cultures are thus likely to be more responsive to sales promotions that contain differential treatment. These mainly involve non-monetary promotions, in which differential treatment may occur by purchase value (e.g., free gifts and reward programs) or by chance (e.g., sweepstakes). In contrast, cultures with lower power distance are less tolerant of inequalities and special privileges (Hofstede, 1991). Consumers in such a culture would have a relatively higher preference for sales promotions that offer equal rewards for everyone. These mainly involve monetary promotions, such as price discounts and coupons, as they are generally available with the same level of benefit offered to everyone. Hypotheses 1A : Monetary promotions are more effective for low power distance cultures relative to high distance power cultures. Hypotheses 1B : Non-monetary promotions are more effective for high power distance cultures relative to low distance power cultures. 3.2 Uncertainty Avoidance Uncertainty avoidance deals with the level of discomfort regarding future uncertainties (Nakata and Sivakumar 2001). Although not equivalent, it is closely related to risk aversion. In high uncertainty avoidance cultures, there is a tendency to prefer stable situations and avoid risk (Usunier, 2000). Thus, to the extent that uncertainty avoidance is related to risk aversion, such cultures would prefer promotions that offer more tangible and immediate rewards (e.g., price discounts). This is expected since such rewards are more certain and involve minimal amounts of risk. On the other hand, cultures with low uncertainty avoidance are more risk tolerant and see opportunities within future uncertainties (Nakata and Sivakumar 2001). In fact, they may even be considered as risk seeking given that cultures

with low uncertainty avoidance have been shown to exhibit higher levels of innovativeness (Steenkamp, Hofstede and Wedel, 1999). Thus, consumers of such a culture will be more accepting of promotions that offer relatively less tangible and long-term rewards (e.g., sweepstakes and loyalty programs). Hypotheses 2A : Monetary promotions are more effective for high uncertainty avoidance cultures relative to low uncertainty avoidance cultures. Hypotheses 2B : Non-monetary promotions are more effective for low uncertainty avoidance cultures relative to high uncertainty avoidance cultures. 3.3 Individualism/Collectivism Individualism refers to the degree of distance in social relationships (Nakata and Sivakumar 2001). It has been suggested that relationships play an important role in the search and choice processes of consumers (Doran, 1994). Thus, the extent of individualism may affect consumer choices between different types of promotions. Individualistic cultures have distant social relationships, in which personal goals are favoured over group needs (Watkins and Liu, 1996). Value is placed on self-interest and independence (Gurhan-Canli and Maheswaran, 2000), as well as pleasure (Triandis and Hui, 1990). In addition, individualistic cultures emphasise differentiation (Aaker and Maheswaran, 1997) and the ability to express ones uniqueness (Watkins and Liu, 1996). Given these characteristics, individualistic cultures might be more receptive to non-monetary promotions since the associated hedonic benefits are entertaining and more experiential. Furthermore, hedonic benefits can provide intrinsic value to individuals and provide an opportunity for self-expression. In contrast, less individualistic (or more collectivistic) cultures are characterised by close relationships and interdependence (Fletcher and Brown, 1999). There is strong emphasis on conforming to in-groups, which are typically close social groups such as family and friends (Hofstede, 1991). At the same time, entry and exit to other groups is difficult and rare (Watkins and Liu, 1996). Thus, collectivistic cultures can be expected to be less responsive to relationship building promotions (e.g., free gifts and reward programs) since they will be reluctant to forge a relationship with an out-group. Instead, collectivistic cultures may be more likely to respond to monetary promotions since the benefits provided are more common (e.g., conform to group norms) and are more readily shared amongst the in-group (e.g., savings and quality). Hypotheses 3A : Monetary promotions are more effective for collectivistic cultures relative to individualistic cultures. Hypotheses 3B : Non-monetary promotions are more effective for individualistic cultures relative to collectivistic cultures. 3.4 Masculinity/Femininity Masculinity refers to the tendency to strive for personal achievement and performance (Cutler, Erdem and Javalgi, 1997; Nakata and Sivakumar, 2001). In more masculine cultures, strong values are placed on materialistic success and assertiveness (Fletcher and Brown, 1999). It can be argued that consumers in masculine cultures are more likely to respond to monetary promotions, since the more tangible and transactional-based benefits can satisfy their need for personal and materialistic success. At the other end of the spectrum, less masculine (or more feminine) cultures emphasise the values of nurturing, caring for others

and the quality of life (Nakata and Sivakumar 2001). There is relatively less emphasis on personal and materialistic gains. Instead, people and relationships are important (Hosfstede, 1991) and group oriented harmony is preferred (Cutler, Erdem and Javalgi, 1997). Thus, feminine cultures are expected to be more responsive to non-monetary promotions since the benefits offered are more relationship focused. Hypotheses 4A : Monetary promotions are more effective for masculine cultures relative to feminine cultures. Hypotheses 4B : Non-monetary promotions are more effective for feminine cultures than masculine cultures. 3.5 Confucian Dynamism The final dimension of Confucian dynamism concerns time orientation and is bipolar. It has been suggested that the way consumers understand and allocate time may help explain differences in consumer behaviour across cultures (Brodowsky and Anderson, 2000). The higher or positive end is related to a future oriented perspective with values placed on persistence and loyalty (Fletcher and Brown, 1999). Consumers in such cultures are more willing to make short-term sacrifices or investments for long term gains (Nakata and Sivakumar 2001). This is supported by research studies which have shown that people with a future orientation have a preference for delayed rewards (Klineberg, 1968). In effect, consumers in cultures high on Confucian dynamism are expected to be more responsive to non-monetary promotions such as sweepstakes and loyalty programs, since many of the rewards are long term and loyalty-based (Foxman, Tansuhaj and Wong, 1988). In contrast, the lower or negative end is characterised by a past oriented perspective, with an emphasis on traditions (Fletcher and Brown, 1999). People in such cultures favour shortterm planning and more immediate financial gains (Nakata and Sivakumar, 2001). This is supported by the fact that people with a past orientation are less likely to save money for the future (Spears, Lin and Mowen, 2001). Thus, consumers of cultures low on Confucian dynamism are expected to react relatively poorly towards non-monetary promotions due to the delayed gratification involved (Foxman, Tansuhaj and Wong, 1988). Instead, they are expected to favour monetary promotions given the benefits are more immediate and transactional. Hypotheses 5A : Monetary promotions are more effective for cultures low on the Confucian dynamism relative to cultures high on the Confucian dynamism. Hypotheses 5B : Non-monetary promotions are more effective for cultures high on the Confucian dynamism relative to cultures low on the Confucian dynamism. The ten hypotheses associated with the five cultural dimensions are summarised in Figure 1. Each cultural dimension is considered one-by-one.

Figure 1: Summary of Hypotheses

Low power distance Cultural Dimensions (considered individually) High uncertainty avoidance Collectivistic Masculine Low Confucian

High power distance Low uncertainty avoidance Individualistic Feminine High Confucian

Sales Promotions and Benefits

Monetary Promotions (with utilitarian benefits)

Non-monetary Promotions (with hedonic benefits)

Congruency Effects

Utilitarian Products

Hedonic Products

4. Measurement To test the hypotheses, there are two pretests and one main experiment. However, it is first necessary to discuss the stimuli and measurement scales. This is summarized in Table 1 and detailed below. Table 1: Summary of Measures Item Sales promotion benefits Measures/Source 18-item benefit scale 3-item overall evaluation scale (Chandon, Wansink and Laurent, 2000) 4-item utilitarian index score (Batra and Ahtola, 1990) Secondary research (Burton, 2001) Secondary research 26-item CVSCALE (Yoo, Donthu and Lenartowicz, 2001) Brand choice (market shares) Area of Application pretest one

Product category stimuli Brand stimuli Promotion stimuli Culture Sales promotion effectiveness

pretest two main experiment pretest two main experiment pretest one main experiment main experiment main experiment

4.1 Sales Promotion Benefits Sales promotion benefits are defined and classified in this study according to the scale developed by Chandon, Wansink and Laurent (2000). The scale indicates six main benefits, which can be classified as either utilitarian or hedonic. Specifically, the benefits of savings, quality and convenience are classified as utilitarian, whilst the benefits of value expression, exploration and entertainment are hedonic. A direct replication of these classifications is appropriate as the scale has been shown to be valid and maintaining scale consistency can enhance the comparability of final results with the original research (Churchill, 1979). A pretest is conducted to confirm the classification of benefits in the Australian context. The measures for the pretest are the same 18-item agree/disagree scales used in the original study. 4.2 Product Category and Brand Stimuli Utilitarian and hedonic products are used as stimuli in the main experiment. In the past, researchers have used a variety of products to represent these two categories. For example, Hirschman and Holbrook (1982) suggested that washing machines and other durables are utilitarian products, whereas movies and fashion items are hedonic. The product categories used in this study are mostly derived from those used by Chandon, Wansink and Laurent (2000). Specifically, laundry detergent, AA batteries and film are selected as utilitarian products, whilst chocolates, ice-cream and biscuits 1 are selected as hedonic products. This degree of replication helps to ensure consistency. A pretest described later is used to verify the utilitarian and hedonic nature of the pre-selected product categories. In measuring the degree of utilitarianism and hedonism of a product, an adaptation of the scale by Batra and Ahtola (1990) is used. Specifically, products are classified as either utilitarian or hedonic based on a utilitarian index score. The index consists of 9-point semantic differential scales on two hedonic items of fun/not fun and pleasant/unpleasant, and two utilitarian items of useful/useless and wise/foolish. Within each of these product categories, brands of relevance to Australian consumers are selected. Furthermore, only brands with a high level of customer-based brand equity are used as the congruency effects are expected to be strongest for high equity brands 2 . High equity brands tend to be well-known national brands with high market shares. These were identified from pretests based on the brands with the highest market shares in Australia for each product category, as reported by Retail World (Burton, 2001). 4.3 Promotion Stimuli Examples of monetary and non-monetary promotions are used as stimuli for both the pretests and the main experiment. The stimuli for monetary promotions are shelf-price discounts and price packs 3 , and for non-monetary promotions they are sweepstakes and free gifts. These promotion techniques are appropriate for several reasons. Firstly, they correspond to those used in earlier research and thus, consistency can again be maintained (e.g., Dickson and Sawyer, 1990; Dhar and Hoch, 1996; Huff and Alden, 1998). The techniques are also appropriate as they are commonly used in Australia and are also common for the pre-selected product categories. In contrast, alternatives such as coupons are inappropriate since coupon
1

Biscuits did not form part of the earlier study but are included as an appropriate example of a hedonic product in Australia. 2 This decision was made because the focus of this study is the impact of culture and not the effect of brand equity per se. 3 Price packs, as defined by Tellis (1998), are monetary promotions that offers savings through multiple packs (e.g., two for the price of one) or enlarged packs (e.g., contains 50% more).

usage is relatively low in Australia. Loyalty programs are also unsuitable given that there are few loyalty programs for the product categories selected for this study. The nature of the preselected promotion techniques (i.e., monetary or non-monetary) is verified in a pretest. Specific examples of the four promotion techniques are used in the main experiment. They are drawn from currently offered promotions in the product categories to ensure realism. This involved the use of a combination of secondary data and judgement. Specifically, the examples are derived from reviews of weekly supermarket catalogues and from direct observations in supermarkets. Judgment is then applied to determine the most typical examples representing each promotion technique. Consideration is also given to the fact that monetary promotions will be preferred over non-monetary promotions of the same nominal value (Campbell and Diamond, 1990). This is due to the time value of money and the relatively higher effort required for non-monetary promotions (Soman, 1998). Thus, another criterion for identifying the examples of each promotion technique is that in general, the nominal value of non-monetary promotions will need to be greater than monetary promotions. 4.4 Culture Culture is measured using a personality-centered approach based on direct value inference (Lenartowicz and Roth, 1999). In particular, use is made of the CVSCALE proposed by Yoo, Donthu and Lenartowicz (2001). This is an adaptation of Hofstedes scale; it consists of 26items, measured by 5-point Likert scales, relating to Hofstedes five cultural dimensions (see Appendix A). It allows culture to be measured at the individual level and then aggregated to form groups at a chosen level for comparison. This is appropriate as it recognises that members of a society may not share the same cultural values (Au, 1999) and it also allows different ethnic groups within one country to be analysed. As suggested by Yoo, Donthu and Lenartowicz (2001), the CVSCALE is useful for analysing cultural values in a heterogeneous country and thus, the scale is particularly relevant for this study. Furthermore, the items of the scale have been adapted to suit the consumer context. This reduces the negative impact of using Hofstedes measures, which were based on work-related values. Finally, the reliability and validity of the CVSCALE have been shown to be robust in multiple studies across different samples, including validation studies involving both student and non-student respondents from across America, Korea, Brazil and Poland (Yoo, Donthu and Lenartowicz, 2001). The CVSCALE has also been applied in cross-cultural research (Yoo and Donthu, 2002). Thus, there is strong evidence to support the use of this scale. Its reliability and validity is further tested in this study. 4.5 Sales Promotion Effectiveness There are various ways to define and measure the effectiveness of sales promotions. The measures typically used are short-term measures, as sales promotions are mostly used to produce short-term effects. This includes measuring the effectiveness of sales promotions by sales volume (Dhar and Hoch, 1996), profits (Hoch, Dreze and Purk, 1994) and consumer usage of the promotion (Babaku, Tat and Cunningham, 1988). However, it has been noted that a brands sales volume is by far the best measure of the performance of a sales promotion (Totten and Block, 1987). For the purposes of this study, the effectiveness of sales promotions is measured by market share, which is a proxy for sales volume. Market shares are calculated based on choices for promotion types, made under the conditions of the quasi-experiment. The effectiveness of sales promotions is then determined by a comparison between the choice shares of promotion types across different products. This is consistent with Chandon, Wansink and Laurent (2000).

10

5. Methodology and Pretests The sample is described, followed by a summary of the pretests. Then the main experiment is introduced. 5.1 Samples The two ethnic groups of Anglo-Australians and Chinese-Australians are selected for investigation. The source countries of these groups differ markedly in terms of Hofstedes (1991) cultural dimensions. Relatively, China is seen as: high power distance, low on uncertainty avoidance, collectivistic, feminine and high on the Confucian dynamism, whereas Australia is: low power distance, high on uncertainty avoidance, individualistic, masculine and low on the Confucian dynamism. Whilst these differences are found at a very broad national level, it is expected that the main differences will also be evident at an ethnic level and hence facilitate the testing of the hypotheses. Previous studies of Anglo-Australians and Chinese-Australians, and NorthAmericans and Canadian-Chinese, support this belief (Bochner and Hesketh, 1994; Doran, 1994). However, rather than merely rely on previous studies, the differences between the two ethnic groups are derived empirically in the present study (see section 6.2). The ethnic samples used for both pretests and the main experiment are drawn from undergraduate students. Although there are criticisms over the validity and generalisability of student samples (Peterson, 2001), the main purpose of this study is theory testing and not effects application. Thus, the use of a homogeneous sample such as students is acceptable and indeed appropriate, as it reduces variability and the impact of irrelevant factors (Calder, Phillips and Tybout, 1981). Furthermore, the use of students is only inappropriate if there are theoretical reasons to suggest that they will systematically respond differently than others to sales promotions. However, there does not appear to be any major theoretical evidence for this, and students are consumers of all the products studied here. The samples are controlled for non-cultural confounding factors. As Foxman, Tansuhaj and Wong (1988) pointed out, both macroeconomic and sociodemographic factors can affect consumers of different cultures in their responses to sales promotions. Macroeconomic factors, such as the level of national economic activity, are effectively controlled by examining only one country and thus, these factors can be treated as constants. In regard to sociodemographic factors, common characteristics considered in cross-cultural studies on sales promotion include age, gender, income and the level of education. These have either been treated as covariates (Steenkamp, Hofstede and Wedel, 1999) or controlled via matched sampling (Bond, 1988; Green, 1995). However, it has also been argued that consumer demographics do not explain any differential effectiveness of a promotion (Totten and Block, 1987). Given these varied findings, in this study a mixed approach to the treatment of confounding sociodemographic factors is adopted. Firstly, the level of education is matched. The samples are restricted to undergraduate students, but of any discipline. This is designed to ensure a common level of education and, at the same time, reduce any bias from knowledge of a particular field. Secondly, the factors of gender, age and income are treated as covariates. Although gender and age were not found to be a significant factor by Chandon, Wansink and Laurent (2000), they remain important to examine as gender and age differences in consumer behaviour are possible, particularly across different cultures (Usunier, 2000). Income is also

11

considered a covariate although students tend to be relatively homogeneous in this respect, and any potential impact is expected to be minimal (Durvalsula et al, 1993). In terms of recruitment, a self-identification process is used to determine the ethnicity of respondents. This involves asking respondents to self-identify their ethnicity based on prompts such as Anglo-Australians, including Anglo-Saxons and Anglo-Celtics4 . Respondents that identify themselves as neither Anglo-Australian nor Chinese-Australian are excluded from analysis. Self-identification is believed to be more relevant for selecting subcultures within a country than other popular measures, such as the country of citizenship (Bochner and Hesketh, 1994). Self-identification represents a persons internal beliefs and hence is said to reflect a persons cultural reality (Hirschman,1981). However, despite the validity of self-identification, it may be confounded with the effect of acculturation (i.e., the extent of assimilation of a new culture by an ethnic minority). It has been suggested that the level of acculturation can largely determine an individuals commitment to the cultural norms of his or her ethnic group (Hirschman, 1981). Thus, acculturation is another confounding factor that is controlled. Based on an adaptation of the acculturation scale suggested by Quester, Karunaratna and Chong (2001), acculturation in this study is analysed using a persons country of birth and the time they have spent living in Australia. 5.2 Pretests Two pretests were completed (see Appendix B for details). The first pretest was designed to confirm the nature of the promotion techniques, and verify the relationships between monetary and non-monetary promotions with utilitarian and hedonic benefits respectively. Short self-administered questionnaires were completed by a random sample of 15 AngloAustralian and 15 Chinese-Australian students. Results confirmed that shelf-price discounts and price packs can be classified as monetary promotions, with sweepstakes and free gifts as non-monetary promotions. They also showed that positive relationships exist between monetary and non-monetary promotions with utilitarian and hedonic benefits respectively. These results largely replicated the findings of Chandon, Wansink and Laurent (2000). The second pretest was designed to verify the utilitarian and hedonic nature of the preselected product categories that are used for the main experiment. It also sought to identify specific brands that are representative of high equity brands in each of the product categories. Using a new random sample of 15 Anglo-Australians and 15 Chinese-Australian, batteries and film were shown to be utilitarian products and chocolates and ice-cream as hedonic products. Brands with which respondents had most experience, in terms of frequency of purchase, were Energiser (batteries), Kodak (film), M&Ms (chocolates) and Peters (icecream). 5.3 Main Experiment The main experiment consists of a self-administered questionnaire, which is designed to test the validity of the CVSCALE and test the five pairs of hypotheses listed in section 3. The questionnaire was pilot tested. In the main experiment, two versions were used to test for ordering effects. Questionnaires were only distributed in undergraduate classes, thereby eliminating the need to screen for education level. After a brief introduction, without disclosing the purpose of the questionnaire, participation was sought on a voluntary basis.
4

However, this is merely a screening exercise. Respondents do not self-select into treatment conditions.

12

Screening for ethnicity did not occur at this stage. Instead all students were invited to participate. The questionnaire took about 10-15 minutes to complete and almost all students choose to participate across all classes. No incentives were offered it was felt that in a study about sales promotion the offer of incentives might bias results. Respondents were randomly assigned one of the two versions of the questionnaire. For both versions, respondents were asked to: i) choose between options A and B for two utilitarian and two hedonic products (Option A was a brand associated with a monetary promotion whilst option B was the same brand but associated with a non-monetary promotion); ii) provide past purchase information for the products and brands involved; iii) complete the CVSCALE items; and iv) complete demographic questions including gender, age, income, ethnicity and acculturation. It should be noted that Chandon, Wansink and Laurent (2000) differed slightly in their research design as they asked respondents to evaluate two different brands within each product category. However, only one brand is used for this study so as to strengthen the promotion manipulations; by asking respondents to choose between two types of promotions for the same brand, any branding effects are eliminated and the focus is on pure-promotional effects. Price was omitted from the earlier study, whereas here it is included to give added realism to the scenarios, but it was held constant and therefore not manipulated. 6. Analysis In this section, a descriptive analysis is first presented. This is followed by an analysis of the CVSCALE and then the presentation of the main results. 6.2 Descriptive Analysis A total of 815 questionnaires were completed, 250 of which were for either AngloAustralians or Chinese-Australians (with an equal split between the two groups). Almost half of the sample was born overseas with Hong Kong (16%), China (8%) and Indonesia (5%) being the most common countries of birth. However, on average respondents have lived in Australia for 15 years (an average of 19 years among Anglo-Australians and 11 years among Chinese-Australians). Demographically the two groups are matched, although AngloAustralians are somewhat older on average and have a slightly higher level of income. With regard to past purchasing behaviour, most respondents have bought the four product categories examined in the past 12 months (batteries-80%, film-81%, chocolate-98% and icecream 95%). Past purchases were also relatively high at the brand level (Energiser-64%, Kodak-77%, M&Ms-78% and Peters-58%). Whilst there are some differences between the two ethnic groups (ice-cream, M&Ms and Peters; p< 0.05), the percentages of past purchases remain relatively high across all products and brands for each ethnic group. This suggests that respondents have sufficient knowledge and purchase experience of the products and brands involved, and it is reasonable to assume they are able to make informed evaluations of the different promotional options presented. 6.2 CVSCALE Analysis Responses to the CVSCALE are used to determine the relative cultural values of both ethnic groups on the five cultural dimensions. However, first the reliability and validity of the CVSCALE is tested. For the whole sample, the reliability alpha of the cultural dimensions ranged from 0.60 to 0.69 (Table 2). Although these results are modest, they are comparable to those reported by Yoo, Donthu and Lenartowicz (2001) and they all satisfy the reliability

13

threshold of 0.6 that is commonly accepted for new scales (Hair et al, 1998). Furthermore, no significant differences were found in the responses between the two versions of the questionnaire. Thus, there appear to be no ordering effects. It should be that reliability levels varied slightly between the ethnic groups. However, the variations are similar to those reported by Yoo, Donthu and Lenartowicz (2001) and in only one case did the reliability alpha fall below 0.6 (0.54 for masculinity among AngloAustralians) (Table 2). Table 2: Reliability Analysis Results Dimension Power Distance Uncertainty Avoidance Collectivism Masculinity Confucian Dynamism Whole Sample 0.65 0.64 0.67 0.60 0.69 AngloAustralians 0.69 0.67 0.61 0.54 0.68 ChineseAustralians 0.62 0.61 0.70 0.67 0.69 Yoo, Donthu and Lenartowicz 0.61 0.71 0.76 0.67 0.69

After reliability testing, factor analysis was used to ascertain the validity of the items (Table 3). Under the specification of 5 factors, the results of exploratory factor analysis provide preliminary support for the CVSCALEs validity. With one exception (D3-personal steadiness and stability), all the items loaded highly on the appropriate factors and no item loaded on more than one factor. Overall, therefore, the results support the independence of the constructs. Furthermore, the five factors explained 45% of the total variance, which exactly matches the figure reported by Yoo, Donthu and Lenartowicz (2001). Table 3: Exploratory Factor Analysis Results CVSCALE Item No. P2 P1 P4 P5 P3 U3 U5 U2 D3 U4 U1 C3 C4 C6 C2 C5 C1 M2 M1 Power Distance 0.69 0.66 0.60 0.58 0.49 0.04 -0.05 0.12 -0.08 -0.01 0.12 0.01 0.21 0.11 -0.20 0.18 0.07 -0.06 0.37 Uncertainty Avoidance 0.12 0.09 -0.10 0.18 -0.26 0.78 0.64 0.53 0.45 0.45 0.44 0.03 0.03 0.00 0.19 0.20 -0.02 0.09 -0.14 Collectivism Masculinity 0.15 -0.04 0.13 0.06 -0.02 0.05 0.06 0.11 -0.04 0.12 -0.01 0.67 0.67 0.67 0.54 0.51 0.50 0.01 0.17 -0.08 0.11 0.13 0.02 0.05 0.09 0.17 -0.19 -0.09 0.23 0.17 0.16 -0.05 0.27 -0.09 0.25 0.02 0.76 0.66 Confucian Dynamism -0.08 -0.16 0.07 -0.02 0.11 0.03 0.17 0.26 0.44 0.09 0.28 0.19 0.21 -0.07 -0.14 -0.17 0.16 0.15 0.15

14

M3 D4 D5 D6 D1 D2

0.11 0.07 0.08 0.15 -0.05 0.22 Extraction Method Rotation Method Kaiser-Meyer-Olkin Measure of Sampling Adequacy Bartlett's Test of Sphericity

0.10 -0.05 0.09 -0.08 0.02 -0.20

0.15 0.29 0.00 0.27 0.28 -0.23

0.64 -0.01 0.07 0.70 0.03 0.68 -0.10 0.67 0.09 0.58 -0.03 0.35 Principal Components Varimax 0.731 0.000

Confirmatory factor analysis was then employed to validate the scale in regard to specific constructs. The measurement model is based on the same specifications as Yoo, Donthu and Lenartowicz (2001), with 5 factors and 26 items, where each item loaded on only one factor and the factors are uncorrelated. Using AMOS 4.0, the key results of the standardised solution are shown in Table 4. The overall fit of the measurement model was excellent: 2 (d.f. = 299) = 543.32; root mean square error of approximation (RMSEA) = 0.06; normed fit index (NFI) = 0.97; comparative fit index (CFI) = 0.99; and incremental fit index (IFI) = 0.99. These results provide strong confirmatory support for the CVSCALE and its use in studying the hypothesised constructs. With regard to composite reliability, all the estimates were above the recommended level of 0.70 (Hair et al, 1998), ranging from 0.79 to 0.85 (Table 4). These results are evidence of the scales convergent validity. In addition, whilst the average variance extracted for each dimension was only moderate at 0.50, they do satisfy the minimum acceptable level (Hair et al, 1998). Thus, the results provide support for the independence of the dimensions (Fornell and Larker 1981). Table 4: Confirmatory Factor Analysis Results CVSCALE Item No. P1 P2 P3 P4 P5 U1 U2 U3 U4 U5 C1 C2 C3 C4 C5 C6 Power Distance 0.61 0.53 0.41 0.52 0.47 Standardised Factor Loadings Uncertainty Collectivism Masculinity Avoidance Confucian Dynamism

0.46 0.46 0.75 0.37 0.62 0.39 0.29 0.66 0.61 0.48 0.59

15

M1 M2 M3 M4 D1 D2 D3 D4 D5 D6 Composite Reliability Variance Extracted

0.61 0.57 0.60 0.36 0.58 0.23 0.49 0.73 0.48 0.69 0.85 0.50

0.83 0.50 2 543.32

0.82 0.50

0.85 0.50

0.79 0.50

Model

d.f. 299

RMSEA 0.06

NFI 0.97

CFI 0.99

IFI 0.99

Having confirmed the reliability and validity of the CVSCALE, responses to the scale are then aggregated for analysis. For the whole sample, an average score for each cultural dimension is computed for both ethnic groups. The score is calculated as the average of the individual items of each cultural dimension answered by the respondents of each ethnic group 5 . This approach reflects the flexibility of the CVSCALE in that it allows culture to be measured at the individual level but analysed at an appropriate aggregate level. Thus, individual respondents may differ from the average of their group but will remain appropriate for analysis. The average scores are then compared to classify the relative cultural values of the two ethnic groups on each dimension (Table 5). Table 5: Average Cultural Scores Power Distance 4.05 Low 3.88 High 2.49 0.01 Uncertainty Avoidance 2.16 Low 2.07 High 1.40 0.16 Collectivism 2.99 Individualistic 2.78 Collectivistic 2.96 0.00 Masculinity 3.28 Feminine 3.03 Masculine 3.16 0.00 Confucian Dynamism 2.11 Low 1.91 High 3.22 0.00

AngloAustralians ChineseAustralians T-value Sig. Pvalue

Although the absolute difference appears small, based on conventional statistical standards there are significant differences between the two ethnic groups on all of the cultural dimensions (p < 0.05), except for uncertainty avoidance. Using the relative averages, AngloAustralians can be classified as relatively low power distance, low on uncertainty avoidance, individualistic, feminine and low on Confucian dynamism, and vice-versa for ChineseAustralians. The classifications largely conform to Hofstedes (1991) results. The exceptions are the contrary results on uncertainty avoidance and masculinity. However, some
5

For example, if there are 5 items for the power distance dimension and 25 respondents for each ethnic group. The average for each ethnic group is equal to: (25 x 5 items scores) / 25

16

inconsistency is acceptable given the distinctiveness of the CVSCALE and the limitations of Hofstedes empirical data. Indeed, as explained in section 4, the purpose of using the CVSCALE is to provide a direct measure of culture and to avoid the limitations of inferring this from past studies such as Hofstedes. It should also be noted that despite the slight inconsistency, the hypotheses can still be tested. This is because the hypotheses involve a comparison of cultures defined only by cultural dimensions. There are no restrictions regarding how Anglo-Australians or ChineseAustralians should score on each dimension. For example, the hypotheses involving masculinity can still be tested by using Anglo-Australians as the feminine culture and Chinese-Australians as the masculine culture. 6.3 Analysis Procedures In order to examine each hypothesis, the results of the experiment are analyzed using two main procedures. Firstly, logistic regression is used to test for the congruency relationships between product and promotion types. The dependent variable is the choice between promotion type (monetary or non-monetary) and the independent variables are product type (i.e., utilitarian or hedonic) and the covariates of gender, age and income. Secondly, choice shares of promotion types are analysed to identify any differences in the choices between ethnic groups. Analysis is undertaken at an ethnic-group level and an individual level, and across different acculturation groupings 6.4 Main Results at an Ethnic Level In testing the hypotheses, the data were analysed at an ethnic level. The ethnic groups are already classified on each cultural dimension as shown in Table 5. For the purposes of analysis, the upper median splits within each ethnic group on each cultural dimension are used. This results in greater variance between the two groups on the dimensions of interest. Logistic regression analysis is performed on each ethnic group for each dimension. Thus, a total of 10 regressions were conducted (Table 6). Results show that the regression models generally have a poor fit since the reduction in the 2 log likelihood values and the R2 values are relatively low. However, the omnibus test of model coefficients indicates that coefficients were significant for 5 of the models (p < 0.05). Within the significant models, product type was consistently shown to have a significant and negative relationship with promotion type: high power distance (B = -1.60, p = 0.00), high uncertainty avoidance (B = -1.33, p = 0.00), collectivist (B = -0.96, p = 0.00), masculine (B = -1.38, p = 0.00) and high Confucian Dynamism (B = -1.06, p = 0.00). These results indicate that for each significant dimension, hedonic products are associated with the choice of monetary promotions and utilitarian products are associated with the choice of non-monetary promotions. The covariates of gender, age and income were generally found to be insignificant. The only exception is that higher income was found to be associated with the choice of non-monetary promotions under the collectivist dimension (B = 1.29, p = 0.04). Table 6: Logistic Regression Results at an Ethnic Level
Model Summary -2 Log R2 Omnibus Test Likelihood Value of Model Coefficients Independent Variables Product Gender Age Income Type

17

Low PD223a Anglo (226)b High PD218 Chinese (242) Low UA238 Anglo (242) High UA241 Chinese (261) Individualist261 Anglo (266) Collectivist244 Chinese (258) Feminine241 Anglo (242) Masculine216 Chinese (236) Low CD255 Anglo (261) High CD236 Chinese (250) a Model 2 Log Likelihood c Nagelkerke R2

0.03c 0.15 0.03 0.12 0.03 0.09 0.01 0.12 0.03 0.09
b d

0.43 0.00 0.42 0.00 0.28 0.00 0.95 0.00 0.25 0.00

-0.23 (0.50)d -1.60 (0.00) -0.48 (0.15) -1.33 (0.00) -0.28 (0.36) -0.96 (0.00) -0.16 (0.63) -1.38 (0.00) -0.39 (0.22) -1.06 (0.00)

0.64 (0.08) 0.39 (0.27) 0.55 (0.88) -0.15 (0.65) 0.30 (0.36) 0.30 (0.35) 0.16 (0.67) 0.41 (0.25) 0.59 (0.07) 0.35 (0.30)

0.15 (0.69) -0.17 (0.66) 0.09 (0.79) -0.16 (0.64) -0.22 (0.50) -0.07 (0.85) -0.05 (0.88) -0.15 (0.70) -0.07 (0.85) -0.26 (0.46)

-0.31 (0.46) 0.58 (0.18) 0.47 (0.18) 0.64 (0.10) 0.61 (0.07) 1.29 (0.04) 0.14 (0.71) 0.78 (0.09) 0.23 (0.50) 0.61 (0.10)

Initial 2 Log Likelihood Significant value

The choice share results for each ethnic group on each dimension are shown in Table 7. The results are reflective of the regression findings, in that hedonic products have a relatively higher choice share of monetary promotions than utilitarian products. Another key result is that for each ethnic split, monetary promotions are preferred over non-monetary promotions across all products and for each product type.

18

Table 7: Choice Shares for Monetary Promotions at an Ethnic Level All Products Power Distance Low-Anglo High-Chinese (sig. p-value) Uncertainty Avoidance Low-Anglo High-Chinese (sig. p-value) Collectivism Individualist-Anglo Collectivist-Chinese (sig. p-value) Masculinity Feminine-Anglo Masculine-Chinese (sig. p-value) Confucian Dynamism Low-Anglo High-Chinese (sig. p-value)
a

Utilitarian Products 81% 70% (0.03) 76% 66% (0.10) 75% 71% (0.52) 80% 73% (0.16) 75% 72% (0.52)

Hedonic Products 85% 92% (0.07) 82% 89% (0.33) 80% 86% (0.16) 82% 91% (0.04) 82% 88% (0.15)

83%a 81% (0.56) 79% 78% (0.44) 78% 79% (0.75) 81% 82% (0.82) 78% 80% (0.67)

Choice share for non-monetary promotions is the complement to 100%

The choice share results also provide a basis to evaluate the hypotheses. As is evident in Table 7, there was no significant difference in the choice shares between ethnic groups across all products. Within product types, differences were found in only 2 out of the possible 10 cases. Firstly, in the case of utilitarian products, low power distance Anglo-Australians were found to have a higher preference for monetary promotions than high power distance Chinese-Australians (81% vs 70%; p < 0.05). This is in line with the prediction of hypotheses H1A and H1B. Secondly, in the case of hedonic products, feminine Anglo-Australians were found to have a lower preference for monetary promotions than masculine ChineseAustralians (82% vs 91%; p < 0.05). This is consistent with hypotheses H4A and H4B. However, these were the only two instances where differences were found. It is evident that, in general, there was no difference in the choice shares between the two ethnic groups across all products and product types, despite differences in cultural values. Thus, there is insufficient evidence to support the hypotheses of this study. The results were confirmed when the same analyses were performed with a quartile-split sample. Although there was greater variance in the cultural values between ethnic groups, no significant difference in choice shares was observed for any of the cases. 6.5 Main Results at an Individual Level In order to provide further understanding, the data was also analysed at an individual level. Specifically, median splits were conducted on each dimension based on the scores of all individuals, regardless of their ethnic background.

19

The two groups on each dimension were analysed using the same logistic regression model as specified earlier. From Table 8 it is evident that the results generally reflect those found at the ethnic level. Firstly, the reduction in the 2 log likelihood values and the R2 values for each regression are again relatively low, suggesting a poor fit for all the models. However, with the exception of the collectivist dimension, model coefficients were found to be significant for the same dimensions identified at the ethnic level (p< 0.05). Similarly, product type was consistently shown to have a significant and negative relationship with promotion type: high power distance (B= -0.85, p= 0.00), high uncertainty avoidance (B= -0.81, p= 0.00), masculine (B= -0.89, p= 0.00) and high Confucian Dynamism (B= -0.68, p= 0.00). These results confirm the findings at the ethnic level that hedonic products are associated with the choice for monetary promotions. In regard to the covariates, they were generally again found to be insignificant. However, there are some results that differed from the ethnic level. Firstly, the model coefficients for low uncertainty avoidance, individualist and low Confucian Dynamism were also found to be significant (p< 0.05). Product type was again negatively related to promotion type. In addition, under the low Confucian Dynamism dimension female responses were positively related with the choice of non-monetary promotions (B= 0.47, p= 0.05).

20

Table 8: Logistic Regression Results at an Individual Level

Model Summary -2 Log R2 Omnibus Test Likelihood Value of Model Coefficients Low Power distance High Power distance Low Uncertainty Avoidance High Uncertainty Avoidance Individualist Independent Variables Product Gender Age Income Type

474a (483)b 509 (524) 479 (489) 505 (519) 491 (503) 497 (506) 484 (492)

0.03c 0.05

0.06 0.01

-0.54 (0.02)d -0.85 (0.00) -0.58 (0.01) -0.81 (0.00) -0.74 (0.00) -0.65 (0.00) -0.50 (0.03)

0.48 (0.06) 0.03 (0.91) 0.40 (0.10) 0.09 (0.69) 0.33 (0.18) 0.18 (0.43) 0.41 (0.14)

0.17 (0.50) -0.15 (0.53) 0.24 (0.33) -0.23 (0.32) 0.02 (0.93) -0.03 (0.88) 0.16 (0.49)

-0.02 (0.93) 0.03 (0.89) -0.09 (0.71) 0.15 (0.52) 0.11 (0.65) -0.04 (0.88) 0.15 (0.54)

0.03 0.04 0.04

0.05 0.01 0.01

0.03 0.03

0.06 0.10

Collectivist

Feminine

Masculine

500 (517)

0.05 0.04 0.03

b d

0.00 0.01 0.05

-0.89 (0.00) -0.71 (0.00) -0.68 (0.00)

0.10 (0.67) 0.47 (0.05) 0.07 (0.75)

-0.19 (0.45) 0.12 (0.63) -0.11 (0.65)

-0.09 (0.72) 0.08 (0.74) 0.10 (0.66)

Low 497 Confucian (506) Dynamism High 494 Confucian (503) Dynamism a Model 2 Log Likelihood c Nagelkerke R2

Initial 2 Log Likelihood Significant value

Choice share results for each dimension at an individual level are shown in Table 9. It is again evident that hedonic products are associated with a higher choice share of monetary promotions. The preference for monetary promotions also dominates across all products and for each product type. In addition, across all products and for each product type, no significant difference was found in the choice shares between the groups on each dimension. In fact, there are many cases with virtually no difference in the choice shares between the groups.

21

Table 9: Choice Shares for Monetary Promotions at an Individual Level All Products Utilitarian Products Hedonic Products

Power Distance Low 81%a 77% 85% High 78% 71% 85% (sig. p-value) (0.24) (0.13) (1.00) Uncertainty Avoidance Low 81% 76% 85% High 79% 72% 85% (sig. p-value) (0.39) (0.26) (1.00) Collectivism Individualist 80% 74% 86% Collectivist 80% 74% 85% (sig. p-value) (0.94) (0.92) (0.80) Masculinity Feminine 81% 77% 84% Masculine 79% 72% 86% (sig. p-value) (0.48) (0.19) (0.62) Confucian Dynamism Low 80% 74% 85% High 80% 74% 85% (sig. p-value) (0.94) (0.92) (1.00) a Choice share for non-monetary promotions is the complement to 100% 6.6 Acculturation Analysis The effects of acculturation are explored by dividing the sample of Chinese-Australian respondents using a median split based on the number of years that respondents have lived in Australia and, in a separate analysis, based on whether the respondent was born in Australia or overseas. These two splits are seen as different ways to examine the same underlying dimension of acculturation (Quester, Karunaratna and Chong, 2001). The four groups were analysed using logistic regression (Table 10). The results are fairly consistent across all groups. Firstly, all the models had relatively poor fit, as shown by the low reductions in the 2 log likelihood value and the low R2 values. Secondly, with the exception of the Australia born group, the model coefficients were found to be significant for all groups (p< 0.05) and the only significant variable was product type, which had a negative relationship with promotion type.

22

Table 10: Logistic Regression Results using Acculturation Splits

Model Summary -2 Log R2 Omnibus Test Likelihood Value of Model Coefficients Independent Variables Product Gender Age Income Type

Years In AustraliaLow Years In AustraliaHigh Overseas Born Australia Born

a c

219a (230)b 236 (258) 342 (368) 113 (121)

0.07c

0.03

-1.01 (0.01)d -1.53 (0.00) -1.31 (0.00) -1.20 (0.01)

0.44 (0.21) 0.05 (0.89) 0.39 (0.17) -0.44 (0.45)

-0.02 (0.92) 0.10 (0.79) 0.03 (0.92) -0.47 (0.43)

0.44 (0.32) 0.35 (0.33) 0.35 (0.28) 0.47 (0.36)

0.13

0.00

0.10

0.00

0.10
b d

0.10

Model 2 Log Likelihood Nagelkerke R2

Initial 2 Log Likelihood Significant value

The choice share results for the four groups are shown in Table 11. The results are consistent with those reported for Chinese-Australians in the earlier analyses (Tables 7 and 9). Firstly, the choice for monetary promotions was found to dominate non-monetary promotions. Secondly, hedonic products were found to be associated with monetary promotions. The results were consistent across all groups, and no significant difference in choice shares was observed. Thus, the results suggest that acculturation does not have an impact on the findings of this study. Table 13: Choice Shares for Monetary Promotions using Acculturation Splits All Products Years Lived in Australia Low High (sig. p-value) Country of Birth Overseas Australia (sig. p-value)
a

Utilitarian Products 76% 67% (0.12) 72% 69% (0.61)

Hedonic Products 90% 90% (0.83) 90% 88% (0.55)

83%a 79% (0.26) 82% 78% (0.46)

Choice share for non-monetary promotions is the complement to 100%

23

7. Summary and Discussion The key findings and contributions of the study can be summarized in three main areas: i) culture and ethnicity, ii) culture and sales promotion, iii) culture and the congruency framework. 7.1 Culture and Ethnicity Clear cultural differences are found at an ethnic group level. With the exception of uncertainty avoidance, the mean scores between Anglo-Australians and Chinese-Australians are significantly different from each other across all cultural dimensions. This is an important finding as most cross-cultural studies only assume or suggest that cultural differences exist at an ethnic level, rather than measure and demonstrate these differences (Doran, 1996; Laroche, Pons and Turnel, 2002). By contrast, the current study provides empirical evidence that confirms the popular assumption. It also further validates the CVSCALE established by Yoo, Donthu and Lenartowicz (2001). The flexibility of the CVSCALE is demonstrated, in that culture can be analyzed at both the ethnic and individual level. Thus, the study provides further evidence for the validity and usefulness of this scale. 7.2 Culture and Sales Promotion Another key contribution of the study is that despite cultural differences between ethnic groups, there is no significant difference in their preferences for sales promotion types. With only 2 exceptions, this result is found to be consistent at an ethnic-group level across all products and for each product type. The absence of cultural effects is also evident at an individual level. The implication of this finding is twofold. Firstly, although cultural differences may exist, these do not appear to affect consumer responses to sales promotion at an ethnic level. This suggests that managers can use standardized sales promotions when targeting different ethnic groups and avoid the use of more costly differentiated strategies. Secondly, the finding highlights the fact that cultural distinctions may be more relevant in some areas of marketing than in others. For example, the distinction between collectivism and individualism was found to account for differences in consumer complaining behavior (Watkins and Liu, 1996) but not in advertising appeals (Cutler, Erdem and Javalgi, 1997). Thus, it would be a mistake to assume that cultural differences will affect all areas of marketing. Hence, the relevance of ethnic marketing as suggested by researchers and practitioners (Jarvis, 2002; Lee, Fairhurst and Dillard, 2002; Quester, Karunaratna and Chong, 2001) needs to be considered within the specific context in which it is applied. 7.3 Culture and the Congruency Framework There are mixed findings in regard to the congruency framework of sales promotion effectiveness. Firstly, the preference for monetary promotions was found to dominate over the preference for non-monetary promotions across all product types. Furthermore, with only a few exceptions, the covariates of gender, age and income were all insignificant in accounting for the choice of promotions. All of these results are consistent with those reported by Chandon Wansink and Laurent (2000). They were also evident across all cultural groups at all levels of analysis and thus, the impact of culture on these results appears to be minimal. However, the interesting finding is that the direction of congruency effects between product and promotion types was opposite to that described by Chandon Wansink and Laurent

24

(2000). In the current study, it was generally found that hedonic products were associated with the choice of monetary promotions whilst utilitarian products were associated with nonmonetary promotions. The contrasting relationships may be explained by the fact that the congruency effects observed by the earlier researchers were only marginal (B = 0.32 for the US sample and B = 0.022 for the French sample). Thus, other factors could have affected the direction of the relationships. For example, the use of monetary incentives for recruiting respondents may have biased their results (whereas incentives were not used in the present study). Another possible explanation is that non-monetary promotions are preferred for utilitarian products because they provide consumers with the experiential benefits that are not provided by the product itself. This is supported by the successful use of non-monetary promotions for many utilitarian products, such as the loyalty program for Unilevers Omo laundry detergent and online competitions for Kelloggs Coco Pops. On the other hand, monetary promotions may be preferred for hedonic products because they can reduce the guilt associated with hedonic consumption. This is hinted at by the fact that promised donations to charity are more effective for promoting luxuries (i.e., hedonic products) than for necessities (Strahilevitz and Myers, 1998). 8. Limitations and Further Research There are several limitations, relating to the focus of the study and the methodology used. Some of these highlight useful directions for future research. One issue is that the cultural dimensions are examined separately. There is no examination of any correlation effects between the dimensions and no assessment of the relative importance of each dimension. Yet, ethnic groups are not expected to conform to any single cultural dimension as they involve a whole set of cultural values (Tan and McCullough, 1985). However, the focus on single cultural dimensions provides a clear conceptual distinction that can facilitate analysis and assist in the interpretation of results. Furthermore, the separate analysis of dimensions is consistent with past studies (Brodowsky and Anderson, 2000; Steenkamp, Hofstede and Wedel, 1999; Watkins and Liu, 1996). It is also noted that due to the many complexities involved in consumer responses to sales promotion, this study has necessarily been limited to a fairly narrow focus in order to isolate the impact of culture. For example, the effects of branding and prices have been kept constant. However, it has been suggested that consumer responses to brands (McCort and Malhotra, 1993) and prices (Laroche, Pons and Turnel, 2002) can differ across cultures. Thus, it would be worthwhile for future research to explore branding and pricing effects along with the impact of culture on consumer sales promotions. The study is also limited by its consideration of culture. This study ignores the fact that people may perceive themselves to belong to more than one ethnic culture and that the strength of identification with a particular ethnic group may differ between its members (Tan and McCullough, 1985). These themes deserve further consideration. Apart from culture, there are also likely to be other factors that will impact on the congruency framework and the effectiveness of sales promotions. For example, it has been shown that the role of guilt can affect the type of promotional benefits preferred (Kivetz and Simonson, 2002) and the nature of the decision can influence the choice between utilitarian and hedonic products (Dhar and Wertenbroch, 2000). Furthermore, the current study only focused on consumer promotions and consumer packaged goods. The congruency effects and the impact of culture may apply differently to business-to-business trade promotions and other types of products (e.g., services and industrial products). Analysis of these conditions would mean

25

going beyond student respondents; something that would increase variance in the data (particularly in the covariates). All of these areas represent opportunities for future research that can help extend our knowledge of sales promotion effectiveness. In terms of the methodology for this study, a quasi-experimental design is adopted and data are analyzed using inferential statistics. It is acknowledged, this is only one of the many possible methodologies that might be used. An alternative would be to observe the choice behavior of consumers at the point of purchase. This would accurately capture choice behavior, although it might be difficult to assign ethnicity. Another alternative is to use scanner data to measure brand choice, as has been adopted in previous studies of consumer promotion (Ehrenberg, Hammond and Goodhardt, 1994; Lemon and Nowlis, 2002; Mela, Gupta and Lehmann, 1997). Typically, this would be accompanied by a usage and attitude questionnaire, which could provide demographic, ethnographic and acculturation information. The study could be further extended by considering the use of alternative measures and stimuli. For example, culture may also be measured using Hofstedes (1990) original scale, or one of the alternatives that has been proposed (e.g., Furrer, Liu and Sudharshan, 2000). The results may then be compared with the CVSCALE to provide a form of triangulation. Another possible extension is to present the promotion scenarios with pictorial aids. The pictorial presentation of both the product and promotional offer may have an impact on consumer responses to the sales promotion scenarios. Finally, the generalizability of the results could be extended by considering other monetary and non-monetary promotions (e.g., coupons, loyalty schemes), and by broadening the list of utilitarian and hedonic products (e.g., other packaged goods or services). This is particularly important given the variety of promotional types found in most supermarkets, across a diverse range of products categories. It would also be worthwhile to explore the responses of other cultural groups within and across a variety of countries (e.g., Italian-Australians, Korean-Australians, North African-French, Hispanic-Americans, etc).

26

Appendix A: The CVSCALE Cultural Dimension Power Distance Values Measurement Items P1. People in higher positions should make most decisions without consulting people in lower positions. P2. People in higher positions should not ask the opinions of people in lower positions too frequently. P3. People in higher positions should avoid social interaction with people in lower positions. P4. People in lower positions should not disagree with decisions by people in higher positions. P5. People in higher positions should not delegate important tasks to people in lower positions. U1. It is important to have instructions spelled out in detail so that I always know what I'm expected to do. U2. It is important to closely follow instructions and procedures. U3. Rules and regulations are important because they inform me of what is expected of me. U4. Standardised work procedures are helpful. U5. Instructions for operations are important. Cl. Individuals should sacrifice self-interest for the group (either at school or the work place). C2. Individuals should stick with the group even through difficulties. C3. Group welfare is more important than individual rewards. C4.Group success is more important than individual success. C5.Individuals should only pursue their goals after considering the welfare of the group. C6.Group loyalty should be encouraged even if individual goals suffer. Ml. It is more important for men to have a professional career than it is for women. M2. Men usually solve problems with logical analysis; women usually solve problems with intuition. M3. Solving difficult problems usually requires an active, forceful approach, which is typical of men. M4. There are some jobs that a man can always do better than a woman. Dl. Careful management of money (thrift) D2. Going on resolutely in spite of opposition D3. Personal steadiness and stability D4. Long term planning D5. Giving up today's fun for success in the future D6. Working hard for success in the future 5-Point Scale 1 = Strongly agree 2 = Agree 3 = Neither agree/disagree 4 = Disagree 5 = Strongly disagree

Uncertainty Avoidance Values

1 = Strongly agree 2 = Agree 3 = Neither agree/disagree 4 = Disagree 5 = Strongly disagree

Collectivism Values

1 = Strongly agree 2 = Agree 3 = Neither agree/disagree 4 = Disagree 5 = Strongly disagree

Masculinity Values

1 = Strongly agree 2 = Agree 3 = Neither agree/disagree 4 = Disagree 5 = Strongly disagree

Confucian Dynamism Values

1 = Very unimportant 2 = Unimportant 3 = Neither important/unimportant 4 = Important 5 = Very important

27

Appendix B: Pretests (To be deleted or printed in smaller type if there are constraints on journal space) B.1 Pretest One: Design and Procedure The first pretest tests and confirms the nature of the promotion techniques. It also verifies the relationships between monetary and non-monetary promotions with utilitarian and hedonic benefits respectively. Short self-administered questionnaires were completed by a random sample of 15 AngloAustralian and 15 Chinese-Australian students. Recruitment was by random interception and participation was voluntary6 . Respondents were first screened for their education level and ethnicity, and only those who passed both screening criteria were retained. Respondents were randomly assigned a questionnaire consisting of either promotion stimuli set A or B. As shown in Table B1, each set includes an example of the four promotion techniques described in section 4. However, the stimuli examples are not specific to any product, as the purpose of the pretest is only to examine the nature of promotion techniques and their relationships with promotional benefits. Table B1: Promotion Stimuli used for Pretest One Promotion Type 1. Shelf-price discounts 2. Price packs 3. Sweepstakes 4. Free gifts Stimuli Set A Special! Save 50 cents. Get two for the price of one. Scratch the panels inside and win instantly! Free bag! Redeem by mail with receipt. Stimuli Set B Take 10% off the marked price. Get 20% extra for free. Win a trip to Hawaii! See details inside pack Buy this product and receive a free toy.

Each respondent was asked to: i) classify each promotion stimuli as either a monetary or nonmonetary promotion, and ii) evaluate each promotion stimuli based on the 18-item scale for measuring promotion benefits. B.2 Pretest One: Results Firstly, the results confirmed the nature of the four promotion techniques. Shelf-price discounts and price packs were correctly identified by the majority of respondents as monetary promotions, and sweepstakes and free gifts were identified as non-monetary promotions. Similar results were obtained between the two sets of stimuli and between the two ethnic groups (see Table B2). Table B2: Nature of Promotion Techniques (% of correct identification) Promotion Technique 1. Shelf-price discounts 2. Price packs 3. Sweepstakes 4. Free gifts Overall Sample 83% 63% 93% 87% Anglo-Australians 93% 60% 100% 87% Chinese-Australians 73% 67% 87% 87%

The total sample size of 30 is deemed sufficient for most statistical analysis (Hair et al, 1998).

28

Secondly, congruency relationships between the promotional types and promotion benefits were examined. Whilst Chandon, Wansink and Laurent (2000) analyzed their results with second-order confirmatory factor analysis, their purpose was scale and theory development. Here the only objective is to verify past findings. Thus, although confirmatory factor analysis may be employed, an alternative and a relatively more efficient set of analyses is used. The utilitarian and hedonic nature of the six types of benefits was first confirmed via reliability analysis. The items for the benefits of savings, quality and convenience produced a relatively high reliability alpha of 0.79. This supports the theory that these are three related utilitarian benefits. Similarly, the items for the benefits of value expressive, exploratory and entertainment had a high reliability alpha of 0.71. This confirms the hedonic nature of these three benefits. A utilitarian and hedonic factor score was then computed for each evaluation of a promotion technique. The scores represent the summated average of all the utilitarian and hedonic items respectively, as classified by Chandon, Wansink and Laurent (2000) from the 18-item scale for promotion benefits. Logistic regression was then applied in which the two factor scores acted as the independent variables. The dependent variable is the classification of a promotion technique as either a monetary or non-monetary promotion. Results are summarized in Table B3. The R2 value of 0.46 was relatively modest; however, the reduction in the -2LL value from an initial 163 to a final 113 suggests a good overall fit of the model. This is confirmed by the low significance values in the omnibus tests of model coefficients (0.00), which is another test of model fit. Table B3: Model Summary Results -2 Log likelihood -2 Log likelihood (initial) (model)* Summary 163 113 Omnibus Tests Chi-Square Df Model 50 2 *Enter method after 4 iterations Nagelkerke R2 0.46 Sig 0.00

From the regression results it is seen that both the utilitarian and hedonic factors were significant explanatory variables for the classification of promotion techniques. Specifically, the coefficient of the utilitarian factor is negative (e.g., B= -6.81, p = 0.00) whilst the coefficient of hedonic factor is positive (e.g., B= 4.10, p= 0.00). These results indicate that as the amount of utilitarian benefits decreases and the amount of hedonic benefits increases, the more likely that the sales promotion technique is a non-monetary promotion. Similarly, the higher the amount of utilitarian benefits and the lower the amount of hedonic benefits, the more likely that the sales promotion technique is a monetary promotion. The same relationships were observed between the two ethnic groups when analysed individually. The key regression results are summarised in Table B4. Table B4: Regression Results Variable Utilitarian Hedonic Constant B -6.81 4.10 0.76 SE 1.24 1.17 0.44 Wald 30.04 12.22 2.92 Df 1 1 1 Sig 0.00 0.00 0.09 Exp (B) 0.00 60.38 2.13

29

In summary, the results of pretest one largely replicated the findings of Chandon, Wansink and Laurent (2000). In particular, the results confirmed that shelf-price discounts and price packs can be classified as monetary promotions, with sweepstakes and free gifts as nonmonetary promotions. Furthermore, the results show that positive relationships exist between monetary and non-monetary promotions with utilitarian and hedonic benefits respectively. B.3 Pretest Two: Design and Procedure The second pretest verifies the utilitarian and hedonic nature of the pre-selected product categories that are used for the main experiment. The pretest also seeks to identify specific brands that are representative of high equity brands in each of the product categories. Following the same recruitment procedure as pretest one, self-administered questionnaires were completed by a new sample of 15 Anglo-Australian and 15 Chinese-Australian students on a voluntary basis. Respondents that passed the same screening criteria as pretest one were retained. Respondents were given short descriptions of the 6 pre-selected product categories and a list of high equity brands from each category. They were then asked to: i) evaluate each product category based on the utilitarian index adapted from Batra and Ahtola (1990), and ii) for each product category, select the brands that they have purchased in the past 12 months. B.4 Pretest Two: Results Firstly, the utilitarian index is calculated as the difference between two average 9-point semantic differential scores. Specifically, it is the difference between the average score of the two hedonic items of fun/not fun and pleasant/unpleasant, and the average score of the two utilitarian items of wise/foolish and useful/useless. The index scores based on individual responses are summated for each product category and then averaged to give a summary measure for the nature of each product. The utilitarian index ranges from 8 to 8, with a more positive number indicating a more utilitarian product. The results confirmed the utilitarian and hedonic nature of each product category (Table B5). Specifically, relatively higher index scores were recorded for laundry detergent (2.70), AA batteries (2.30) and film (1.38). This provides support for their utilitarian nature. On the other hand, relatively lower index scores were evident for chocolates (-1.40), ice-cream (-1.52) and biscuits (-0.77), which confirms their hedonic nature. The distinctions for each product category were also consistent across both ethnic groups. Table B5: Utilitarian Index Scores Product 1. Laundry Detergent 2. AA Batteries 3. Film 4. Chocolates 5. Ice-Cream 6. Biscuits Utilitarian Index 2.70 2.30 1.38 -1.40 -1.52 -0.77 Anglo-Australians 3.37 2.40 1.47 -2.03 -1.97 -0.83 Chinese-Australians 2.03 2.20 1.30 -0.77 -1.07 -0.70

With regard to identifying the specific brands to represent each product category, frequency tests were performed on the brands purchased by respondents in the last 12 months. The brands with the highest frequency were then selected to be representative of each category

30

and they include: Omo (laundry detergent), Energiser (AA batteries), Kodak (film), M&Ms (chocolates), Peters (ice-cream), and Arnotts (biscuits). These findings were consistent across both ethnic groups (Table B6). Table B6: Frequency Results Product Highest 2nd Highest 1. Laundry Detergent Omo (20)* Cold Power (13) 2. AA Batteries Energiser (23) Everready (14) 3. Film Kodak (29) Fuji (14) 4. Chocolates M&Ms (20) Cadbury (19) 5. Ice-Cream Peters (18) Cadbury (15) 6. Biscuits Arnotts (29) Westons (11) * All frequencies are out of a total of 30 respondents In summary, the results confirmed the utilitarian and hedonic nature of each product category. However, for the purposes of the main experiment, laundry detergent is excluded as analysis of the results indicated that a relatively small number of respondents had purchased brands from this product category. This suggests that laundry detergent is an inappropriate stimulus since the sample appears to lack the purchase experience to form evaluations. Biscuits are also excluded to ensure an even distribution of product types, with 2 utilitarian and 2 hedonic product stimuli 7 . This is consistent with the research design of Chandon, Wansink and Laurent (2000). The final configuration is shown in Table B7. Table B7: Product Stimuli Product 1. AA Batteries 2. Film 3. Chocolates 4. Ice-Cream Nature Utilitarian Utilitarian Hedonic Hedonic Brand Energiser Kodak M&Ms Peters

Biscuits are selected for exclusion over the other two hedonic products, as it was the least hedonic of all the hedonic products.

31

References Aaker, Jennifer L. and Durairaj Maheswaran (1997), The Effect of Cultural Orientation on Persuasion, Journal of Consumer Research, 24 (December), 315-328. Albaum, Gerald and Robert A. Peterson, (1984), Empirical Research in International Marketing, Journal of International Business Studies, 15 (Spring/Summer), 161-173. Au, Kevin Y. (1999), Intra-Cultural Variation: Evidence and Implications for International Business, Journal of International Business Studies, 30 (4), 798-813. Babakus, Emin, Peter K. Tat, William Cunningham (1988), Coupon Redemption: A Motivational Perspective, Journal of Consumer Marketing, 5 (2), 37-43. Babin, Barry J., William R. Darden, and Mitch Griffin (1994), Work and/or Fun: Measuring Hedonic and Utilitarian Shopping Value, Journal of Consumer Research, 20 (March), 64456. Batra, Rajeev and Olli T. Ahtola (1990), Measuring the Hedonic and Utilitarian Sources of Consumer Attitudes, Marketing Letters, 2 (2), 159-70. Blattberg, Robert C and Scott A. Neslin (1990), Sales Promotion: Concepts, Methods, and Strategies. Englewood Cliffs, NJ: Prentice Hall. Bochner, Stephen and Beryl Hesketh (1994), Power Distance, Individualism/Collectivism, and Job-Related Attitudes in a Culturally Diverse Work Group, Journal of Cross-Cultural Psychology, 23 (June), 233-57. Bond, Michael Harris (1988), Finding Universal Dimensions of Individual Variation in Multicultural Studies of Values: The Rokeach and Chinese Value Surveys, Journal of Personality and Social Psychology, 55 (6), 1009-15. Bridges, Eileen, Renee Florsheim A and John Claudette (1996), A cross-cultural comparison of response to service promotion, The Service Industries Journal, 16 (July), 265-287. Brodowsky, Glen H. and Beverlee B. Anderson (2000), A Cross-Cultural Study of Consumer Attitudes Toward Time, Journal of Global Marketing, 13 (3), 93-109. Burton, David (2001) 2001 Annual Report. Market Sizes and Shares, Retail World, December, 33-75. Calder, Nobby J., Lynn W. Phillips and Alice M. Tybout (1981), Designing Research for Application, Journal of Consumer Research, 8 (September), 197-207. Campbell, Leland and William D. Diamond (1990), Framing and Sales Promotions: The Characteristics of a 'Good Deal, Journal of Consumer Marketing, 7 (4), 25-31. Chandon, Pierre, Brian Wansink and Gilles Laurent (2000) A Benefit Congruency Framework of Sales Promotion Effectiveness, Journal of Marketing, 64 (October), 65-81.

32

Churchill, Gilbert A. (1979), A Paradigm for Developing Better Measures of Marketing Constructs, Journal of Marketing Research, 16 (February), 64-73. Clark, Terry (1990) International Marketing and National Character: A Review and Proposal for an Integrative Theory, Journal of Marketing, 54 (October), 66-79. Cutler, Bob D., S. Altan Erdem and Rajshekhar G. Javalgi (1997), Advertisers Relative Reliance on Collectivism-Individualism Appeals: A Cross-Cultural Study, Journal of International Consumer Marketing, 9 (3), 43-55. Dawar, Niraj and Philip Parker (1994), Marketing Universals: Consumers Use of Brand Name, Price, Physical Appearance, and Retailer Reputation as Signals of Product Quality, Journal of Marketing, 58 (April), 81-95. Dhar, Ravi and Klaus Wertenbroch (2000), Consumer Choice Between Hedonic and Utilitarian Goods, Journal of Marketing Research, 37 (February), 60-71. Dhar, Sanjay K. and Stephen J. Hoch (1996), Price Discrimination Using In-Store Merchandising, Journal of Marketing, 60 (January), 17-30. Dickson, Peter R. and Alan G. Sawyer (1990), The Price Knowledge and Search of Supermarket Shoppers, Journal of Marketing, 54 (3), 42-53. Donthu, Naveen and Boonghee Yoo (1998), Cultural Influences on Service Quality Expectations, Journal of Services Research, 1 (2), 178-186. Doran, Kathleen Brewer (1994), Exploring Cultural Differences in Consumer Decision Making: Chinese Consumers in Montreal, Advances in Consumer Research, 21, 318-322. Dowling, Grahame R. and Mark Uncles (1997), Do Customer Loyalty Programs Really Work? Sloan Management Review, 38 (Summer), 71-82. Durvalsula, Srinivas., J. Craig Andrews, Steven Lysonski and Richard G. Netemeyer (1993), Assessing the Cross-national Applicability of Consumer Behaviour Models: A Model of Attitude toward Advertising in General, Journal of Consumer Research, 19 (March), 62636. Ehrenberg, Andrew, Kathy Hammond and Gerald Goodhardt (1994), The After-Effects of Price-Related Consumer Promotions, Journal of Advertising Research, July-August, 11-21. Fletcher, Richard and Linden Brown (1999), International Marketing: An Asia-Pacific Perspective. Sydney, Australia: Prentice Hall. Fornell, Claes and David F. Larker (1981), Evaluating Structural Equation Models with Unobservable Variables and Measurement Error, Journal of Marketing Research, 18 (February), 39-50. Foxman, Ellen R., Patriya S. Tansuhaj and John K. Wong (1988), Evaluating Cross-National Sales Promotion Approach Strategy: An Audit Approach, International Marketing Review, 5 (Winter), 7-15.

33

Furrer, Olivier, Ben Shaw-Ching Liu and D. Sudharshan (2000), The Relationship Between Culture and Service Quality Perceptions, Journal of Services Research, 2 (4), 355-371. Goodenough, Ward H. (1971), Culture, Language and Society, Reading, MA: Modular Publications. Green, Corliss L. (1995), Differential Responses to Retail Sales Promotion Among AfricanAmerican and Anglo-American consumers, Journal of Retailing, 71(Spring) 83-92. Gurhan-Canli, Zeynep and Durairaj Maheswaran (2000), Cultural Variations in Country of Origin Effects, Journal of Marketing Research, 37 (August), 309-17. Hair, Joseph F., Rolph E. Anderson, Ronald L. Tatham and William C. Black (1998), Multivariate Data Analysis (Fifth Edition). New Jersey: Prentice Hall. Hirschman, Elizabeth C. and Morris B. Holbrook (1982), Hedonic Consumption: Emerging Concepts, Methods and Propositions, Journal of Marketing, 46 (Summer), 92-101. ---- (1981), American Jewish Ethnicity: Its Relationship to Some Selected Aspects of Consumer Behaviour, Journal of Marketing, 45 (Summer), 102-110. Hoch, Stephen J., Xavier Dreze, and Mary Purk (1994), EDLP, Hi-Lo, and Margin Arithmetic, Journal of Marketing, 58 (October), 16-27. Hofstede, Geert (1991), Cultures and Organizations: Software of the Mind. Berkshire, England: McGraw-Hill Book Company. Huff, Lenard C. and Dana L. Alden (1998), An Investigation of Consumer Responses to Sales Promotions in Developing Markets: A Three-Country Analysis, Journal of Advertising Research, 38 (May-June), 47-57. Jarvis, Steve (2002), Ethnic sites draw new ad wave, Marketing News, August 5, 4. Kale, Sudhir H. and McIntyre, Roger P. (1991), Distribution Channel Relationships in Diverse Cultures, International Marketing Review, 5 (3), 31-45. Kivetz, Ran and Itamar Simonson (2002), Earning the Right to Indulge: Effort as a Determinant of Customer Preferences Toward Frequency Program Rewards, Journal of Marketing Research, 39 (May), 155-70. Klineberg, Stephen L. (1968), Future Time Perspective and the Preference for Delayed Reward, Journal of Personality and Social Psychology, 8 (3), 253-257. Kroeber, Alfred L. and Clyde, Kluckholn (1952) Culture: A Critical Review of Concepts and Definitions, Anthropological Papers, 4, Peasboody Museum. Laroche, Michel, Frank Pons and Andree Turnel (2002), A Cross-Cultural Comparison of Direct Mail Receptivity, Journal of International Consumer Marketing, 14 (4), 5-24.

34

Lee, Eun-Ju, Ann Fairhurst and Susan Dillard (2002), Usefulness of Ethnicity in International Consumer Marketing, Journal of International Consumer Marketing, 14 (4), 25-48. Lemon, Katherine N. and Stephen M. Nowlis (2002), Developing Synergies Between Promotions and Brands in Different Price-Quality Tiers, Journal of Marketing Research, 39 (May), 171-85. Lenartowicz, Tomasz and Kendall Roth (1999), A Framework for Culture Assessment, Journal of International Business Studies, 30 (4), 781-98. Lyons, Kate (2002), Sales Promotion Grows Up, Professional Marketing, March, 6-7. McCort, Daniel John, Naresh K. Malhotra (1993), Culture and Consumer Behaviour: Toward an Understanding of Cross-Cultural Consumer Behaviour in International Marketing, Journal of International Consumer Marketing, 6 (2), 91-127. Mela, Carl F., Sunil Gupta, and Donald R. Lehmann (1997), The Long-Term Impact of Promotion and Advertising on Consumer Brand Choice, Journal of Marketing Research, 34 (May), 248-61. Millet, Michael (2002), Rapid social change divides the country, The Sydney Morning Herald, June 18, 1. Mittal, Banwari (1994), An Integrated Framework for Relating Diverse Consumer Characteristics to Supermarket Coupon Redemption, Journal of Marketing Research, 31 (November), 533-44. Mooij, Marieke de and Geert Hofstede (2002), Convergence and Divergence in Consumer Behavior: Implications for International Retailing, Journal of Retailing, 78 (1), 61-69. Nakata, Cheryl and K. Sivakumar (2001), Instituting the Marketing Concept in a Multinational Setting: The Role of National Culture, Journal of Academy of Marketing Science, 29 (Summer), 255-75. Osgood, Charles E. and Percy H. Tannenbaum (1955), The Principle of Congruity in the Prediction of Attitude Change, Psychological Review, 62 (1), 42-55. Peterson, Robert A. (2001), On the Use of College Students in Social Science Research: Insights from a Second-Order Meta-Analysis, Journal of Consumer Research, 28 (December) 450-61. Quester, Pascale G., Amal Karunaratna and Irene Chong (2001), Australian Chinese Consumers: Does Acculturation Affect Consumer Decision Making? Journal of International Consumer Marketing, 13 (3), 7-28. Roehm, Michelle L., Ellen Bolman Pullins and Harper A. Roehm Jr. (2002), Designing Loyalty-Building Programs for Packaged Goods Brands, Journal of Marketing Research, 39 (May), 202-13.

35

Roth, Martin S. (1995), The Effects of Culture and Socioeconomics on the Performance of Global Brand Image Strategies, Journal of Marketing Research, 32 (May), 163-75. Schwartz, Shalom H. and Wolfgang Bilsky (1987), Toward a Psychological Structure of Human Values, Journal of Personality and Social Psychology, 53 (3), 550-62. ---- and ---- (1990), Toward a Theory of the Universal Content and Structure of Values: Extensions and Cross-Cultural Replications, Journal of Personality and Social Psychology, 58 (5), 878-91. Sondergaard, Mikael (1994), Research Note: Hofstedes Consequences: A Study of Reviews, Citations and Replications, Organization Studies, 15 (3), 447-56. Spears, Nancy, Xiaohua Lin and John C. Mowen (2001), Time Orientation in the United States, China and Mexico: Measurement and Insights for Promotional Strategy, Journal of International Consumer Marketing, 13 (1), 57-75. Steenkamp, Jan-Benedict E.M., Frenkel ter Hofstede, and Michel Wedel (1999), A Crossnational Investigation into the Individual and National Cultural Antecedents of Consumer Innovativeness, Journal of Marketing, 63 (April), 55-69. Strahilevitz, Michal and John G. Myers (1998), Donations to Charity as Purchase Incentives: How Well They Work May Depend on What You Are Trying to Sell, Journal of Consumer Research, 24 (4), 434-46. Tan, Chin Tiong and James McCullough (1985), Relating Ethnic Attitudes and Consumption Values in an Asian Context, Advances in Consumer Research, 12, 122-125. Tellis, Gerald J. (1998), Advertising and Sales Promotion Strategy, Sydney Australia, Addison-Wesley. Totten, John C. and Martin P. Block (1987), Analyzing Sales Promotion: Text and Cases, Wellington, Chicago: Commerce Communications. Triandis, Harry C. (1989), The Self and Social Behaviour in Differing Cultural Contexts, Psychological Review, 96 (July), 506-20. ---- and Harry C. Hui (1990), Multimethod Probes of Individualism and Collectivism, Journal of Personality and Social Psychology, 59 (5), 1006-20. Usunier, Jean-Claude (2000), Marketing Across Cultures (3rd Edition), England: Prentice Hall. Watkins, Harry S. and Raymond Liu (1996), Collectivism, Individualism and In-Group Membership: Implications for Consumer Complaining Behaviours in Multicultural Contexts, Journal of International Consumer Marketing, 8 (3/4), 69-96.

36

Yoo, Boonghee and Naveen Donthu (2002), The effects of marketing education and individual cultural values on marketing ethics of students, Journal of Marketing Education, 24 (2), 92-104. ----, ---- and Tomasz Lenartowicz (2001), Measuring Cultural Values: Development and Validation of CVSCALE, working paper. Georgia State University.

37