THE INFLUENCE OF CONSUMERS’ ETHNIC IDENTITY ON ADVERTISING RESPONSES

BY

ELISABETH A. DER KINDEREN (SURINAME)

“This paper was submitted in partial fulfillment of the requirements for the Masters of Business Administration (MBA) degree at the Maastricht School of Management (MSM), Maastricht, the Netherlands, July 2005.”

The Influence of Consumers’ Ethnic Identity on Advertising Reponses

ii

The Influence of Consumers’ Ethnic Identity on Advertising Reponses

DEDICATION This research is dedicated to my parents for their unconditional support during my whole ‘MBA-journey’.

iii

The Influence of Consumers’ Ethnic Identity on Advertising Reponses

ACKNOWLEDGEMENT

A number of people made this research possible. I am grateful to all who have contributed to and supported my work. I would like to express my appreciation first to my supervisor Mr. Peter Coekelebergh who provided me with his kind and useful feedback

I sincerely recognize the contribution of Drs. Wim Zeggen who helped me a lot and I am thankful to him for sharing his knowledge and being so patience with me.

Thanks to my fellow students of MBA-intake 1, I cannot put in words my gratitude for their friendship and support during the past two years. I also appreciate the indefatigable support from my family and friends.

Special mention goes out to Eng. Patrick Krolis who introduced me to the MBA program and who has been supportive throughout the program.

Last but not least, I would like to extend my thanks to Mr. Hans Lim A Po, director F.H.R Lim A Po Institute for Social Studies and Mrs. Ollye Chin A Sen for making this MBA program a success and an experience worthwhile.

iv

The Influence of Consumers’ Ethnic Identity on Advertising Reponses

ABSTRACT Nearly all research on the relation between ethnicity and consumers’ response to advertisements has been done in the United States. The ethnic, racial and cultural composition of the population of Suriname is different from the population of the United States. This raises the question if findings of these researches - demonstrating the effects of consumers’ ethnic identity on responses to advertisements - are cross culturally robust. This research examines the cross culturally robustness of these findings by employing a ‘2 x 6 between subjects’ design to investigate the influence of consumers’ ethnic identity on advertising responses in Suriname. The focus of the research is on advertising responses of members of Suriname’s two largest ethnic groups: Hindustanis and Creoles. The types of advertising responses investigated are attitudinal responses: Attitude toward the Advertisement, Attitude toward the Model, Identification with the Model and Targetedness. The hypotheses test whether or not ethnicity of the model featured in the advertisement and the level of cultural embeddedness of advertisements influence advertising responses. Furthermore, the hypotheses test whether or not multi-cultural cues in an advertisement moderate the effect of model’s ethnicity on advertising responses and if viewers’ strength of ethnic identification affects these advertising responses. In total 126 subjects were involved with the experiment of which 49 identified themselves as Creole and 47 as Hindustani. Findings of this study indicate that Creoles as well as Hindustanis identify themselves more with the model of own ethnicity and feel more targeted by the advertisement featuring a model of similar ethnicity than by the advertisement featuring a model of dissimilar ethnicity. Findings also suggest that, contrary to expectations, neither Creoles nor Hindustanis have more favorable advertising responses to the high culturally embedded advertisements than to the low culturally embedded advertisements; only Hindustanis feel more targeted by a high Hindustani culturally embedded advertisement. Additionally, multi-cultural advertising cues do moderate the effect of a model’s ethnicity on viewers’ advertising responses; suggesting that an advertiser may be able to reach multiple groups of different cultural or ethnic backgrounds. Creoles as well as Hindustanis feel more targeted and have a more favorable attitude towards the multi-cultural embedded advertisement featuring a model of dissimilar ethnicity than towards an advertisement without these multi-cultural cues.

Keywords: Ethnic Identity, Advertising, Cultural Embeddedness

v

The Influence of Consumers’ Ethnic Identity on Advertising Responses

TABLE OF CONTENTS ACKNOWLEGDEMENT……………………………………………………...iv ABSTRACT…………………………………………………………………….v

CHAPTER ONE INTRODUCTION 1.1

Introduction................................................................................................................4

1.2

Suriname context .......................................................................................................4

1.2.1

Ethnic groups .........................................................................................................5

1.2.2

Advertising in Suriname.........................................................................................5

1.3

Problem Statement .....................................................................................................7

1.4

Objectives of Research ..............................................................................................7

1.5

Research Questions....................................................................................................8

1.6

Scope and Limitations of Research............................................................................9

1.7

Approach and Research Method..............................................................................10

1.8

Relevance of Research.............................................................................................10

1.9

Structure of Research Paper.....................................................................................10

CHAPTER TWO LITERATURE REVIEW 2.1

Introduction..............................................................................................................11

2.2

Marketing to Ethnic Consumers ..............................................................................11

2.3

Ethnic Identity..........................................................................................................12

2.3.1

Definition of Ethnic Identity.................................................................................13

2.3.2

Components of Ethnic Identity.............................................................................14

2.3.3

Types of Ethnic Identity .......................................................................................15

2.3.4

Previous Research on the Impact of Ethnic Identity............................................16

2.4

Theoretical Foundation ............................................................................................18

2.4.1

Distinctiveness Theory .........................................................................................18

2.4.2

Identification Theory............................................................................................21

2.4.3

Intercultural Accommodation ..............................................................................23 1

The Influence of Consumers’ Ethnic Identity on Advertising Responses

2.5

Advertising Effectiveness ........................................................................................26

2.5.1

Definition of Advertising Effectiveness................................................................26

2.5.2

Advertising Response Models ..............................................................................26

2.5.3

Techniques for Measuring Advertising Effectiveness ..........................................27

2.5.4

Copy Testing ........................................................................................................29

2.6

Summary ..................................................................................................................31

CHAPTER THREE METHODOLOGY 3.1

Introduction..............................................................................................................33

3.2

Hypotheses...............................................................................................................33

3.3

Design ......................................................................................................................36

3.4

Subjects ....................................................................................................................37

3.5

Stimuli material........................................................................................................37

3.6

Procedure .................................................................................................................38

3.7

Measurements ..........................................................................................................39

3.7.1

Ethnic salience .....................................................................................................39

3.7.2

Ethnic identity ......................................................................................................39

3.7.3

Attitude towards the Advertisement .....................................................................40

3.7.4

Attitude towards the Model..................................................................................40

3.7.5

Identification with the Model ...............................................................................41

3.7.6

Targetedness ........................................................................................................41

3.8

Analysis....................................................................................................................41

3.9

Summary ..................................................................................................................42

CHAPTER FOUR RESULTS 4.1

Introduction..............................................................................................................43

4.2

Salience of Ethnicity ................................................................................................43

4.3

Advertising Responses.............................................................................................43

4.3.1

Interaction and Main Effects................................................................................44

4.3.2

Hypotheses Testing..............................................................................................44

4.4

Summary ..................................................................................................................54 2

The Influence of Consumers’ Ethnic Identity on Advertising Responses

CHAPTER FIVE CONCLUSIONS AND RECOMMENDATIONS 5.1

Introduction..............................................................................................................55

5.2

Conclusions on The Influence of Ethnic Identity on Advertising Responses (in

Suriname).............................................................................................................................55 5.3

Conclusions on Cross Culturally Robustness of Previous Research .......................56

5.4

Limitations of the Study...........................................................................................58

5.5

Recommendations for Effective Advertisements ....................................................58

5.6

Recommendations for Future Research ...................................................................59

LIST OF REFERENCES………………………………………………………61

APPENDIX 1……………………………………………………………………I APPENDIX 2.………………………………………………………………..VIII APPENDIX 3…………………………………………………………………XX APPENDIX 4………………………………………………………………XXIX

TABLE OF FIGURES Figure 1. Strategies for Marketing to Ethnic Consumers .......................................................12 Figure 2. A Partial Model of Intercultural Accommodation...................................................24

Table 1. Mean Responses and Standard Deviations for Creole Identifiers .............................45 Table 2. Mean Responses and Standard Deviations for Hindustani Identifiers .....................46

3

The Influence of Consumers’ Ethnic Identity on Advertising Responses

CHAPTER ONE INTRODUCTION

1.1 Introduction An important challenge for a marketer in a multicultural marketplace is to make an advertisement that is relevant to many people as possible, without offending or alienating others. In the process of creating targeted advertisements for multicultural marketplaces, marketers look for meaningful characteristics by which to divide a single heterogeneous market into separate homogenous segments (Grier and Brumbaugh 2002). Ethnicity is one of those characteristics. In many cases, ethnic consumers have no visible preferences for certain product attributes or features, but the communication element regarding a product may vary greatly among consumers, in terms of values, languages, advertising appeals and media usage patterns. Therefore, when designing advertising messages and selecting channels, marketers need to ensure that one is sensitive to the traditions and feelings of ethnic consumers (Chui 1997).

The title of the paper already indicates that this research paper deals with the subject of consumer’s ethnic identity in relation to their advertising responses. In order to put the subject in a Surinamese context the next paragraph will discuss briefly the different ethnic groups in Suriname and provides a general overview of advertising in Suriname. This first chapter is an introductory chapter. After the paragraph on Surinamese context subsequently the problem statement, the research objectives, scope, research questions, research methodology and relevance will be discussed. Chapter one will finish off with an overview of the structure of this research paper.

1.2 Suriname context As mentioned above, this paragraph is dedicated to discuss briefly the different ethnic groups in Suriname and to provide an impression on advertising in Suriname.

4

The Influence of Consumers’ Ethnic Identity on Advertising Responses

1.2.1 Ethnic groups The population of Suriname is very much ethnically diverse. Suriname's population consists of: •

Hindustani (37%): the ‘East Indians’, their ancestors emigrated from northern India in the latter part of the 19th century;

•

Creole (31%): African or mixed African and European ancestry.

•

Javanese (15%): their ancestors emigrated from Java (Indonesia) to Suriname in the last decennia of the 19th century;

•

Maroons (10%): the Maroons are the descendents of escaped slaves who fought themselves free and established viable, autonomous communities along the major rivers of Suriname rainforest interior in the 17th and 18th centuries;

•

Amerindian (2%): the indigenous people, the original inhabitants of Suriname;

•

Chinese (2%);

•

White (1%) and other (2%)

Dutch is still the official language, but in the streets you hear a "lingua franca" called Sranang Tongo, which has elements from English, African, Dutch and many other languages. But the original mother-languages are also spoken within cultural groups, such as Surinamese-Hindustani (Sarnami), Indonesian (Javanese), Chinese and several Amerindian and Maroon languages. Christianity, Hinduism and Islam are all practiced religions. Most ethnic groups tend to maintain their own language, culture and religion, which is apparent also throughout the existing media, politics and (social) organizations.

1.2.2 Advertising in Suriname The most common forms of advertising in Suriname are TV advertisements, radio advertisements, print advertisements, billboards and wall paintings. There are currently about fifteen television channels, more than twenty five radio stations, four daily newspapers and several periodicals and magazines. About twenty-five percent of a newspaper consists of advertisements and ten to eighteen minutes per hour are used for commercial purposes on the radio. This figure is lower for television. Radio advertising is relatively cheap, as well as television advertising. Print advertising is one of the most expensive forms of advertising in Suriname.

5

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Media organizations, as well as advertising agencies and other practitioners in the field of advertising declare that there has been an increase in advertising over the last five years. Although there are no numbers available to support this statement, they state that the number of advertisers as well as the frequency of advertising has increased.

Surinamese consumers are being exposed to locally produced advertisements as well as international produced advertisements. These international or ‘foreign’ advertisements promoting an international brand or product are most of the time in English and feature models with whom probably only a small part of the viewers can identify based on the ethnicity of the model. For example, TV commercials for Dutch beer (e.g. Grolsch, Bavaria) feature usually only white models, or Coca-Cola commercials aired in Suriname - which have been produced in the Caribbean or Latin- America - feature typically black or Latin models. In locally produced advertisements one often tries to represent the diverse ethnic groups by using several models with different ethnicity or one has a tendency to use models in advertisements that represent their own ethnic group. In other words, an advertisement for a company that is owned by a person with a Hindustani ethnicity features usually models with a Hindustani ethnicity.

A media- and advertising research of 2002 provides insight on the ‘viewing -, reading and listening habits’ of consumers in Suriname. This research was conducted among 600 respondents living in different areas of Suriname by the Institute for Social Scientific Research (IMWO). The findings of this research demonstrate that: 92% of the population reads one or more newspapers, or magazine available in Suriname. The main reason for selecting a newspaper or magazine to read is the news and other information the reader is able to obtain from that newspaper or magazine. The advertising in the newspapers or magazines does not influence the selection of a newspaper of magazine; 99% of the population listens to one or more radio channels. One listens during the whole day to the radio. The news- and informative character of the preferred radio station is the most important reason for the majority for listening to a radio channel. The second most import reason is the ethnic language spoken on a channel. The combination of these two reasons explains the popularity of different radio channels in different areas of Suriname, where the population is dominated by certain ethnic group;

6

The Influence of Consumers’ Ethnic Identity on Advertising Responses

98% of the population watches one or more Surinamese television channels. One mainly watches television in the evening. The choice for a certain channels depends on the programs shown. As well as for selecting a radio or television channel the advertising on the different channels does not have any impact on the selection. Other important findings of this research were that 86% of the respondents were able to recall an advertisement out of the newspaper, 89% remembered an advertisement on the radio and that 92% of the respondents remembered a television advertisement. Advertisements in newspapers were mainly remembered because of a rich illustrative lay-out. The likeability of the advertisement and the frequency of being exposed to an advertisement were the main factors for recalling advertisements on radio and television. Further more this research also concluded that respondents prefer advertisement in Dutch (official language) above English, Sranang Tongo or one of the ethnic languages. Respondents seemed not to have a preference for types of music, gender and ethnicity of the models. It is important to mention that the respondents were directly addressed with these questions, in contradiction to the methods employed in the majority of research dealing with these issues.

1.3 Problem Statement Nearly all research on the relation between ethnicity and consumers’ response to advertisements has been done in the United States.

The ethnic, racial and cultural

composition of the population of Suriname is different from the population of the United States. This raises the issue whether or not findings of these researches - demonstrating the effects of consumers’ ethnic identity on responses to advertisements - are robust cross culturally.

1.4 Objectives of Research The objectives of this research are: •

To provide an (partial) answer to the question raised in the problem statement by investigating the effect of consumers’ ethnic identity on response to advertisements in Suriname

•

To make recommendations for effective advertisements in Suriname as a multicultural marketplace. 7

The Influence of Consumers’ Ethnic Identity on Advertising Responses

1.5 Research Questions In order to understand the effect of consumers’ ethnic identity on advertising response in Suriname it is important to know whether or not ethnicity can be perceived as a distinctive salient trait for Surinamese consumers as part of an ethnic diverse population. Following the distinctiveness theory of McGuire and colleagues (1978) it is to be expected that ethnicity is more salient to people living in ethnically mixed society than to people living in a segregated society, and that consumers of a numerical minority group will be more likely to deem their ethnic identity salient than members of numerical majority group would.

RQ 1:

Is ethnicity a distinctive salient trait for Surinamese consumers?

The authors of the study “A Tale of two cities: Distinctiveness Theory and Advertising Effectiveness” (Deshpande and Stayman 1994) found that members of minority (versus majority) groups were more likely than majority groups to have their ethnicity salient. Furthermore they found that members of minority groups find an advertisement spokesperson of their own ethnicity to be more trustworthy which led to more positive attitudes towards the advertisement and to the brand being advertised. The results of the study by Whittler and Spira (2002) “Model’s Race: A Peripheral cue in Advertising Messages” suggests that black viewers’ differential responses to an advertisement featuring black and white models are moderated by the level of their identification with black culture, Specifically, blacks who identify strongly with the black culture (strong black identifiers) have more favorable evaluations of advertisement, product and model when the model is black, whereas weak black identifiers’ evaluations do not differ on models’ race. RQ 2a:

Does the ethnicity of the model featured in the advertisement affect viewers’ responses to the advertisement?

RQ 2b:

Does viewers’ level of ethnic identification affect viewers’ responses -- influenced by the ethnicity of the model featured in the advertisement - to the advertisement?

The study “Ethnic Identification on Adolescents’ Evaluations of Advertisements” (Osei Appai 2001) does not perform the research with only a model’s ethnicity taken into account, but also examines the effects of different cultural advertising cues. The study suggests that 8

The Influence of Consumers’ Ethnic Identity on Advertising Responses

the presence of black characters in advertisements, and in some cases black cultural cues, significantly influenced black and white adolescents’ responses. The main effects indicated that black adolescents felt more targeted by, and rated more favorably, black characters in culturally embedded advertisements than they did white characters in culturally embedded advertisements.

RQ 3a:

Does the level of cultural embeddedness of advertisements affect viewers’ responses to the advertisement?

RQ 3b:

Does viewers’ level of ethnic identification affect viewers’ responses to high and low cultural embedded advertisements?

A research by Greenlee and Oakenfull (2000) suggests that an advertiser may be able to reach multiple target segments with one advertising appeal that combines cues viewed favorably by consumers from different cultural backgrounds. Does this indicate that the effect of a model’s ethnicity on consumers’ responses to advertisements can be moderated by using multi-cultural cues in the advertisement?

RQ 4:

Do multi-cultural cues in the advertisement moderate the effect of a model’s ethnicity on viewers’ responses to the advertisement?

1.6 Scope and Limitations of Research Even though this research will explore ethnic identification and the advertising responses of members of the different ethnic groups in Suriname, this study will focus on the advertising responses of members of the two largest ethnic groups, that is the Hindustani and the Creole group. The types of advertising responses to be investigated in this study are limited to four attitudinal responses: attitude toward the advertisement; attitude toward the model; identification with the model and targetedness. It should be clear that this study will only be a small part of a much more complex study.

9

The Influence of Consumers’ Ethnic Identity on Advertising Responses

1.7 Approach and Research Method The literature study will be based on previous researches and underlying theories published in academic journals. In order to provide answers to the research questions and address the issue raised in the problem statement a “2 x 6 between-subjects” (see chapter three) experiment will be performed.

1.8 Relevance of Research Since there is little to no knowledge on this subject based on actual research in Suriname it is expected that the findings of this research should be of interest to both academics and practitioners in the field of marketing and advertising.

1.9 Structure of Research Paper This research paper consists of five chapters. Chapter one is an introductory chapter discussing the background of this research, the research objectives and research questions together with the relevance of this research. Chapter two will provide a brief overview of marketing strategies to ethnic consumers, followed by the discussion on the key concept ‘Ethnic Identity’. The relation between ethnic identity and advertising responses will be explored by reviewing findings of previous research. Thereafter three theories used by researchers to interpret and understand these findings will be explored. In order to complete the literature review relevant advertising key concepts will be discussed. Chapter three will provide the hypotheses to be tested and explain the employed methodology of research in this study. The results of the research will be presented and discussed in chapter four and this paper will be concluded by chapter five discussing the conclusions, the limitations of the study and by providing recommendations for effective advertisements in Suriname. Also recommendations for future research will be done.

10

The Influence of Consumers’ Ethnic Identity on Advertising Responses

CHAPTER TWO LITERATURE REVIEW

2.1 Introduction This literature review commence with a brief overview of marketing strategies to ethnic consumers followed by the discussion on the key concept ethnic identity. The relation between ethnic identity and advertising responses will be explored by reviewing findings of previous research. Thereafter three theories used by researches to interpret and understand these findings will be provided. In order to complete the literature review relevant advertising key concepts will be discussed.

2.2 Marketing to Ethnic Consumers Diversity of the population requires that marketers understand the needs from each group, as their demographics, media usage, consumption patterns of each group. For marketers to design effective marketing strategies in order to reach these groups, one should examine the meanings of concepts such as race, ethnicity, ethnic identification and their interaction with the marketing mix variables (Cui 1997). An important question marketers have to deal with nowadays is whether or not standardized marketing strategies can be used in a market with ethnic diverse consumers without adapting the marketing strategy to the different ethnic target groups. The study “Marketing Strategies in a Multi-ethnic environment” by Geng Cui (1997) provides a framework (see figure 1.) of alternative strategies for marketing to consumers of diverse ethnic background. This framework compares four distinctive approaches to marketing to ethnic consumers: standardized marketing, product adaptation, advertising adaptation and ethnic marketing, and discusses the optimal environment for adopting each strategy (Cui 1997). It deals primarily with two elements of the marketing mix; product and promotion (communication). In order to determine which of the four strategies a viable strategy is, the marketer needs to answer two essential questions: 1) Are the needs and product preferences of ethnic minorities similar to or different from those of the majority population? And 2) are the media usage patterns and advertising responses of ethnic minorities similar to or different from those of the majority population? (Cui 1997).

11

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Figure 1. Strategies for Marketing to Ethnic Consumers COMMUNICATION

PRODUCT

Similar:

Different:

Standardize

Customize

I

III

Total Standardization:

Advertising

Similar:

Use current marketing

Adaptation:

Standardize

mix without

Use current marketing

II

IV

Different:

Product Adaptation:

Ethnic Marketing:

Customize

Use current marketing

Use totally new

mix except product

marketing mix Source: Geng Cui (1997).

In many cases, ethnic consumers have no visible preferences for certain product attributes or features. However the communication element regarding a product may vary greatly among consumers, in terms of values, languages, advertising appeals and media usage patterns. The language barrier often requires marketers to translate or re-create their messages. Different advertising responses and media usage patterns mandate modification of advertising messages and selection of channels (Cui 1997). However when designing advertising messages and selecting channels, marketers need to ensure that these decisions are sensitive to the traditions and feelings of ethnic consumers (Cui 1997). The sensitivity of those traditions and feelings are determined by the ethnic identity of a consumer.

2.3 Ethnic Identity This segment will provide the definition of ethnic identity, descriptions of the components of ethnic identity together with an overview of the different types of ethnic identity. This part of the literature review will be finished with a review of previous research on the impact of ethnic identity on consumers ethnic attitudes, consumption (of ethnicassociated) products and response to ethnic cues (e.g. ethnic actors) in advertising.

12

The Influence of Consumers’ Ethnic Identity on Advertising Responses

2.3.1 Definition of Ethnic Identity The first part of the construct, the term ethnic, (originally also ethnik[e] and even ethnique) is an old term in English and glossed in the Oxford English Dictionary as "heathen, pagan, uncouth, neither Christian nor Jewish". The term has Latin and Greek origins. Ethnicus and ethnikas in Latin both mean nation. Ethos in Greek means custom, disposition or trait. Taken the two terms ethnikas en ethos together can mean something like ‘people living together (nation) who share and acknowledge common customs’. The second part of the construct, identity, is derived from the Latin word identitas; the word is formed from idem meaning same (Trimble and Dickson). A more precise meaning of identity given by Simpson and Weiner (1989) and cited by Trimble and Dickson is “the sameness of a person or thing at all times in all circumstances; the condition or fact that a person or thing is itself and not something else”.

The discussion on ethnic identity is interdisciplinary and, therefore highly diversified in its underlying theories and assumptions. Definitions of ethnic identity vary according these underlying theories. There is no widely agreed on definition of ethnic identity. The psychologist Jean Phinney notes in her study “Ethnic Identity in Adolescents and Adults: Review of Research” (1990) that researchers appeared to share a broad understanding of the term ethnic identity, but the specific aspects that they emphasized differ widely. These differences are due to the diversity how researchers have conceptualized ethnic identity and in the questions the have sought to answer. The most widely used definition of ethnic identity in the psychology and the definition to be used in this research, is developed by Phinney, “Ethnic identity is a dynamic multidimensional construct that refers to one’s identity, or sense of self as a member of an ethnic group. Ethnic identity is not a fixed categorization, but rather a fluid and dynamic understanding of self and ethnic background. Ethnic identity is constructed and modified as individuals become aware of their ethnicity, within the large (social cultural) setting” (2003). Phinney (2003) states that “an ethnic group claims a common ancestry and of which members share at least a similar culture, race, religion, language, kinship or place of origin”. According to Smith (1991) “an ethnic group may be defined as a reference group called upon people who share a common history and culture, who may be identifiable because they share similar physical features and values, and who - through the process of interacting with each other and establishing boundaries with others - identify themselves as being a member of that group”.

13

The Influence of Consumers’ Ethnic Identity on Advertising Responses

2.3.2 Components of Ethnic Identity Previous research addressing research questions on ethnic identity can basically be divided in two types of studies. The first type concerns studies with the focus on what might be called the state of ethnic identity, which is a person’s identification at a given time. The second types of studies are those focused on stages of ethnic identity, or changes over time in a person’s identification (Phinney 1990). Elements being used assessing the state of one’s ethnic identity are: ethnicity and self-identification, sense of belonging, attitude towards the ethnic group and ethnic involvement. Following a brief description of the different components of ethnic identity based on the examination of previous researches by Phinney (1990): • Ethnicity and Self Identification: Self-identification refers to the ethnic label that one uses for oneself. If a person selfIdentifies as a member of a particular ethnic group, then he or she is willing to be perceived and treated as a member of that group. Thus self-ascribed ethnic labels are the obvious manifestations of individuals' identification with a particular ethnicity (Fishman 1999).

However, among those who are racially distinct, by features or skin color, or

whose culture (language, dress, customs, etc.) clearly distinguishes them from other groups, self-identification is at least partly imposed. • Sense of Belonging: People may use an ethnic label when specifically asked for one and yet may not have a strong sense of belonging to the group chosen. Therefore it is important to asses to feeling of belonging. A sense of belonging to one’s own group can also be defined in contrast to another group by the experience of exclusion, contrast or separateness from other groups. • Attitudes towards One’s Ethnic Group: In addition to their self-identification and a sense of a belonging, people can have both positive and negative attitudes towards their own ethnic group. Positive attitudes include pride, in and pleasure, satisfaction and contentment with one’s own group. The absence of positive attitudes can be seen as a denial of one’s ethnic identity.

14

The Influence of Consumers’ Ethnic Identity on Advertising Responses

• Ethnic Involvement: Involvement in the social life and cultural practices of one’s ethnic group is seen as one of the most important components of one’s ethnic identity. Indicators of ethnic involvement are language, friendship, social organizations, religion, cultural traditions and politics.

The ethnic identity of an individual can be measured by the extent to which the individual has a sense of belonging to the ethnic group, a positive or negative attitude towards the group and uses the signs, symbols and language of the culture associated with the ethnic membership group. The range of identification may vary from little or no ethnic identification with one’s membership group to high identification with the group in question (Smith 1991). Strong or high ethnic identifiers are likely to display attitudes and behaviors that are consistent with the core cultural values of their ethnic group.

2.3.3 Types of Ethnic Identity Ethnic identity is to a large extent defined by context (Phinney 1990). Amada M. Padilla cited Heller in Handbook of Language & Ethnic Identity that ethnicity becomes important only when ethnic groups come into contact with each other. “In isolation, it means nothing to be English, Arab, French, or Mexican. Ethnicity may be more meaningful in certain intergroup contexts than in other situations, moreover, the ethnic label that an individual chooses to wear may differ according to social context” (Fishman 1999). There are different researches providing evidence that one’s ethnic identity varies in strength according context. Saylor and Aries (1999) examined in their study maintenance and change in ethnic identity among U.S. college students over the course of an academic year as they moved through a significant change in social context from home to a residential college. Although measures of ethnic identity at the beginning and end of the year were highly correlated, results of the study showed that the students showed a significant increase in ethnic identity and greater involvement in ethnic behaviors and practices by the end of the year. Joining ethnic organizations provided support for ethnic identity. Tiffany Yip and Andrew J. Fuligni provided with their study (2002), examining the links among Chinese American adolescents’ ethnic identity and their ethnic behaviors, ethnic identity salience and psychological well-being, support that feelings of ethnic identity vary considerably on a daily basis and that situational characteristics can be responsible for 15

The Influence of Consumers’ Ethnic Identity on Advertising Responses

fluctuations in feelings of ethnic salience. For example, participation in ethnic behaviors such as speaking Chinese, eating Chinese Food and interacting with other Chinese people can influence momentary feelings of one’s ethnic identification. This momentary state of ethnic identification is called by others ethnic-self awareness, a temporary state during which a person is more sensitive to his of her ethnicity, or also explained as the momentary salience of people’s ethnic identification (Forehand and Deshpande 2001). Ethnic self-awareness is similar to the term felt-ethnicity, introduced by Stayman and Deshpande (1989). They distinguished felt-ethnicity from objective ethnic membership by arguing that ethnicity is “not just who one is, but how one feels in and about a particular situation” (Stayman and Deshpande 1989). Donthy and Cherian (1994) identified a similar distinction in types of ethnic identification: enduring- and episodic ethnic identification. “Enduring identification refers to the base-level of ethnic group membership and episodic identification refers to situational surges of ethnic feelings brought on by relevant cultural events “(e.g. Stayman and Deshpande 1989; Wooten and Galvin 1993). This paper concentrates only on the enduring ethnic identification of consumers.

2.3.4 Previous Research on the Impact of Ethnic Identity Research has demonstrated that individuals vary in their degree in identification with their ethnic group. Researchers have also demonstrated that consumer strength of ethnic identity is correlated with ethnic attitudes, consumption (of ethnic-associated) products and response to ethnic cues (e.g. ethnic actors) in advertising (Appaih 2001; Donthy and Cherian 1994; Green 1999; Whittler and DiMeo 1991; Whittler and Spira 2002; Young-Kyung 2001). The authors of the study “Impact of Strength of Ethnic Identification on Hispanic Shopping Behavior” argue that the strength of ethnic identification moderates retail shopping behavior of Hispanic consumers. The study explores the differences in retail shopping behavior of strong and weak Hispanic identifiers. The results show that strong Hispanic identifiers prefer more than weak Hispanic identifiers Hispanic vendors in general, and for low-involvement services in particular. Other results show that: strong Hispanic identifiers are more brand loyal (especially to brands purchased by their parents and friends) than weak Hispanic identifiers; and strong identifiers are more affected by advertisements (especially those targeted on Hispanics) than weak identifiers (Donthu and Cherian 1994). The study by Corliss Green (1999) examines the effects of strength of ethnic identification, media placement, and racial composition of advertisements on attitudes and 16

The Influence of Consumers’ Ethnic Identity on Advertising Responses

purchase intentions among African-Americans. Results indicate that strong ethnic identifiers generally have more positive evaluations of advertisements that feature African-American in positions of dominance and are place in racially targeted media, whereas weak ethnic identifiers have more positive evaluations of advertisements that feature whites in position of dominance and are placed in non targeted media. In addition, the type of product advertised, in this case two personal care products: foundation, a race-based product, and perfume, a racially neutral product, effects audience members’ evaluations. The study of Whittler and DiMeo examines viewers’ processing of ethnic cues in advertising stimuli. The findings indicated that, regardless of their attitude towards blacks, whites were less likely to purchase the products and had less favorable attitudes towards the products and the advertisements when the advertisements featured black actors (Whittler and Dimeo 1991). This study was conducted among 160 white adults. Findings of study by Whittler and Spira, among a total of 160 black adults, suggest that black’ differential responses to advertisements featuring black and white models are moderated by their identification with black culture. Specifically, blacks with a high ethnic identification have more favorable evaluations of the advertisement, product, and model when the model is black than when the model is white. Evaluations of blacks with a low identification do not differ depending on the model’s ethnicity (2002). A similar study is the study by Osei Appaih (2000) that examines whether the strength of ethnic identity influences black and white adolescents’ responses to advertisements featuring models of different races. The researcher digitally manipulated the race of characters in the advertisements as well as the number race-specific cultural cues while maintaining all other visual features of the advertisements. One hundred seventy-three black and white adolescents evaluated black or white character advertisements. The findings demonstrate that black adolescents who have a strong black ethnic identity perceive themselves more similar to and identify more strongly with black character advertisements than black adolescents with weaker ethnic identities do. Other results suggest that white adolescents, despite their ethnic identity, find black character advertisements with varying degrees of black cultural cues as appealing as similar white character advertisements (Osei Appaih 2001).

Researchers have used several theories and or models that may explain the correlation between ethnic identity and how people deal with ethnic cues in advertising. Three underlying theories are: Distinctiveness theory (McGuire et al. 1978); Kelman’s (1961) Identification Theory and the Intercultural Accommodation model (Holland and Gentry 17

The Influence of Consumers’ Ethnic Identity on Advertising Responses

1997).

2.4 Theoretical Foundation In the next sections the three theories: Distinctiveness theory, Identification theory and the Intercultural Accommodation model will be explained together with applications in several researches.

2.4.1 Distinctiveness Theory The central prediction of the Distinctiveness theory is that an individual’s distinctive trait in relation to other people in the environment will be more salient to the individual than more common traits (McGuire et al. 1978). The theory was developed as a response to a perceived dominance of a “reactive approach” in self-concept research that studied participants’ reactions to dimensions of the self which have been selected a priori by the researcher. McGuire and his colleagues felt that it was more appropriate to investigate individuals’ self-concept by exploring the dimensions that one spontaneously chooses to describe oneself instead of those chosen by the researcher (McGuire, McGuire, Child and Fujioka 1978). They hypothesized that the spontaneous salience of a personal characteristic to the self-concept is evoked by the distinctiveness of that characteristic in a given situation, and that it will become part of one’s identity at that moment. For example, a black woman will likely to be more aware of her race when she is associating with white women, but more aware of her gender when she is associating with black men. So, the postulate implies that one is conscious of oneself insofar one is different and perceives oneself in terms of these distinctive features, and by changing a person’s social context so that different characteristics become distinctive, a person’s self-concept can be changed in a predictable way (McGuire, McGuire, Child and Fujioka 1978). The study “Salience of Ethnicity in the Spontaneous Self-concept” (McGuire et al 1978) investigated ethnicity as a dimension of the self-concept. The distinctiveness postulate was used to predict the salience of ethnicity in the spontaneous self-concept. The postulate implies that ethnicity is more salient in people’s self-concepts in an ethnically mixed society than in a segregated society, and that in an integrated society, ethnicity is more salient in the selfconcepts of members of the minority group than of the majority group (McGuire et al 1978). The measure of salience of ethnicity used in this study was whether children mentioned 18

The Influence of Consumers’ Ethnic Identity on Advertising Responses

spontaneously their ethnicity in response to a nondirective “Tell us about yourself” question. The findings of the study were supportive of the distinctiveness postulate. The findings suggest that people do partly act like information processing-machines, noticing aspects of their environment insofar as they contain information and that one is more conscious of one’s ethnicity (or religion, sex , etc.) to the extent that one’s own kind on this variable is in the minority in one’s usual social environment and that ethnicity (or religion, etc.) becomes more salient in people’s self-concept as their social milieu becomes more heterogeneous in this regard; An integrated society heightens rather than erases consciousness of ethnicity and feelings of difference between the integrated groups (McGuire, McGuire, Child and Fujioko 1978). The distinctiveness postulate has been applied to consumer behavior and advertising contexts and supports the notion that numeric minorities respond more favorably to marketing efforts designed to resonate with their distinctive characteristics. Previous research has found that consumer distinctiveness results in a heightened sensitivity to targeting efforts, more identification with and trust of a similar source, and increased favorability towards advertisement and brand (Aaker, Brumbaugh and Grier 2000; Appaih 2001; Deshpande and Stayman 1994; Forehand and Deshpande 2001; Grier and Brumbaugh 1999). One of the first research to apply the distinctiveness theory within an advertising context is the study “A Tale of Two Cities: Distinctiveness theory and Advertising Effectiveness” by Deshpande and Stayman (1994). They integrated prior research on strength of ethnic identification, the effect of race of actors in ads, and distinctiveness theory to derive a model and a set of hypothesis concerning how distinctiveness applies to advertising effects. The study tested that: ethnic situation affects ethnic salience; ethnic salience moderates the influence of the ethnicity of the spokesperson on spokesperson trustworthiness; and that these changes in spokesperson trustworthiness are expected to impact brand attitude. The research was conducted among 205 adult subjects. The subjects’ ethnic situation was varied by recruiting subjects from one city, San Antonio, in which Hispanics are in the majority (and Anglos in the minority) and a second city, Austin, in which Anglos are in the majority (and Hispanics in the minority). The researchers have found strong support for McGuire’s distinctiveness theory within an advertising context. More specifically, minority group consumers are more likely to spontaneously evoke their ethnicity when they are in a numerical minority rather than a majority in their cities. Hispanic Americans in the study were more likely to spontaneously mention their ethnicity when they lived in Austin (where they constitute a minority of the 19

The Influence of Consumers’ Ethnic Identity on Advertising Responses

city’s population) than when they lived in San Antonio (where they constitute a majority) and vice versa for Anglos. Furthermore, there appears to be a carryover between ethnic identity and responsiveness to ethnic elements of advertising. Hispanic (the traditional “minority” group) consumers were more likely to believe that a Hispanic spokesperson was trustworthy when they lived in Austin than they lived in San Antonio. Hispanic consumers in Austin were also more likely than those in San Antonio to have positive attitudes towards a brand for which the ad featured a Hispanic spokesperson. These results apply also for the Anglos (the traditional “majority” group). Based on these findings the authors concluded that ethnically targeted stimuli (e.g. an advertisement featuring a spokesperson of the same ethnicity as a targeted viewer) were more effective among numerically distinctive consumers than nondistinctive consumers (Deshpande and Stayman 1994).

Other researchers integrated distinctiveness theory into research on ethnicity and target marketing (Aaker, Brumbaugh and Grier 2000; Grier and Brumbaugh 1999). The research by Grier and Brumbaugh “Noticing Cultural Differences: Ad Meanings created by Target and Non-target markets” (1999) examined the meanings consumers create from marketing efforts targeting black heterosexuals (distinctive based on ethnicity), white heterosexuals (nondistinctive), and white homosexuals (distinctive based on sexual orientation), when consumers are (are not) members of the intended target market. Results provide empirical support for the proposition that target and non-target consumers differ in the way the construct meanings from targeted advertisements. Target market consumers are likely to take a positive referential interpretive attitude and create meanings that relate the advertisement positively to themselves. Non-target market consumers are likely to take a negative referential interpretive attitude when reading an advertisement targeted to another group. Target and non-target members differ in their ability to interpret specific cultural cues embedded in the advertisement. One of the findings of this study is that the meanings target and non-target consumers create from targeted advertisements depends, among other things, on the distinctiveness of the viewer. Members of numerically less prevalent groups, blacks and gay viewers in the study, were particularly aware of being targeted and incorporated that awareness into the meanings they derived from targeted advertisements. Black and gay viewers were able to create favorable links between themselves and the advertisement more readily than white viewers due to the relative infrequency with which those groups are targeted and their heightened salience of distinctive characteristics. Their ability to use the cultural cues embedded in the advertisements added meanings for them that were not 20

The Influence of Consumers’ Ethnic Identity on Advertising Responses

available to non-target viewers. On the contrary white viewers were less likely to make meaningful links between the advertisements and themselves, even when they were the target market. Apparently “whiteness” is neither a salient nor a meaningful characteristic for those non-distinctive viewers, and targeting on the basis of this characteristic does not factor into the meaning they create (Grier and Brumbaugh 1999). Aaker, Brumbaugh and Grier (2000) also explored this moderating role of distinctiveness on the effects of targeted advertising. Their research examined the intended and unintended effects of target marketing on members of the advertiser’s target market as well as members of the non-target market. The results demonstrate that the effects of targeted are moderated by viewer distinctiveness. Unfavorable non-target market effects are stronger for members of non-distinctive (e.g. whites, heterosexuals) versus distinctive (e.g. blacks, homosexuals) groups, and that favorable target market effects are stronger for members of distinctive versus non-distinctive groups. Favorable target market effects occur for distinctive viewers because of heightened levels of felt similarity with a source, while favorable target market effects for non-distinctive viewers result from targetedness based on some aspects of the entire configuration of advertisement cues. Unfavorable non-target market effects occur for nondistinctive viewers because of perceived dissimilarity with a source, while unfavorable nontarget market effects occur for distinctive viewers because of perceived exclusion from the intended target market (Aaker, Brumbaugh and Grier 2000) These results suggest that distinctive and non-distinctive viewers differ in the process by which their attitudes are formed in response to targeted advertisements. The next step is to understand what felt similarity and targetedness represent in relation to consumer attitudes towards targeted marketing efforts. According Kelman (1961) these attitudes are formed by the processes of identification and internalization. In the next section the identification theory will be discussed and very briefly the theory of internalization.

2.4.2 Identification Theory Identification occurs when one adopts behavior derived from another person or group because this behavior is associated with a satisfying self-defining relationship (a role relationship that forms a part of the person’s self-image) to this person or group (Kelman 1961). For example, a young man may be persuaded by a sneaker advertisement featuring another young man because he feels that the spokesperson has similar needs, goals and a common lifestyle. In such cases, persuasion may occur because one accepts the message of a 21

The Influence of Consumers’ Ethnic Identity on Advertising Responses

similar person and desires to maintain positive self-esteem in light of their shared traits (Aaker et al. 2000). The Identification theory maintains that people automatically assess their level of similarity with a source during an interaction and make similarity judgments. When viewers perceive that the source possesses characteristics similar to their own, such as race, they begin to infer that the source will also share other characteristics, all of which lead to greater identification and therefore adopt the attitude, or behavior of the source. Thus felt similarity occurs through the process of identification. Therefore, distinctive viewers’ feelings of similarity with a source who shares the distinctive personally-relevant trait (e.g. ethnicity) should lead to identification. However non-distinctive viewers should not feel similar because the trait they share is not as personally relevant or salient (Aaker et al. 2000). Non-distinctive viewers create positive attitudes towards targeted advertisement through the process of internalization. Internalization occurs when one accepts influence because the induced behavior is congruent with his value system. It is the content of the induced behavior that is intrinsically rewarding. One adopts it because he finds it useful for the solution of a problem or because it is congenial to his own orientation, or because it is demanded by his own values. The characteristics of the source do play an important role in internalization, but the crucial dimension here is the source’s credibility. A person may adopt the recommendations of an expert because he finds them relevant to his own problem and congruent with his own values (Kelman 1961). For example, a different young man may be persuaded by the same sneaker advertisement as in previous example, because he feels the spokesperson is knowledgeable about which sneakers are most effective on the basketball court (Aaker et al. 2000). The study “The Effects of Preference Heterogeneity and Source Characteristics on Ad Processing and Judgements about Endorsers” (Feick and Higie 1992) provides insights about consumers’ use of source characteristics in developing attitudes and intentions about the source of the communications and the recommended product. The results suggest that for products and services characterized by higher preference heterogeneity, “the man on the street” testimonials are likely to be effective if the source is perceived to be similar to the target audience. Restaurants, night clubs, plays, movies and other goods and services for which tastes vary could benefit by a clear identification of target audience characteristics and the choice of similar spokesperson. For low preference heterogeneity products and services, the results suggest the use of experienced individuals. For services such as plumbers and dry cleaners the match of the spokesperson to the target audience does

22

The Influence of Consumers’ Ethnic Identity on Advertising Responses

not seem to be as important. For these products, communication of the spokesperson’s experience is more important (Feick and Higie 1992). The authors of the studies “The Effects of Actor’s Race in Commercial Advertising: Review and Extension” (Whittler 1991), “Viewers’ Reaction to Racial Cues in Advertising Stimuli” (Whittler and DiMeo 1991), “Model’s Race: A Peripheral Cue in Advertising Messages?” (Whittler and Spira 2002) and “Ethnic Identification on Adolescents’ Evaluations of Advertisements” (Osei Appaih 2001) use the identification theory to interpret their findings. Corlin Green (1991) uses the Intercultural Accommodation theory as a framework to understand viewers’ attitudes towards advertisements.

2.4.3 Intercultural Accommodation What is meant by Intercultural Accommodation? In the research by Holland and Gentry (1997) the term is used to indicate those efforts on the part of communicators to make themselves more similar to members of another cultural group in order improve communication. In marketing, Intercultural Accommodation could be manifested at various strategic levels. For example, in marketing communications, the domain of accommodation behavior would include such things as using spokespersons of similar ethnic background in advertisements, hiring ethnic salespeople, or using language, music, art, national flags or other cultural symbols as part of the brand or promotion. The term intercultural is used to convey to idea that this communication is occurring across at least two cultures (Holland and Gentry 1997). Thus, Intercultural Accommodation involves communicators of one group borrowing cultural symbols from another group in order to appear more similar, enhance communication, and gain approval (Holland and Gentry 1997). Intercultural Accommodation finds its roots in the Speech Accommodation Theory (SAT) which emerged in the early seventies. SAT has frequently been used to understand communication patterns across cultures (Holland and Gentry 1997). The basis of SAT is that, when people are motivated to seek approval or improve communication, they make an effort to adopt the language or speech patterns of the other party, their speech patterns will converge. When a communicator wishes to distance him/herself from the other or, a pattern of divergence will develop (Holland and Gentry 1997). Koslow, Shamdasani and Touchstone (1994) examined in their study “Exploring Language Effects in Ethnic Advertising: A Sociolinguistic Perspective”, the effects of 23

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Spanish-language usage in advertising in one predominant subculture, United States Hispanics. The study proposes that use of Spanish language in advertising to Hispanic consumers triggers the dynamics of accommodation theory. For accommodation to occur Hispanics must perceive the choice and use of Spanish in the advertisement as an indicator of the advertiser’s respect for the Hispanic culture and desire to break down cultural barriers through reduction of linguistic dissimilarities. Furthermore, the study proposes that more effortful accommodation results in a more favorable response. Thus, increasing the amount of Spanish in an advertisement should increase the positive effects (Koslow 1994). The results of this study demonstrates that Spanish-language usage increases Hispanic consumers’ perceptions of the advertiser’s sensitivity, which in turn have a positive influence on their affect towards the advertisement. Holland and Gentry developed a model of Intercultural Accommodation to understand how consumers react to targeting attempts and how their reactions affect the effectiveness of such attempts. The central concept of this model is that consumer’s affective and cognitive responses to marketer’s targeting efforts will determine the outcomes of that effort. The model demonstrates a variety of antecedents affecting the consumer’s response to the accommodation attempt. However the consumer’s strength of ethnic identification is one of the most important. If people do not identify strongly with their ethnic group, it is unlikely that ethnic group membership will be a predictor of their behavior or response to an advertisement. Ethnic group members who identify strongly with their heritage are likely to have stronger emotional response to the use of their cultural symbols in marketing communication than those who identify less with their heritage. However, this response may be either positive or negative (Holland and Gentry 1997).

Figure 2. A Partial Model of Intercultural Accommodation ANTECEDENTS

CONSUMER RESPONSE TO THE

CONSEQUENCES

ACCOMMODATION ATTEMPT ATTITUDE TOWARDS THE AD

STRENGTH OF ETHNIC

ATTRIBUTIONS

ATTITUDE TOWARDS

IDENTIFICATION

AFFECT

THE BRAND

MESSAGE RECALL

Source: Holland and Gentry (1997) 24

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Holland and Gentry demonstrated in their study (1997) that: attitudes towards the ad and attitudes towards the brand were significantly affected by the use of cultural symbols in advertisements; the differential impact of positive and negative uses of cultural symbols in an advertisement on ad and brand evaluations is an indication of the impact of ad effectiveness; cultural images that evoke negative affect can result in lowered brand evaluations and that advertisements with cultural images that evoke positive affect can significantly improve brand evaluations. Holland and Gentry also demonstrated that consumers with strong ethnic identification, so called strong ethnic identifiers, also react stronger to advertisements with cultural images than weak ethnic identifiers.

Ossei Appiah (2001) demonstrated in his study that black adolescents overall rated black culturally embedded advertisements (advertisements with black cultural images and or cues) more favorably than white culturally embedded advertisements (advertisements with white cultural images and cues). Findings from this study suggest that the presence of black characters in advertisements, and in some cases black cultural cues, significantly influenced black (and white) adolescents’ responses. It was indicated that black adolescents felt more targeted by, and rated more favorably, black characters in cultural embedded advertisements than white characters in culturally embedded advertisements. Although these findings are supportive of the concept of intercultural accommodation, the study failed to demonstrate that the strength of ethnic identification would influence black adolescents’ attitude towards the advertisements. However, high black ethnic identifiers did perceive themselves to be more similar to black characters in black culturally embedded advertisements than black adolescents with low black identity did.

The terms advertising responses and attitudes towards advertisements have often been used discussing the correlation between ethnic identity and advertising or the underlying theories of this correlation. In order to have full understanding of this discussion the next segment discusses the related concepts of advertising effectiveness.

25

The Influence of Consumers’ Ethnic Identity on Advertising Responses

2.5 Advertising Effectiveness Besides providing a definition of advertising effectiveness this closing part of the literature review will discuss advertising response models and techniques for measuring advertising effectiveness.

2.5.1 Definition of Advertising Effectiveness In the literature the term advertising effectiveness and advertising effects are often used interchangeably. According Beerli and Martin Santana (1999) advertising effectiveness refers to the measurement of the results of an advertising campaign or of particular advertisement, which must in turn to be defined in terms of the achievement of the advertising objectives which the advertiser set for the campaign or advertisement. Wright-Isaak, Faber and Horner (Fishman 1999) make a distinction between advertising effectiveness and advertising effects; advertising effectiveness involves assessments of actual advertising campaigns in natural settings and is cumulative over time and affects feelings, attitudes, and behaviors; and advertising effects involves responses to individual ads. Assessments of effectiveness are typically made over longer time spans than measures of effects. Assessments of the effectiveness of advertising campaigns involve multiple exposures to ads and multiple executions within campaigns. In contrast, measuring effects of individual ads involves limited numbers of executions and exposures. Effects are as likely to be evaluated in an experimental setting as in a natural field setting. Effectiveness must be determined within a complex environment where other marketing activities and competitive actions greatly add to the difficulty of assessing advertising's value (Wright- Isaak et al. in Fisbman 1999).

2.5.2 Advertising Response Models The evaluation of the effectiveness of an advertising campaign or effects of an advertisement should be carried out with a view to the advertising objectives established and based on the responses of individuals towards the advertising, analyzed by means of advertising response models. While the literature proposes a number of advertising response models which attempt to explain the behavior of the consumer towards advertising, there is no single universally

26

The Influence of Consumers’ Ethnic Identity on Advertising Responses

accepted model, but rather a number of different models which deal with the same process from different perspectives and which take into account the influence of a great number of variables in the process (Beerli and Santana 1999). After reviewing existing models Beerli and Santana (1999) identified that all of these models do have the same basis; a sequence of three stages which a consumer experience when being exposed to an advertisement and making a purchase decision. These three stages of learning, feeling and doing are related to the three main functions of advertising: to inform, to create attitudes and to induce action. The advertising response models can be grouped into four categories: •

Global models of advertising response: Developing and attempting to explain the process to which individuals are subjected from the moment they are exposed to an advertisement to the moment they take action;

•

Cognitive response models, which only analyze how individuals process information;

•

Attitude models, which explore the attitude component alone;

•

Behavior models which attempt to analyze how individuals act.

In the world of advertising the attitude models together with the global models stand out the most in terms of relevance and application. Attitude models are based on the analysis of the process of attitude formation and change, and of the factors which influence such attitudes (Beerli and Santana 1999). According to Beerli and Santana the major models are those of Fishbein and Azjen (1975), Mitchell (1981), Petty and Cacioppo (1981), Edel and Staelin (1983), Greenwald and Leavitt (1984) and MacKenzie and Lutz (1989). In this research paper an attitude model will be applied in order to answer the research questions.

2.5.3 Techniques for Measuring Advertising Effectiveness Besides the many advertising response models to explain consumers behavior towards advertisement the literature offers, many different techniques for measuring advertisement effectiveness are also being discussed in the literature. Beerli and Santana (1999) defined three categories of techniques based on the three stages which underlie the advertising response models. The three categories are named: cognitive, affective and conative techniques.

27

The Influence of Consumers’ Ethnic Identity on Advertising Responses

•

Cognitive techniques:

Attempt to measure an advertisement’s ability to attract attention, be remembered and communicate the desired message and also to analyze the levels of knowledge and understanding and individual possesses about the advertisement. 1. Awareness measurements assess an individual’s awareness of the existence of a product, brand or company. 2. Memory tests measure the intensity of the impact of a message through the capacity of the individual to recall and/or recognize it. •

Affective techniques:

These techniques measure the attitude which an advertisement is able to generate in the individual, or change of attitude or a reinforcement of an already existing attitude. Attitude towards the advertisement can be defined as a predisposition to respond in a favorable or unfavorable manner to a particular advertising stimulus during a particular exposure (MacKenzie and Lutz 1989). A positive attitude created towards an advertisement is thus a predisposition which increases the probability of the advertised object being purchased, although the purchase decision can still be influenced by a number of other factors. A negative attitude towards an advertisement can be a reason enough for a consumer not to buy the advertised object. The most commonly used affective tests are liking, attitude towards the advertisement, attitude towards the brand and persuasion. 1. The importance of liking in advertising testing is based on the idea that an advertisement must be at the very least pleasing to the individual if it is to be effective. Later in this literature study the importance of liking will be discussed in more detail. 2. The aim of measuring attitude towards an advertisement is to detect the immediate spontaneous reaction of the individual towards an advertisement. 3.

The objective of measuring attitude towards brand is to assess whether or not an individual’s opinion about the brand is positive in comparison with competitors brands.

4. Persuasion test determine the ability of an advertisement to provoke a change in the individual’s attitude towards the advertised brand.

28

The Influence of Consumers’ Ethnic Identity on Advertising Responses

•

Conative techniques:

Conative techniques measure the response behavior of the individual in terms of purchase and repurchase and in terms of predisposition to act in the desired manner. The most important techniques are intention to purchase, inquiry tests and sales variations. 1. Intention to purchase measures the probability that an individual will try or buy the advertised brand or product in the near future as a result of exposure to the advertisement. 2. Inquiry tests evaluate how well the objectives of campaigns aiming to influence individual behavior achieve their established objectives. 3. Sales variations measurements try to measure the sales as a direct result of advertising. The affective techniques will be used in this research in order to measure the impact of one’s ethnic identity on their response or attitude towards advertisement.

2.5.4 Copy Testing While the techniques for measuring advertising effectiveness basically describe what to measure, copy testing is known as a procedure or method on how to measure advertising effectiveness. The procedure of copy testing lays out the methodology and the data collection instruments to be used. Copy testing can be done before or after an ad is released; to ensure that will surface no surprises after the advertisement is released, or to ensure that the advertising is well received. Copy testing is usually conducted as a large scale quantitative research study where the target audience is exposed to the advertising. A copy testing must meet a series of requirements in order to be considered a good measurement instrument. Important requirements are validity, reliability, sensitivity and independency in its applications. Variables which have an influence on advertising effectiveness should be included in a copy testing. Such variables are: •

Individual involvement with the product

•

General attitude towards advertising

•

Degree of credibility given to advertising

(Beerli and Santana 1999)

29

The Influence of Consumers’ Ethnic Identity on Advertising Responses

The Advertising Research Foundation‘s Copy Research Validity Project, hereafter referred to as the CRPV is a comprehensive investigation on the predictive validity of TV Commercial copy-testing measures. In the CRPV 35 advertising pre-test or “copy-test” measures were examined for their ability to predict the sales effectiveness of five pairs of TV commercials with a “winner” and a “loser” in each pair for consumer product brands. The most widely drawn conclusion from the CRPV is that a single measure,” ad likeability”, is the best predictor of an ad’s sales effectiveness (Rossiter and Eagleson 1994; Haley and Baldinger 2000.) This finding in the study, the strong relationship between attitude towards the ad (likeability) and its effects on sales has been intensively discussed by researchers and practitioners. It was generally believed that the function of advertising was to sell and that whether the copy was likable or not had little to do with it (Haley and Baldinger, 2000). Haley and Baldinger discuss in their article The ARF Copy Research Validity (2000) other support for the value of likeability provided by Jim Spaeth of Viewfacts, Inc., Gordon Brown of Millward Brown, Inc., Alex Biel of the Center for Research and Development, and Esther Thorson. Spaeth reported that liking was an excellent predictor of sales generated from five pairs of TIME magazine commercials concerning a variety of subscription offers. Gordon Brown of Millward Brown, Inc. cited a high correlation between likeability and his awareness index which, in turn, he finds to be the strongest correlate of sales increases reflected in tracking studies. He has three years of data and a large number of cases. Alex Biel of the Center for Research and Development summarized an Ogilvy study spanning 57 products in 11 categories that showed that likable ads were twice as persuasive as the average ad. So the case for likeability is documented from other applied sources. Academic researchers have also shown evidence, in the context of research on "Attitude towards the Ad" and its effect on brand attitudes and purchase intentions, that likeability is an important mediator of message effects. Esther Thorson, in a 1990 review of "Consumer Processing of Advertising" for Current Issues & Research in Advertising, sums them up by saying "These studies leave little doubt that the individual's response to the ad itself is a powerful predictor of ad impact." (Haley and Baldinger 2000). However the CRVP report concluded that the measure of likeability is the best predictor for sales effectiveness, the study does not imply that this measure of likeability should be considered as a stand-alone measure of advertisement effectiveness. Other measures such as persuasion and recall are also important measures.

30

The Influence of Consumers’ Ethnic Identity on Advertising Responses

2.6 Summary There are four distinctive approaches to marketing to ethnic consumers: standardized marketing, product adaptation, advertising adaptation and ethnic marketing. In marketing to ethnic consumers and specific in applying the strategy of advertising adaptation one needs to ensure that decisions regarding designing advertising messages and selecting channels, are sensitive to the traditions and feelings of ethnic consumers (Cui 1997). The sensitivity of those traditions and feelings are determined by the ethnic identity of a consumer. “Ethnic identity is a dynamic multidimensional construct that refers to one’s identity, or sense of self as a member of an ethnic group. Ethnic identity is not a fixed categorization, but rather a fluid and dynamic understanding of self and ethnic background. Ethnic identity is constructed and modified as individuals become aware of their ethnicity, with in the large (social cultural) setting”. “An ethnic group claims a common ancestry and of which members share at least a similar culture, race, religion, language, kinship or place of origin” (Phinney 2003). Research has demonstrated that individuals vary in their degree in identification with their ethnic group. Research has also demonstrated that consumer strength of ethnic identity is correlated with ethnic attitudes, consumption (of ethnic-associated) products and response to ethnic cues (e.g. ethnic actors) in advertising. Several theories and or models have been used to explain the correlation ethnic identity and how people deal with ethnic cues in advertising. Three theories are: Distinctiveness theory (McGuire et al. 1978); Kelman’s (1961) Identification Theory and the Intercultural Accommodation model (Holland and Gentry 1997). The central prediction of the Distinctiveness theory is that an individual’s distinctive trait in relation to other people in the environment will be more salient to the individual that more common traits (McGuire et al. 1978). The distinctiveness postulate has been applied to consumer behavior and advertising contexts and supports the notion that numeric minorities respond more favorably to marketing efforts designed to resonate with their distinctive characteristics. Previous research has found that consumer distinctiveness results in a heightened sensitivity to targeting efforts, more identification with and trust of a similar source, and increased favorability towards advertisement and brand (Aaker, Brumbaugh and Grier 2000; Appaih 2001; Deshpande and Stayman 1994; Forehand and Deshpande 2001; Grier and Brumbaugh 1999). The Identification theory maintains that people automatically assess their level of similarity with a source during an interaction and make similarity judgments. When viewers

31

The Influence of Consumers’ Ethnic Identity on Advertising Responses

perceive that the source possesses characteristics similar to their own, such as race, they begin to infer that the source will also share other characteristics, all of which lead to greater identification and therefore adopt the attitude, or behavior of the source. Intercultural Accommodation indicates those efforts on the part of communicators to make themselves more similar to members of another cultural group in order improve communication. Advertising effectiveness refers to the measurement of the results of an advertising campaign or of particular advertisement, which must in turn to be defined in terms of the achievement of the advertising objectives which the advertiser set for the campaign or advertisement. The evaluation of the effectiveness of an advertising campaign or effects of an advertisement should be based on the responses of individuals towards the advertising, analyzed by means of advertising response models. Besides the many advertising response models to explain consumers behavior towards advertisement the literature offers, many different techniques for measuring advertisement effectiveness are also being discussed in the literature. Beerli and Santana (1999) defined three categories of techniques based on the three stages which underlie the advertising response models. The three categories are named: cognitive, affective and conative techniques. While the techniques for measuring advertising effectiveness basically describe what to measure, copy testing is known as a procedure or method on how to measure advertising effectiveness. The procedure of copy testing lays out the methodology and the data collection instruments to be used. The measure of likeability is the best predictor for sales effectiveness.

32

The Influence of Consumers’ Ethnic Identity on Advertising Responses

CHAPTER THREE METHODOLOGY

3.1 Introduction In chapter one the problem statement addresses the issue whether or not findings of previous research - which primarily has been done in the United States - demonstrating the effects of consumers’ ethnic identity on responses to advertisements, are cross culturally robust? In order to address the problem statement and answer the developed research questions, this study investigates the effect of consumers’ ethnic identity on advertising responses in Suriname. Therefore an experiment has been conducted to measure how people with different ethnic identities in Suriname respond to advertisements featuring ethnic models and cues, in terms of: attitude towards the advertisement; attitude towards the model; identification with the model and targetedness. Though the experiment included all ethnic groups in Suriname, the focus of the analyses is on subjects representing the two largest ethnic groups in Suriname: Hindustanis and Creoles. This chapter will elaborate on the hypotheses which have been developed, the design of the experiment, subjects, stimuli material used in the experiment, procedure and the type of analysis used.

3.2 Hypotheses Based on the research questions 2a, 2b, 3a, 3b and 4 defined in chapter one and the overall discussion in chapter two several hypotheses have been developed and will be tested.

The following hypotheses have been derived from the research questions: “Does the ethnicity of the model featured in the advertisement affect viewers’ responses to the advertisement?” and “Does viewers’ level of ethnic identification affect viewers’ responses influenced by the ethnicity of the model featured in the advertisement - to the advertisement?”: H2.1a:

Strong and weak ethnic identifiers have a more favorable attitude towards the advertisement with a model of similar ethnicity than towards an advertisement with a model of dissimilar ethnicity.

33

The Influence of Consumers’ Ethnic Identity on Advertising Responses

H2.1b:

Strong ethnic identifiers have a more favorable attitude towards an advertisement featuring a model with similar ethnicity than weak identifiers

H2.2a:

Strong and weak ethnic identifiers respond more favorable to a model with similar ethnicity featured in an advertisement than to a model with dissimilar ethnicity.

H2.2b:

Strong ethnic identifiers respond more favorable to the model with similar ethnicity than weak ethnic identifiers.

H2.3a:

Strong and weak ethnic identifiers identify themselves more with a model of similar ethnicity featured in the advertisement than with a model of dissimilar ethnicity.

H2.3b:

Strong ethnic identifiers identify themselves more with the model of a similar ethnicity than weak ethnic identifiers.

H2.4a:

Strong and weak ethnic identifiers feel more targeted by an advertisement featuring a model of similar ethnicity than by an advertisement featuring a model of dissimilar ethnicity.

H2.4b:

Strong ethnic identifiers feel more targeted by an advertisement featuring a model with similar ethnicity than weak ethnic identifiers

Based on the next two research questions: “Do cultural embedded advertisements affect viewers’ responses to the advertisement?”, “Does viewers’ level of ethnic identification affect viewers’ responses to high and low cultural embedded advertisements?”, and again on the discussion, the following eight hypotheses have been developed: H3.1a:

Strong and weak ethnic identifiers respond more favorable to a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than to a low culturally embedded advertisement

H3.1b:

Strong ethnic identifiers respond more favorable to high culturally embedded advertisement with ethnic cues related to their ethnicity than weak ethnic identifiers.

34

The Influence of Consumers’ Ethnic Identity on Advertising Responses

H3.2a:

Strong and weak ethnic identifiers respond more favorable to the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than in a low cultural embedded advertisement

H3.2b:

Strong ethnic identifiers respond more favorable to the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than weak ethnic identifiers

H3.3a:

Strong and weak ethnic identifiers identify themselves more with the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than in a low cultural embedded advertisement

H3.3b:

Strong ethnic identifiers identify themselves more with the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than weak ethnic identifiers

H3.4a:

Strong and weak ethnic identifiers feel more targeted by a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than by a low cultural embedded advertisement

H3.4b:

Strong ethnic identifiers feel more targeted by a high culturally embedded advertisement with ethnic cues related to viewer’s ethnicity than weak identifiers

The last research question: “Do multi-cultural cues in the advertisement moderate the effect of a model’s ethnicity on viewers’ responses to the advertisement?” led to the following hypotheses:

H4.1:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement respond more favorable to the advertisement with multi-cultural cues than to an advertisement without these cues.

H4.2:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement respond more favorable to this model in an advertisement with multi-cultural cues than in an advertisement without these cues

35

The Influence of Consumers’ Ethnic Identity on Advertising Responses

H4.3:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement identify themselves more with this model in an advertisement with multi-cultural cues than in an advertisement without these cues

H4.4:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement feel more targeted by an advertisement with multi-cultural cues than by an advertisement without these cues

3.3 Design The study employed a 2 x 6 (“two by six”) between-subjects (ANOVA) design to test the hypotheses for diverse ethnic groups with a focus on the Hindustani and Creole groups. The term “between subjects” reflects the comparisons between different groups of subjects. Each number in the description “2 x 6” represents an independent variable and the value of the number represents the number of levels of that independent variable. For example, 3 x 2 would mean two independent variables, one with three levels and one with two levels. Thus in this experiment there were two independent variables. In describing an ANOVA (Analysis of Variance) design, the term factor is used as a synonym of independent variable1. The two independent variables, or factors in this experiment were: Strength of Ethnic Identity and Cultural Embeddedness. The factor Strength of Ethnic Identity has two levels: strong and weak. The factor Cultural Embeddedness has, corresponding to the six advertisements used in the experiment (see paragraph 3.5), six levels. The six advertisements varied in model’s ethnicity and level of cultural embeddeness. The six levels of the factor Cultural Embeddedness were: low Hindustani cultural embeddedness (advertisement with Hindustani model without ethnic cues), high Hindustani cultural embeddedness (advertisement with Hindustani model and ethnic cues), Hindustani multi-cultural embeddedness (advertisement with Hindustani model and multi-cultural cues), low Creole cultural embeddedness (advertisement with Creole model without ethnic cues), high Creole cultural embeddedness (advertisement with Creole model and ethnic cues) and Creole multicultural embeddedness (advertisement with Creole model and multi-cultural cues). The four dependent variables were: Attitude towards the Advertisement, Attitude towards the Model, Identification with the Model and Targetedness.

1

http://psych.rice.edu/online_stat/chapter12/anova_designs.html.

36

The Influence of Consumers’ Ethnic Identity on Advertising Responses

The design of this experiment is similar to designs employed in previous research as discussed in the literature review.

3.4 Subjects Subjects were selected from the group employees at Fernandes Bottling Company N.V., Surinamese franchise of The Coca-Cola Company. The two main reasons to select subjects from this group were: the group is a relative good representation of the diverse ethnic groups in Surinamese society, and it is assumed that all subjects selected from this group would have a similar attitude towards the brand Coca-Cola (which is used in the stimuli material as described in the next paragraph). Therefore any influence by subjects’ attitude towards the brand on advertisement responses measured in this experiment will remain neutral when comparing results between the subjects. In total 126 subjects participated in the experiment: 49 (39%) Creoles, 47 (37%) Hindustani, 16 (13%) ‘Others’ (Javanese, Maroon, Chinese, Amerindians) and 14 (11%) participants did identify themselves with an ethnic group. 21% female and 79% male. All participants were older than 20 years.

3.5 Stimuli material The stimuli material consisted of six professionally prepared full-color print advertisements. The six advertisements varied in the ethnicity of the model featured in the advertisement and the level of cultural embeddeness. There were three advertisements featuring a Hindustani model of which one advertisement did not feature any ethnic cues (advertisement xh1), one advertisement featuring ethnic cues (advertisement xh2) and one advertisement featuring multi-cultural cues (advertisement xh3) (see appendix 1: figure 1, 2, and 3). The three other advertisements were featuring a Creole model of which also one advertisement was without any ethnic cues (advertisement ck1), one with ethnic cues (advertisement ck2) and one with multi-cultural cues (advertisement ck3) (see appendix 1: figure 4, 5 and 6). The six advertisements were digitally manipulated to vary the ethnicity of the model and the different types of ethnic cues in each advertisement while all other visual features remained constant. For this experiment it was chosen to feature a Hindustani model and a Creole model in the advertisements because they represent the two largest ethnic groups in Suriname. Effort 37

The Influence of Consumers’ Ethnic Identity on Advertising Responses

was made to use equivalent Hindustani and Creole cues in the high culturally embedded versions of the advertisement. Celebrities representing the ethnic group of the model, ethnical clothing and traditional music instruments were used as ethnic cues in the advertisements (see appendix 1: figure 2 and 5). The so called multi-cultural cues in advertisement 3 and 6 (see appendix 1: figure 3 and 6) consisted of typical Surinamese symbols and images All six advertisements were promoting the brand Coca-Cola. The main reason to use this brand in the experiment is that Coca-Cola has itself proven to be adopted by almost all cultures worldwide and even so by all cultures or ethnic groups in Suriname.

3.6 Procedure The experiment was conducted at the site of the organization from which subjects were chosen to participate. The experiment took place in four sessions with twenty-one participants each session, one session with 16 participants and several sessions with smaller groups (due to circumstances it was not achievable to have six sessions with twenty-one participants each session as planned). The participants were told that they would be participating in an advertising survey designed to obtain their opinion on proposed advertisements in order for the advertising agency to improve the look, the style and content of those advertisements. Each participant was given a booklet containing two advertisements and a questionnaire (see appendix 2) which were randomly assigned. The questionnaire consisted of six sets of questions. In the first set the subjects were asked to answer the question: “Please tell us about yourself in your own words”. Subsequently subjects were asked to answer the questions on the two advertisements. The first four questions of this set were corresponding to the first advertisement and the last four questions to the second advertisement. Each set of four questions were immediately answered by the subject after reviewing the corresponding advertisement. After finishing all the questions on the advertisements they were asked to answer the following question: “Please tell us what you are not” The next set of questions concerned the measure of the subject’s ethnic identity followed with a set of questions on age and level of education. Finally a post-experimental question assessed whether participants were aware of the true purpose of the study. Once the questionnaires were completed subjects who did not identify themselves with an ethnic group or were aware of the true purpose of the study were excluded from the research from further analysis. Also incomplete questionnaires were excluded. Of the 126 38

The Influence of Consumers’ Ethnic Identity on Advertising Responses

participants, 9 (7.14%) subjects did not identify themselves with an ethnic group and 0 subjects indicated to be aware of the true purpose of the study. 3 (2.38%) questionnaires were incomplete.

3.7 Measurements As mentioned in the introduction of this chapter this experiment measures how people with different ethnic identities in Suriname respond to advertisements featuring ethnic models and cues. The responses to advertisements measured in this research are: attitude towards the advertisement; attitude towards the model; identification with the model and targetedness. These measurements are types of affective techniques. They measure the attitudes which an advertisement is able to generate in the individual (paragraph 2.5.3), which are reliable indicators for the sales effectiveness of an advertisement (paragraph 2.5.4). In order to define the effect of consumers’ ethnic identity on responses to advertisement it is essential to determine the ethnic group one identifies with and the strength of that ethnic identification. And in order to understand the effect of consumers’ ethnic identity on advertising response it is important to know whether or not ethnicity can be perceived as a distinctive salient trait for Surinamese consumers. Therefore measurements of ethnic identity and ethnic salience were included. The following six paragraphs provide a description of these measurements.

3.7.1 Ethnic salience Ethnic salience was measured by using the spontaneous self-concept measure according Mc Guire (1978). For this measure the subjects were asked to describe aspects of them self in response to the open-ended question. “Please tell us about yourself in your own words” and “Please tell us what you are not”. The percentage identifying their background was used as the measure of ethnic salience.

3.7.2 Ethnic identity The strength of ethnic identification was measured by using The Multigroup Measure of Ethnic Identity (MEIM), developed by Phinney. MEIM was originally published in the article “The Multigroup Ethnic Identity Measure: A new scale for use with adolescents and young 39

The Influence of Consumers’ Ethnic Identity on Advertising Responses

adults from diverse groups” by Phinney (1992)2. It has subsequently been used in dozens of studies and has consistently shown good reliability, typically alphas above .80 across a wide range of ethnic groups and ages. In this research a revised version of the measure will be used. This measure comprises two factors. The first factor is ethnic identity search (a developmental and cognitive component) and the second factor is affirmation, belonging and commitment (an affective component). The items (see appendix 2)1, 2, 4, 8 and 10 of the measure have to do with ethnic identity search and the items 3, 5, 6, 7, 9, 11, 12 concern the factor affirmation, belonging and commitment. Items 13, 14, and 15 are used only for purposes of identification and categorization by ethnicity. Each item was measured using a 4point scale ranging from strongly disagree (1) to strongly agree (4). An ethnic identity scale was developed by averaging the mean scores from each of the 12 items. The median ethnic identity scores for Hindustani, Creole and the group ‘Others’ were respectively 2.81, 2.98 and 3.01. Strong and weak ethnic identifiers scored either above or below the median for their group.

3.7.3 Attitude towards the Advertisement Participants’ attitude towards the advertisement was measured by using nine 7-point semantic

differential

scales:

good/bad,

pleasant/unpleasant,

favorable/unfavorable

like/dislike, effective/not effective, interesting/not interesting, attractive/unattractive and persuasive/not persuasive, positive/negative (MacKenzie and Lutz, 1989; Holbrook and Batra, 1987). These scales have been successfully used in other similar studies and demonstrated strong evidence of reliability (e.g. Deshpande and Stayman, 1994; Green 1999, Appaih 2001). The single index of “Attitude towards the Advertisement” used in the analysis was obtained by averaging the scores to these scales.

3.7.4 Attitude towards the Model Participants rated their likeability of the model on six scales. These six 7-point semantic differential scales were anchored by warm/cold, likable/unlikable, sincere/insincere, friendly/unfriendly, trustworthy /untrustworthy and credible/not credible. The last two scales

2

http://www.calstatela.edu/academic/psych/ftp/meim.doc

40

The Influence of Consumers’ Ethnic Identity on Advertising Responses

mentioned are an extension on the 4 scales measure by Whittler and DiMeo (1991). The scores on the six scales were averaged into a single measure of likeability of the model.

3.7.5 Identification with the Model Participants also indicated on a 7-point Likert scale how strongly they identified with the model in the advertisement by expressing their extent of agreement with the following statements: “ A person whom I want to be like,” “My type of person,” and “A person who speaks for a group of which I am a member” (Whittler and Spira, 2002). The scores were averaged over the three scales.

3.7.6 Targetedness To measure whether or not the subject felt targeted by the advertisement they had to answer three felt targetedness questions. “I feel the advertisement was intended for people like me”, “I don’t believe I was in the target market the company created the advertisement for”, and “The advertiser made that advertisement to appeal people like me”. (Aaker, Brumbaugh and Grier 2000) These felt targetedness items were measured on 7-point Likert scale anchored by disagree completely (1) and agree completely (7).

3.8 Analysis An experiment that includes two independent variables and one dependent variable is appropriately analyzed using a two-way analysis of variance (two-way ANOVA). ANOVA is used to uncover the main and interaction effects of the independent variables on a dependent variable. A main effect is the direct effect of an independent variable on the dependent variable. An interaction effect is de joint effect of two or more independent variables on the dependent variable. The key statistic in ANOVA is the F-test of difference of group means, testing if the means of the groups formed by values of the independent variable (or combinations of values for multiple independent variables) are different enough not to have occurred by chance. If the group means do not differ significantly then it is inferred that the independent variables did not have an effect on the dependent variable. If the F-test shows that overall the independent variables are related to the dependent variable then Multiple comparison tests of 41

The Influence of Consumers’ Ethnic Identity on Advertising Responses

significance are used to explore just which group means of the independent variables differ significantly from the others3. To test the hypotheses for the different ethnic groups a series of two-way ANOVA, Multiple comparisons tests have been conducted for each of the four dependent variables. The results of the experiment are presented and discussed in the following chapter.

3.9 Summary This study investigates the effect of consumers’ ethnic identity on advertising responses in Suriname. Therefore an experiment has been conducted to measure how people with different ethnic identities in Suriname respond to advertisements featuring ethnic models and cues, in terms of: attitude towards the advertisement; attitude towards the model; identification with the model and targetedness. The study employed a 2 x 6 (“two by six”) between-subjects (ANOVA) design. The two independent variables, or factors in this experiment were: Strength of Ethnic Identity and Cultural Embeddedness with respectively two and six levels. The four dependent variables were: Attitude towards the Advertisement, Attitude towards the Model, Identification with the Model and Targetedness. Also measures of ethnic salience and ethnic identity were included in the experiment. To uncover the main and interaction effects of the independent variables on a dependent variable and to test the hypotheses for the different ethnic groups a series of two-way ANOVA and Multiple comparisons tests have been conducted for each of the four dependent variables.

3

http://www2.chass.ncsu.edu/garson/pa765/anova.htm

42

The Influence of Consumers’ Ethnic Identity on Advertising Responses

CHAPTER FOUR RESULTS

4.1 Introduction In this chapter the results of the experiment will be presented and discussed. First the results of the measure on the salience of ethnicity will be presented, followed by the advertising responses of the Creole group and the Hindustani group to the six different advertisements. In order to analyze the impact of the factors Strength of Ethnic Identity and Cultural Embeddedness on the advertising responses and to test the hypotheses series of twoway ANOVA and Multiple comparisons tests have been performed.

4.2 Salience of Ethnicity From the 112 subject who identified themselves with an ethnic group 17 subjects mentioned their ethnicity when answering the question “Tell us about yourself” or “Tell us what you are not”. Ethnicity is a salient trait for 12.77% (6 out of 47) of the Hindustani subjects, 14.29% (7 out of 49) of the Creoles and for 25% (4 out of 16) of the group consisting of subjects identifying with the Javanese, Maroon, Chinese or Amerindian ethnicity.

4.3 Advertising Responses In the next section the interaction and main effects of the independent variables: Cultural Embeddedness and Strength of Ethnic Identity on the dependent variables: Attitude towards the advertisement, Attitude towards the Model, Identification with the Model and Targetedness will be discussed for the Creole group and for the Hindustani group. All results of the series of two-way ANOVA for the two ethnic groups are available in appendices 3 and 4. The means for all independent and dependent variables for the two groups are presented in table 1 and 2.

43

The Influence of Consumers’ Ethnic Identity on Advertising Responses

4.3.1 Interaction and Main Effects In the case of the Creole group the ANOVA demonstrates an interaction effect of the two independent variables on the dependent variables Identification with the Model (F5,84 = 2.142, p< 0.1) and Targetedness (F5,84 =2.323, p< 0.1) (see appendix 3). These interactions effects indicate that the advertising responses Identification with the Model and Targetedness are influenced by the joint effect of the cultural embeddedness of the advertisement and the strength of ethnic identity of the viewer. The series of ANOVA demonstrate significant main effects for the dependent variable Cultural Embeddedness on each of the four dependent variables: Attitude towards the Advertisement (F5,86 = 7.910, p< 0.05), Attitude towards the Model (F5,86 = 4.601, p< 0.05), Identification with the Model (F5,86 = 11.304, p< 0.05) and Targetedness (F5,86 = 12.071, p< 0.05) (see appendix 3). The analysis does not demonstrate a main effect for the dependent variable Strength of Ethnic Identity on one of the four dependent variables. For the Hindustani group the results of the ANOVA reveals interaction effects for: Attitude towards the Advertisement (F5,82 = 4.086, p< 0.05), Attitude towards the Model (F5,82 = 2.582, p< 0.05), Identification with the Model (F5,82 = 2.864, p< 0.05) and Targetedness (F5,82 = 5.027, p< 0.05) (see appendix 4). The analyses reveal significant main effects for the dependent variable Cultural Embeddedness on all four dependent variables: Attitude towards the Advertisement (F5,82 = 4.891, p< 0.05), Attitude towards the Model (F5,82 = 4.277, p< 0.05), Identification with the Model (F5,82 = 6.270, p< 0.05) and Targetedness (F5,82 = 17.461, p< 0.05) (see appendix 4). The analyses do not reveal a main effect for the dependent variable Strength of Ethnic Identity on one of the four dependent variables.

4.3.2 Hypotheses Testing As mentioned in chapter 3 (paragraph 3.8) and in the introduction of this chapter series of two-way ANOVA and Multiple comparisons tests have been conducted to test the hypotheses. Only in case of a significant main effect (see previous paragraph) further investigation through Multiple comparisons tests has been performed in order to determine whether or not the hypothesis is supported.

44

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Table 1. Mean Responses and Standard Deviations for Creole Identifiers

Total

H2.1a:

Identifiers Identifiers

Weak Creole

Strong Creole

AD

Attitude

Attitude

Identification

towards the

towards

with the Model

Advertisement

the Model

M

SD

M

SD

Targetedness

M

SD

M

SD

ck1

5.5271

.9505

5.4757

.9824

4.3343

1.7222

5.2857

.9512

ck2

5.8567

1.0342

6.2533

.4883

5.4833

1.1927

5.8056

1.3924

ck3

5.8567

.9806

6.0722

.9705

5.2475

1.1652

4.8322

1.2767

xh1

4.6667

1.7512

5.0833

1.0685

1.9340

.9256

2.5017

.5889

xh2

4.0570

.7845

4.5170

1.1882

3.1670

1.5089

3.7660

1.4231

xh3

5.3089

1.2166

4.8344

1.6566

3.9622

1.1599

4.4078

1.6565

ck1

5.5533

.8557

5.4256

1.0138

4.4822

.8188

4.7789

1.1116

ck2

5.1050

1.2234

6.0988

.6532

5.0763

1.1988

4.5838

1.8152

ck3

5.7343

.9141

5.1614

.5637

4.1429

.6624

5.3814

.7561

xh1

4.3071

1.3911

5.0714

1.0793

3.6229

1.0432

3.1914

1.3204

xh2

3.5078

1.3005

4.9244

1.2053

2.6289

1.0976

2.1300

1.1666

xh3

4.8413

1.0683

5.2300

.6569

3.4575

1.2213

4.3975

1.1302

ck1

5.5419

.8672

5.4475

.9669

4.4175

1.2448

5.0006

1.0184

ck2

5.5029

1.1573

6.1806

.5587

5.2918

1.1764

5.2306

1.6096

ck3

5.8031

.9225

5.6738

.92044

4.7320

1.0919

5.0725

1.0850

xh1

4.4731

1.5100

5.0769

1.0287

2.9192

1.2890

2.8731

1.0698

xh2

3.7968

1.0672

4.7100

1.1812

2.9121

1.3229

2.9911

1.5238

xh3

5.0888

1.0683

5.0206

1.2657

3.7247

1.1801

4.4029

1.3896

Strong and weak ethnic identifiers have a more favorable attitude towards the advertisement with a model of similar ethnicity than towards an advertisement with a model of dissimilar ethnicity.

The hypothesis is not supported in the case of the Creole group. Although the results demonstrate a significant main effect for cultural embeddedness on the dependent variable Attitude towards the Advertisement (F5,86 = 7.910, p< 0.05) and strong and weak Creole identifiers have a more favorable attitude towards the advertisement featuring a model of similar ethnicity (M= 5.5419) than towards the advertisement featuring a model of dissimilar 45

The Influence of Consumers’ Ethnic Identity on Advertising Responses

ethnicity (M= 4.4731, p > 0.05), the difference between means is not significant according the Multiple comparisons (see table 1 and appendix 3). The hypothesis is also not supported in the case of the Hindustani group. According to the Multiple comparisons Hindustani identifiers overall do not have a significant more favorable attitude towards the advertisement with a Hindustani model (M= 5.1411) than towards the advertisement with a Creole model (M= 4.4180, p < 0.05) (see table 2 and appendix 4).

Table 2. Mean Responses and Standard Deviations for Hindustani Identifiers

Identifiers

Hindustani

H2.1b:

Identifiers

Hindustani

Total

Weak

Strong

AD

Attitude

Attitude

Identification

towards the

towards

with the Model

Advertisement

the Model

M

SD

M

SD

M

SD

Targetedness

M

SD

ck1

4.5713

.9323

4.7288

.9529

3.4588

.3547

2.7500

.6609

ck2

4.5633

.1.0901

4.6856

1.2654

4.0933

1.1878

4.1656

ck3

5.2325

.7159

5.3500

.7349

3.6650

2.4738

4.8763

xh1

6.0467

.4483

5.6289

.8999

4.9956

..7573

5.1856

xh2

5.6733

.8755

5.7450

.5434

5.7167

.7080

5.9450

.9748

xh3

5.2580

.7073

4.8500

.4877

4.4510

1.2855

4.8330

.8492

ck1

4.2429

.3379

4.5743

.4173

3.8571

.6344

3.5214

.6635

ck2

4.5500

.4474

4.5850

.6101

4.4800

.3390

4.0088

.5257

ck3

5.6329

1.4201

5.2614

.4492

4.2357

1.0145

5.5171

1.1996

xh1

4.3260

.9535

5.3460

1.1068

4.3500

.5250

3.9500

.7859

xh2

5.1140

.8755

5.2340

.8959

4.9960

.7290

5.3020

.6912

xh3

5.8986

.7073

6.1900

.6050

6.2371

.7642

6.1414

1.0126

ck1

4.4180

.7157

4.6567

.7314

3.6447

.5270

3.1100

.75212

ck2

4.5571

.8257

4.6382

.9629

4.2753

.8918

4.0918

.8264

1.0540 1.0368 .9577

ck3

5.4193

1.0602

5.3087

.5988

3.9313

1.8941

5.1753

1.1241

xh1

5.1411

1.1502

5.4800

.9968

4.6558

.7088

4.5353

1.0549

xh2

5.4191

1.0638

5.5127

.7348

4.9960

.7777

5.6527

.91373

xh3

5.5216

1.3706

5.4018

.8562

6.2371

1.4034

5.3718

1.0126

Strong ethnic identifiers have a more favorable attitude towards an advertisement featuring a model with similar ethnicity than weak identifiers

46

The Influence of Consumers’ Ethnic Identity on Advertising Responses

There is no significant main effect for the independent variable Strength of Ethnic Identity found (see appendix 3). A closer look at the means confirms that the difference between the attitudes towards advertisement of strong and weak Creole identifiers is not significant (see table 1). Therefore the hypothesis is not supported in the case of the Creole group. Also in the case of the Hindustani group the hypothesis is not supported. As discussed in paragraph 4.2, there is no main effect for the independent variable Strength of Ethnic Identity on the dependent variable Attitude towards the Advertisement found. H2.2a:

Strong and weak ethnic identifiers respond more favorable to a model with similar ethnicity featured in an advertisement than to a model with dissimilar ethnicity.

The ANOVA for the dependent variable Attitude towards the Model does demonstrate a main effect for cultural embeddedness (F5,86 = 4.379, p< 0.05) for the Creole group. This main effect is not established by differences of mean responses between the low Creole culturally embedded advertisement and the low Hindustani culturally embedded advertisement, because the results do not show a significant difference between the attitude of strong and weak Creole identifiers towards the Creole model (M= 5.4475) and the attitude towards the Hindustani model (M= 5.0769) (see table 1 and appendix 3). Therefore the hypothesis is not supported in the case of the Creole group. For the Hindustani group the ANOVA demonstrates a main effect for cultural embeddedness on the dependent variable Attitude towards the Model (F5,82 = 4.277, p< 0.05). The Multiple comparisons does not demonstrate that the difference between the attitude of Hindustani identifiers towards the Hindustani model (M= 5.4800) and the attitude towards the Creole model (M= 4.6567, p< 0.05) is significant (see table 2 and appendix 4). The hypothesis is not supported in case of the Hindustani group. H2.2b:

Strong ethnic identifiers respond more favorable to the model with similar ethnicity than weak ethnic identifiers.

The analysis demonstrates that there is no significant main effect for the independent variable Strength of Ethnic Identity on Attitude towards the Model. Again a closer look at the means confirms that the difference between the attitudes of strong Creole identifiers and

47

The Influence of Consumers’ Ethnic Identity on Advertising Responses

weak Creole identifiers towards the model is not significant. Thus the hypothesis is not supported in the case of the Creole group. Also in the case of the Hindustani group the hypothesis is not supported. As discussed in paragraph 4.2, there is no main effect for the independent variable Strength of Ethnic Identity on the dependent variable Attitude towards the Model found. H2.3a:

Strong and weak ethnic identifiers identify themselves more with a model of similar ethnicity featured in the advertisement than with a model of dissimilar ethnicity.

A significant main effect for cultural embeddedness on the variable Identification with the Model is found (F5,84 = 11.304 p 0.05) (see table 2 and appendix 4).

49

The Influence of Consumers’ Ethnic Identity on Advertising Responses

H3.1b:

Strong ethnic identifiers respond more favorable to high culturally embedded advertisement with ethnic cues related to their ethnicity than weak ethnic identifiers.

The results do not demonstrate a main effect for the strength of ethnic identity on Attitude towards the Advertisement for the two ethnic groups. This indicates that the hypothesis is not supported in the case of both ethnic groups. H3.2a:

Strong and weak ethnic identifiers respond more favorable to the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than in a low culturally embedded advertisement

The hypothesis is not supported in the case of the Creole group. As discussed earlier a main effect for cultural embeddedness on Attitude towards the Model is found. Despite of the main effect the Multiple comparisons test reveal that Creoles overall do not respond significantly more favorable to the model featured in a high Creole culturally embedded advertisements (M= 6.1806) than to the model in a low Creole culturally embedded advertisement (M= 5.4475) (see table 1 and appendix 3). In the case of the Hindustani group the hypothesis is not supported. The main effect for cultural embeddedness for Attitude towards the Model as discussed earlier, is not a result from the difference between the attitude towards the model in a high Hindustani culturally embedded advertisement (M= 5.5127) and the attitude towards the model in a low Hindustani culturally embedded advertisement (M= 5.4800) (see table 2 and appendix 4).

H3.2b:

Strong ethnic identifiers respond more favorable to the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than weak ethnic identifiers

The results do not demonstrate a main effect for the strength of ethnic identity on Attitude towards the Model for the two ethnic groups. This indicates that the hypothesis is not supported in the case of both ethnic groups. H3.3a:

Strong and weak ethnic identifiers identify themselves more with the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than in a low cultural embedded advertisement

50

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Further investigation of the main effect for cultural embeddedness on Identification with the Model reveals that Creole identifiers overall do not identify themselves significantly more with the model in a high Creole culturally embedded advertisement (M=5.2918) than in a low Creole culturally embedded advertisement (M = 4.4175) (see table 1 and appendix 3). Hence, the hypothesis is not supported in the case of the Creole group. Also in the case of the Hindustani group the hypothesis is not supported. The Multiple comparisons test demonstrates that strong and weak Hindustani identifiers do not identify themselves significantly more with the model in a high Hindustani culturally embedded advertisement (M= 5.3891) than in a low Hindustani culturally embedded advertisement (M= 4.6558, p< 0.05) (see table 2 and appendix 4). H3.3b:

Strong ethnic identifiers identify themselves more with the model featured in a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than weak ethnic identifiers

The results do not demonstrate a main effect for the strength of ethnic identity on Identification with the Model for the two ethnic groups. This indicates that the hypothesis is not supported in the case of both ethnic groups. H3.4a:

Strong and weak ethnic identifiers feel more targeted by a high culturally embedded advertisement with ethnic cues related to viewers’ ethnicity than by a low cultural embedded advertisement

The hypothesis is not supported in the case of the Creole group. Although a main effect for cultural embeddedness on Targetedness is found, the results fail to support of the hypothesis. The results show that strong and weak Creole identifiers do not feel significantly more targeted by the high Creole culturally embedded advertisement (M= 5.2306) than by the low Creole culturally embedded advertisement (M= 5.0006) (see table 1 and appendix 3). In the case of the Hindustani group the hypothesis is supported. Besides the main effect the results of the tests show a significant difference between the means of Targetedness for the high Hindustani culturally embedded advertisement (M= 5.6523) and low Hindustani culturally embedded advertisement (M= 4.5353, p< 0.05) (see table 2 and appendix 4). The significant difference in means indicates that Hindustani overall feel more targeted by the

51

The Influence of Consumers’ Ethnic Identity on Advertising Responses

advertisement high in cultural embeddedness than by the advertisement low in cultural embeddedness. H3.4b:

Strong ethnic identifiers feel more targeted by a high culturally embedded advertisement with ethnic cues related to viewer’s ethnicity than weak identifiers

The results do not demonstrate a main effect for the strength of ethnic identity on Targetedness for the two ethnic groups. This indicates that the hypothesis is not supported in the case of both ethnic groups. H4.1:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement respond more favorable to the advertisement with multi-cultural cues than to an advertisement without these cues.

This hypothesis is partially supported in the case of the Creole group. Exploration of the means of the low culturally embedded advertisement (M= 4.4731) and the multi-cultural embedded advertisement (M= 5.0888) with models of dissimilar ethnicity indicates that the results fail to support the hypothesis, but investigation of the means for the high culturally embedded advertisement (M= 3.7968) and again the multi-cultural embedded advertisement with the Hindustani model (p< 0.05) indicates that the hypothesis is supported (see table 1 and appendix 3). In the case of the Hindustani group the hypothesis is fully supported. Multiple comparisons demonstrates that strong and weak Hindustani identifiers have a more favorable attitude towards the multi-cultural embedded advertisement with a Creole model (M= 5.4193) than towards the low culturally embedded (M= 4.4180) or high culturally embedded advertisement with a Creole model (M= 4.5571, p< 0.05) (see table 2 and appendix 4). H4.2:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement respond more favorable to this model in an advertisement with multi-cultural cues than in an advertisement without these cues.

The hypothesis is not supported in the case of the Creole group. Creoles overall seem not to have a more favorable attitude towards the model of dissimilar ethnicity being featured in an advertisement with multi-cultural cues (M= 5.0206) than in an advertisement without these cues (M= 5.0769, p> 0.05) (see table 1 and appendix 3). 52

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Also In the case of the Hindustani group the hypothesis is not supported. The results indicate that Hindustanis overall do not have a significantly more favorable attitude towards the Creole model being featured in a multi-cultural embedded advertisement (M= 5.3087) than in low Creole culturally embedded advertisement (M= 4.6567) or high Creole culturally embedded advertisement (M= 4.6382 p> 0.05) (see table 2 and appendix 4).

H4.3:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement identify themselves more with this model in an advertisement with multi-cultural cues than in an advertisement without these cues.

The hypothesis is not supported in the case of the Creole group. According to the Multiple comparisons test Creoles overall do not identify themselves significantly more with the model of a dissimilar ethnicity in the multi-cultural embedded advertisement (M= 3.7247) than with the model in the high (M= 2.9121) or low culturally embedded advertisement (M= 2.9192, p> 0.05) (see table 1 and appendix 3). Strong and weak Hindustani identifiers do not identify themselves more with the Creole model in a multi-cultural embedded advertisement (M= 3.9313 than with the model in the high (M= 4.2753) or low culturally embedded advertisement (M= 3.6447, p> 0.05). Hence, in the case of the Hindustani group the hypothesis is not supported (see table 2 and appendix 4).

H4.4:

Strong and weak ethnic identifiers with dissimilar ethnicity as the model featured in the advertisement feel more targeted by an advertisement with multi-cultural cues than by an advertisement without these cues

The analysis indicates that strong and weak Creole identifiers feel more targeted by the multi-cultural embedded advertisement featuring a model with dissimilar ethnicity (M= 4.4029) than by the advertisement with the same model without these multicultural cues (low Hindustani culturally embedded advertisement M= 2.8731; high Hindustani culturally embedded advertisement M= 2.911, p< 0.05) (see table 1 and appendix 3). Hence, the hypothesis is supported in the case of the Creole group. The hypothesis is also supported in the case of the Hindustani group. The multicomparisons test indicates that Hindustani overall feel more targeted by the Creole multicultural embedded advertisement (M= 5.1753) than by the low and high Creole culturally embedded advertisements (M= 3.1100; M= 4.0918, p< 0.05) (see table 1 and appendix 4).

53

The Influence of Consumers’ Ethnic Identity on Advertising Responses

4.4 Summary Ethnicity appeared to be a salient trait for 12.77% of the Hindustani subjects, 14.29% of the Creoles and for 25% of the group consisting of subjects identifying with the Javanese, Maroon, Chinese or Amerindian ethnicity. For both groups (Hindustanis and Creoles) an interaction effect is found on the dependent variables Identification with the Model and Targetedness. In the case of the Hindustani group the results also reveal an interaction effect on Attitude towards the Advertisement and Attitude towards the Model. Only for the independent variable Cultural Embeddedness main effects on all four types of advertising responses for both groups are found. The results do not reveal main effects for the independent variable Strength of Ethnic Identity. In the case of the Creole group as well as in the case of the Hindustani group the hypotheses with regard to Attitude towards the Advertisement and multi-cultural cues in the advertisements are supported. Also in both cases the hypotheses on the influence of the ethnicity of the model in relation to the advertising response Identification with the Model are supported by the results. The results also supported the hypotheses on Targetedness in relation to the ethnicity of the model and multi-cultural cues for both groups. In the case of the Hindustani group is the hypotheses with regard to felt Targetedness and high culturally embedded advertisements also supported

54

The Influence of Consumers’ Ethnic Identity on Advertising Responses

CHAPTER FIVE CONCLUSIONS AND RECOMMENDATIONS

5.1 Introduction The objectives of this study are: to address the issue - which was raised in the problem statement - “Are findings of previous research robust cross culturally?”, and to provide recommendations for effective advertisements in Suriname as a multicultural marketplace. To be able to meet the objectives of this study research questions and hypotheses were developed, an experiment was conducted, results were analyzed and hypotheses were tested, which will lead to some conclusions to address the research questions. This chapter will first present the conclusions and based on these conclusions the question raised in the problem statement will be addressed. Thereafter some limitations of the study will be discussed followed by recommendations on effective advertisements. The chapter will be completed with recommendations for future research.

5.2 Conclusions on The Influence of Ethnic Identity on Advertising Responses (in Suriname) Ethnicity appears to be a salient trait for Surinamese consumers. The findings support the prediction of the Distinctiveness theory that a numerical minority group will be more likely to deem their ethnic identity salient than members of numerical majority group would (Mc Guire et al 1978). The next paragraph will discuss this in more detail. The findings that Creoles as well as Hindustanis identify themselves more with the model of own ethnicity and feel more targeted by the advertisement featuring a model of similar ethnicity than by the advertisement featuring a model of dissimilar ethnicity supports the expectation that the ethnicity of the model featured in the advertisement does affect viewers’ responses to the advertisement. These findings also support the Identification theory: when viewers perceive that the source, in this case the model, possesses characteristics similar to their own, such as ethnicity, it will lead to greater identification (Kelman 1961). The fact that one feels more targeted by an advertisement featuring a model of similar ethnicity supports the theory on Intercultural accommodation maintaining it improves communication when the communicators, in this case the advertisers, make

55

The Influence of Consumers’ Ethnic Identity on Advertising Responses

themselves more similar to members of a cultural group (Holland and Gentry 1997).When a person feels targeted by advertisement it is likely that one at least takes the time to read or view the advertisement more closely. Different than expected, neither Creoles nor Hindustanis have more favorable advertising responses to the high culturally embedded advertisements than to the low culturally embedded advertisements. With the exception that Hindustanis feel more targeted by a high Hindustani culturally embedded advertisement. The main effect for cultural embeddedness of the advertisements on the advertising responses for the Creoles as well as for the Hindustanis is mainly a result of differences in means of responses to advertisements which were not further analyzed for addressing the purpose of this study. Creoles as well as Hindustanis feel more targeted and have a more favorable attitude towards the multi-cultural embedded advertisement featuring a model of dissimilar ethnicity than towards an advertisement without these multi-cultural cues. Hence, multi-cultural advertising cues do moderate the effect of a model’s ethnicity on viewers’ responses to the advertisement to a certain extent and therefore an advertiser may be able to reach multiple groups from different cultural or ethnic backgrounds. The finding that one has a more favorable attitude towards the multi-culturally embedded advertisement featuring a model of dissimilar ethnicity than towards an advertisement without these multi-cultural cues suggests that a person is more likely to be effected by the advertisement to purchase the advertised product than the person would be in case of an advertisement without these cues. This study failed to prove that the level of ethnic identification does influence viewers’ responses to advertisements directly; no main effect for the variable Strength of Ethnic Identity was found. Nevertheless, the interaction effects do indicate a joint effect of strength of ethnic identity and cultural embeddedness.

5.3 Conclusions on Cross Culturally Robustness of Previous Research As pointed out in chapter one (paragraph 1.3) nearly all research on the relation between ethnicity and consumers’ advertising responses has been done in the United States. This raises the issue whether or not findings of these researches - demonstrating the effects of consumers’ ethnic identity on responses to advertisements - are cross culturally robust since the ethnic, racial and cultural composition of the population of Suriname is different from the population of the United States. The population of Suriname does not, like the population of the United States, consist of clearly distinct majority (whites) and minority groups (blacks, 56

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Hispanics). As described in chapter one (paragraph 1.2) the population of Suriname is much ethnically diverse; it consists of many ethnic groups of which the two largest groups are numerically very close. Since it is difficult to speak in terms of majority and minority groups for Suriname, findings such as: “members of minority (versus majority) groups were more likely than majority groups to have their ethnicity salient” (Deshpande and Stayman 1994) are supported by this study to the extent that the results indicate that ethnicity for the numerically largest ethnic group in Suriname (Hindustanis) is less salient than for smaller ethnic groups. In other words: the smaller (numerically) the group is, the more salient ethnicity is likely to be. Furthermore, this study found support that findings such as: “members of minority groups find an advertisement spokesperson of their own ethnicity to be more trustworthy” (Deshpande and Stayman 1994), “whites had less favorable attitudes towards the advertisements when the advertisements featured black actors” (Whittler and DiMeo), “Black adolescents overall were more likely to believe that they were the intended audience of black culturally embedded advertisement.” (Appaih 2000) are robust cross culturally. This study shows that ethnicity of the models featured in the advertisement affects Surinamese, more specific Creole and Hindustani, viewers’ responses to advertisements in a similar way. However, this study did not find evidence for the cross culturally robustness of findings with regard to consumers’ strength of ethnic identity such as: “Strong Hispanic identifiers are more affected by advertisements (especially those targeted on Hispanics) than weak identifiers.”(Donthu and Cherian 1994), “Strong black identifiers have more positive evaluations of advertisements that feature African-American in positions of dominance than weak black identifiers.” (Corliss Green (1999), “Blacks’ differential responses to advertisements featuring black and white models are moderated by their identification with black culture.” (Whittler and Spira),”Black adolescents who have a strong black ethnic identity perceive themselves more similar to and identify more strongly with black character advertisements than black adolescents with weaker ethnic identities do.” (Appaih 2000). Then again, the results did show in several cases of advertising responses an interaction effect of strength of ethnic identity and cultural embeddedness (see chapter four). The presence of these interaction effects does indicate the joint effect of the strength of ethnic identity and cultural embeddedness on advertising responses.

57

The Influence of Consumers’ Ethnic Identity on Advertising Responses

5.4 Limitations of the Study The fact that the experiment was conducted among employees from one company puts constraints to generalizing the results and conclusions to the Surinamese population, specially the results and conclusions on the salience of ethnicity as a distinctive trait. Since the subjects of the experiment knew each other it is likely that one’s ethnicity was less salient during the context of the experiment than in a situation the subjects do not know each other. Future research should investigate the salience of ethnicity as a distinctive trait in a different context and among different groups of subjects. One explanation why the level of ethnic identity did not have a significant effect may be due to the cultural cues used in the advertisements. The culturally embedded advertisements may not have been strong enough to sufficiently elicit the effect of ethnic identity. Future research should investigate how consumers evaluate more powerful ethnic-specific culturally embedded advertisements based on their ethnic identity. Future research should also consider conducting a study with a larger sample. Given test statistics are dependent on sample size; studies like this one may fail to find statistically significant results even when they exist because of a small sample size. For example, subjects overall respond more favorable or stronger to high culturally embedded advertisement than to low culturally embedded advertisements, but the results do not show a significant difference in means. Also strong and weak identifiers do demonstrate difference in response, as expected, but the group means apparently do not differ significant.

5.5 Recommendations for Effective Advertisements The findings of this study should be of interest for practitioners in de field of marketing and advertising because it discloses answers to questions which often arise in the process of designing an advertisement. Since the ethnicity of the model featured in the advertisement influences advertising responses of the viewer in a way that one seems to have more favorable or stronger feelings with regard to the advertisement featuring a model of similar ethnicity, one should carefully consider the ethnicity of the model to be featured in the advertisement. When designing an advertisement targeting a specific ethnic group, the advertisement should feature model(s) with similar ethnicity as the target group in order to make certain that the target group at least feels targeted. The use of ethnic specific advertising cues does not

58

The Influence of Consumers’ Ethnic Identity on Advertising Responses

significantly impact advertising responses favorably, but it is recommended to be cautious in case of using such cues when the advertisement is not targeting a specific ethnic group, because it might alienate people who do not relate to those cues. If the advertisement is not designed to target a particular ethnic group the advertiser could use a model representing a Hindustani or a Creole model - as they represent one of the two largest groups – to ensure that at least a large part of the population feels targeted or has favorable feelings towards the advertisement, but to ensure that the advertisements elicit favorable response across different ethnic groups one should combine advertising cues which will be favorably being viewed by people of different ethnic background.

5.6 Recommendations for Future Research It should be clear that this study is only a very small part of a much more complex study and probably raises more questions than currently have been answered. Interesting research questions for future research which are a direct result of the findings of this study are: Which other cues, besides the ones used in this study, will elicit (more) favorable feelings towards the advertisement with people of different cultural background?; What is the effect of featuring models of mixed ethnicity in advertisements?; What is the impact of using several models with different ethnicity in an advertisement?; When do people feel alienated by an advertisement? Designing an advertisement is only a small part of developing an advertising campaign. An interesting area for future research in Suriname on advertising in relation to ethnicity is the area of media and media placement. For example, the medium in which an advertisement is placed may have impact on viewers’ responses to the advertisement. Placing an advertisement in a newspaper or magazine targeting particularly an ethnic group may lead to different levels and types of ethnic identity and thus advertising effectiveness than placing the very same advertisement in a mainstream medium targeting a broader audience (e.g. De Ware Tijd). These effects may also occur for targeted radio, television and other media, whereby the medium provides cues that prime or make salient ethnic identities, and the advertisements embedded therein can reap the advantages or the disadvantages of these effects (Grier and Brumbaugh 2002). As described in paragraph 1.2.2, there are many television stations, radio stations and newspapers in Suriname of which many are targeting ethnically. It would be of interest of many advertisers or other practitioners in the field of advertising to find out about

59

The Influence of Consumers’ Ethnic Identity on Advertising Responses

these effects in order to increase the effectiveness of their advertisements and advertising campaigns.

60

The Influence of Consumers’ Ethnic Identity on Advertising Responses

LIST OF REFERENCES Aaker, Jennifer L., Brumbaugh, Anne M., Grier, Sonya A., “Non-target Markets and Viewer Distinctiveness: The Impact of Target Marketing on Advertising Attitudes”, Journal of Consumer Psychology, 9, 3 (2000): 124-139.

Appiah, Osei, “Ethnic Identification on Adolescents’ Evaluations of Advertisements, Journal of Advertising Research, (September/October), 41, 5 (2001): 7-22.

Cui, Geng, “Marketing Strategies in a Multi-ethnic Environment”, Journal of Marketing Theory and Practice, (winter), 5, 1 (1997): 122-134.

Deshpande, Rohit, Stayman, Douglas M., “A tale of two cities: Distinctiveness theory and advertising effectiveness”, Journal of Marketing Research, (February), 31, 1 (1994): 57-64.

Dimofte, Claudiu V., Forehand, Mark R., Deshpande, Rohit, “Ad Schema as Elicitor of Ethnic Self-Awareness and Differential Advertising Response”, Forthcoming in Journal of Advertising.

Donthu, Naveen, Cherian, Joseph, “Impact of Strength of Ethnic Identification on Hispanic Shopping Behavior”, Journal of Retailing, 70, 4 (1994): 383-393.

Feick, Lawrence and Robin A. Higie, “The Effects of Preference Heterogeneity and Source Characteristics on Ad Processing and Judgements about Endorsers”, Journal of Advertising, XXI, 2 (1992): 9-22.

Fishman, Joshua A., ed., Handbook of Language & Ethnic Identity (New York: Oxford University Press, 1999)

Forehand, Mark R., Deshpande, Rohit, “What We See Makes Us Who We Are: Priming Ethnic Self-Awareness and Advertising Response”, Journal of Marketing Research, (August), 28, 3 (2001): 336-348.

61

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Green, Corliss L., “Ethnic Evaluations of Advertising: Interaction Effects of Strength of Ethnic Identification, Media Placement, and Degree of Racial Composition”, Journal of Advertising, (spring), 28, 1 (1999): 49-64.

Greenlee, Timothy, Oakenfull, Gillian, “All the colors of the Rainbow: The Relationship between Gay Identity and Response to Advertising Content”, Miami University.

Grier, Sony A, Brumbaugh, Anne M., “Consumer Distinctiveness and Advertising Persuasion”, Research Paper no. 1735, Stanford Graduate School of Business, (June), (2002).

Grier, Sonya A., Deshpande, Rohit, “Social Dimensions of Consumer Distinctiveness: The Influence of Social Status Identity and Advertising Persuasion”, Journal of Marketing Research, (May), 38, 2 (2001); 216-224.

Grier, Sonya A., Brumbaugh, Anne M., “Noticing Cultural Differences: Ad Meanings Created by Target and Non-target Markets”, Journal of Advertising, (spring), 28, 1 (1999): 79-93.

Haley, Russel I., Baldinger, Allan L., “The ARF Copy Research Validity”, Advertising Research Foundation, (2000): 114+.

Holbrook, Morris B., Batra, Rajeev, “Assessing the Role of Emotions as Mediators of Consumer Responses to Advertising”, Journal of Consumer Research, (December), 14, (1987): 404-420.

Holland, Jonna L., Gentry, James W., “Ethnic Consumers Reaction to Targeted Marketing: A Theory of Intercultural Accommodation, Journal of Advertising, (spring), 28, 1 (1999): 6577.

Holland, Jonna L., Gentry, James W., “The Impact of Cultural Symbols on Advertising Effectiveness: A Theory of Intercultural Accommodation, Advances in Consumer Research, 24 (1997): 483-489.

62

The Influence of Consumers’ Ethnic Identity on Advertising Responses

IMWO (Instituut voor Maatschappij Wetenschappelijk Onderzoek), “Consumenten Onderzoek in de Mediasector”, Anton de Kom Universiteit van Suriname, 2002.

Kelman, Herbert C., “Process of Opinion Change”, Public Opinion Quarterly, (spring), 25, (1961): 58-78.

Koslow, Scott, Prem N. Shamdasani, Ellen E. Touchstone, “Exploring Language Effects in Ethnic Advertising: A Sociolinguistic Perspective”, Journal of Consumer Research, (March), 20, (1994):575-585.

MacKenzie, Scott B., and Richard J. Lutz, “An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pretesting Context”, Journal of Marketing, (April), 53, (1989): 48-65.

McGuire, William J. and Claire V. McGuire, Pamela Child and Terry Fujioko, “Salience of ethnicity in the spontaneous self-concept as a function of one’s ethnic distinctiveness in the social environment”, Journal of Personality and Social Psychology, 36, 5 (1978): 511-520.

Phinney, Jean S., “Ethic Identity in Adolescents and Adults Review of Research”, Psychological Bulletin, (November), 108, 3, (1990): 499-514

Rossiter, John R., Eagleson, Geoff, “Conclusions from the ARF’s Copy Research Validity Project”, Journal of Advertising, 34, 3 (1994):19+

Saylor, Elizabeth S., Aries, Elizabeth, “Ethnic Identity and Change in Social Context”, Journal of Social Psychology, 139, 5 (1999): 549-566.

Smith, Elsie J., “Ethnic Identity Development: Toward the development of a Theory within the Context of Majority/Minority Status”, Journal of Counseling & Development, (September/October), 70, (1991): 181-188.

Stayman, Douglas M., Deshpande, Rohit, “Situational Ethnicity and Consumer Behavior”, Journal of Consumer Research, (December), 16, (1989): 361-371.

63

The Influence of Consumers’ Ethnic Identity on Advertising Responses

Trimble, Joseph E., Dickson, Ryan, “Ethnic Identity”, in C.B. Fisher & Lerner, R.M. (Eds.; in press), Applied developmental science: An encyclopedia of research, and programs. Thousand Oaks: Sage

Wooten, David B., Galvin, Tiffany, “A Preliminary Examination of the Effects of ContextInduced Felt Ethnicity on Advertising Effectiveness”, Advances in Consumer Research, 20 (1993): 253-256.

Whitler, Tommy E., “The Effects of Actors’ Race in Commercial Advertising: Review and Extension”, Journal of Advertising, 20, 1 (1991): 54-60

Whittler, Tommy E., Spira, Joan Scattone, “Model’s Race: A Peripheral Cue in Advertising Messages?”, Journal of Consumer Psychology, 12, 4 (2002): 291-301

Whittler, Tommy E., DiMeo, Joan, “Viewers’ Ractions to Racial Cues in Advertising Stimuli”, Journal of Advertising Research, 31, (1991): 37-46.

Yip, Tiffany and Andrew J. Fuligni, “Daily Variation in Ethnic Identity, Ethnic Behaviors, and Psychological Well-Being among American Adolescents of Chinese Descent”, Child Development, (September/October), 73, 5 (2002): 1557-1572

Youn-Kyung, Kim, “The Effects of Ethnicity and Product on Purchase Decision Making”, Journal of Advertising Research, (March/April), 41, 2 (2001): 39-48.

http://www.calstatela.edu/academic/psych/ftp/meim.doc (Feb, 2005)

http://www2.chass.ncsu.edu/garson/pa765/anova.htm (June 17, 2005)

http://psych.rice.edu/online_stat/chapter12/anova_designs.html (June 15, 2005)

64

APPENDIX 1 ADVERTISEMENTS

I

Figure 1. Advertisement XH1

II

Figure 2. Advertisement XH2

III

Figure 3. Advertisement XH3

IV

Figure 4. Advertisement CK1

V

Figure 5. Advertisement CK2

VI

Figure 6. Advertisement CK3

VII

APPENDIX 2 QUESTIONNAIRE

VIII

Paramaribo, June 2005 Dear Participant, The purpose of this survey is to obtain your opinion on six proposed advertisements in order for the advertising agency to improve the look, the style and content of those advertisements. You have received 2 advertisements and 1 questionnaire. Please follow the instructions for carefully and answer the questions subsequently. It is important that you answer all questions. No answer is right or wrong. We do appreciate your honest opinion. Your answers are anonymous and will be treated as confidential. Good Luck!! Ellen der Kinderen

IX

SEGMENT A. Please tell us about your self in your own words?

X

SEGMENT B. Look at the first advertisement and answer the questions B1, B2, B3 and B4 B1. Please indicate for each line your opinion by placing a tick in the appropriate box EXAMPLE:

The food is good

The food is bad

X

1.

The ad is good

The ad is bad

2.

I react favorable toward the ad

I react unfavorable toward the ad

3.

I like the ad

I dislike the ad

4.

I believe the ad is effective

I believe the ad is not effective

5.

I find the ad interesting

I find the ad is not interesting

6.

The ad is attractive to me

7.

The ad is unattractive to me

The ad is persuasive

The ad is not persuasive

Score B1:

XI

B2. Please indicate for each statement the extent of your agreement or disagreement by placing a tick in the appropiate column. EXAMPLE:

I disagree completely I disagree completely

I disagree somewhat

I disagree

I disagree somewhat

I neither agree or disagree

I disagree

I neither agree or disagree

3.

I agree somewhat

2.

I agree somewhat

I feel the advertisement was intended for people like me I do not believe I was in the target market the company created the advertisement for The advertiser made that advertisement to appeal people like me

I agree

1.

X

I agree completely

I believe that everybody liked the food

I agree

I agree completely

1.

Score B2:

XII

B3. Please indicate for each line your opinion by placing a tick in the appropriate box

1.

The model seems a warm person

2.

I like the model

The model seems a cold person I do not like the model

3.

The model is sincere

The model is not sincere

4.

The model seems friendly

The model seems unfriendly

5.

The model looks trustworthy

6.

I believe the model

The model looks not trustworthy I do not believe the model Score B3:

B4. Please indicate for each statement the extent of your agreement or disagreement by placing a tick in the appropiate column.

I disagree completely

I disagree

I disagree somewhat

I neither agree or disagree

I agree somewhat

I agree

I agree completely

1.

A person whom I want to be like

2.

My type of person

3.

A person who speaks for a group of which I am a member

Score B4:

XIII

Look at the second advertisement (advertisement B5, B6, B7 and B8

) and answer the questions

B5. Please indicate for each line your opinion by placing a tick in the appropriate box

1.

The ad is good

The ad is bad

2.

I react favorable toward the ad

I react unfavorable toward the ad

3.

I like the ad

I dislike the ad

4.

I believe the ad is effective

I believe the ad is not effective

5.

I find the ad interesting

I find the ad is not interesting

6.

The ad is attractive to me

The ad is unattractive to me

The ad is persuasive

7.

The ad is not persuasive

Score B5:

B6. Please indicate for each statement the extent of your agreement or disagreement by placing a tick in the appropiate column.

I disagree completely

I disagree

I disagree somewhat

I neither agree or disagree

6.

I agree somewhat

5.

I agree

I agree completely

4.

I feel the advertisement was intended for people like me I do not believe I was in the target market the company created the advertisement for The advertiser made that advertisement to appeal people like me Score B6:

XIV

B7. Please indicate for each line your opinion by placing a tick in the appropriate box

1.

The model seems a warm person

2.

I like the model

The model seems a cold person I do not like the model

3.

The model is sincere

The model is not sincere

4.

The model seems friendly

The model seems unfriendly

5.

The model looks trustworthy

6.

I believe the model

The model looks not trustworthy I do not believe the model Score B7:

B8. Please indicate for each statement the extent of your agreement or disagreement by placing a tick in the appropiate column.

I disagree completely

I disagree

I disagree somewhat

I neither agree or disagree

I agree somewhat

I agree

I agree completely

4.

A person whom I want to be like

5.

My type of person

6.

A person who speaks for a group of which I am a member

Score B8:

XV

SEGMENT C. Please tell us what you are NOT?

XVI

SEGMENT D. In this country, people come from many different cultures and there are many different words to describe the different backgrounds or ethnic groups that people come from. Some examples of the names of the ethnic groups are Hindustani, Creole, Javanese, Maroons, Chinese, Amerindian, and Caucasian. These questions are about your ethnicity or ethnic group and how you feel about it or react to it. D1. Please fill in: In terms of ethnic group, I consider myself to be___________________

D2. Please indicate for each statement the extent of your agreement or disagreement by placing a tick in the appropiate column.

4.

I have spent time trying to find out more about my ethnic group, such as its history, traditions, and customs. I am active in organizations or social groups that include mostly members of my ethnic group. I have a clear sense of my ethnic background and what it means for me. I think a lot about how my life will be affected by my ethnic group membership.

5.

I am happy that I am a member of the group I belong to.

6.

I have a strong sense of belonging to my own ethnic group.

7.

I understand pretty well what my ethnic membership means to me. In order to learn more about my ethnic background, I have often talked to other people about my ethnic group.

8. 9.

I disagree strongly

3.

I disagree

2.

I agree

I agree strongly

1.

I have a lot of pride in my ethnic group.

10. I participate in cultural practices of my own group, such as special food, music or customs. 11. I feel a strong attachment towards my own ethnic group. 12. I feel good about my cultural or ethnic background.

Score D2:

XVII

D3. 1. My ethnicity is: 1. Hindustani

2. Creole

3. Javanese

4. Maroon

5. Amerindian

6. Chinese

7. Caucasian

8. Other:

2 My father’s ethnicity is (use number above) : _____ 3. My mother’s ethnicity is (use number above): _____

SEGMENT E. 1. Please indicate your age by placing a tick in the appropriate box. 1. 15-19 years 2. 20-29 years 3. 30-39 years 4. 40-49 years 5. > 49 years

2. Please indicate your level of education by placing a tick in the appropriate box. 1. < 500 SRD 2. 500-1.000 SRD 3. 1000 – 1.500 SRD 4. 1.500 – 2.000 SRD 5. > 2.000 SRD

3. Please indicate your (gross) income by placing a tick in the appropriate box. 1. L.O 2. MULO/LTS 3. NATIN/IMEO 4. HAVO/VWO 5. HBO/ADEK

XVIII

SEGMENT F. Please tell us in your own words the purpose of this survey?

XIX

APPENDIX 3 CREOLE GROUP ANOVA AND MULTI COMPARISONS

XX

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Level of Ethnic Identification (S/W)

ck1 ck2 ck3 xh1 xh2 xh3 s w

16 17 16 13 19 17 50 48

Tests of Between-Subjects Effects Dependent Variable: Attitude towards the Advertisement Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 55.758a 2424.532 50.063 3.296

df 11 1 5 1

Mean Square 5.069 2424.532 10.013 3.296

F 4.004 1915.324 7.910 2.604

Sig. .000 .000 .000 .110

1.665

5

.333

.263

.932

108.864 2633.357 164.622

86 98 97

1.266

a. R Squared = .339 (Adjusted R Squared = .254)

a Levene's Test of Equality of Error Variances

Dependent Variable: Attitude towards the Advertisement F 1.295

df1

df2 11

86

Sig. .241

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXI

Multiple Comparisons Dependent Variable: Attitude towards the Advertisement

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) .0389 -.2613 1.0688 1.7450* .4531 -.0389 -.3002 1.0299 1.7061* .4141 .2613 .3002 1.3300* 2.0063* .7143 -1.0688 -1.0299 -1.3300* .6762 -.6157 -1.7450* -1.7061* -2.0063* -.6762 -1.2920* -.4531 -.4141 -.7143 .6157 1.2920* .0389 -.2613 1.0688 1.7450* .4531 -.0389 -.3002 1.0299 1.7061* .4141 .2613 .3002 1.3300 2.0063* .7143 -1.0688 -1.0299 -1.3300 .6762 -.6157 -1.7450* -1.7061* -2.0063* -.6762 -1.2920* -.4531 -.4141 -.7143 .6157 1.2920*

Std. Error .39189 .39778 .42011 .38176 .39189 .39189 .39189 .41453 .37561 .38591 .39778 .39189 .42011 .38176 .39189 .42011 .41453 .42011 .40497 .41453 .38176 .37561 .38176 .40497 .37561 .39189 .38591 .39189 .41453 .37561 .35466 .31653 .47160 .32702 .35116 .35466 .36327 .50416 .37245 .39381 .31653 .36327 .47811 .33634 .35986 .47160 .50416 .47811 .48512 .50171 .32702 .37245 .33634 .48512 .36912 .35116 .39381 .35986 .50171 .36912

Based on observed means. *. The mean difference is significant at the .05 level.

XXII

Sig. 1.000 .986 .123 .000 .856 1.000 .972 .140 .000 .891 .986 .972 .025 .000 .457 .123 .140 .025 .555 .674 .000 .000 .000 .555 .011 .856 .891 .457 .674 .011 1.000 1.000 .421 .000 .969 1.000 1.000 .560 .001 .995 1.000 1.000 .164 .000 .580 .421 .560 .164 .948 .981 .000 .001 .000 .948 .020 .969 .995 .580 .981 .020

95% Confidence Interval Lower Bound Upper Bound -1.1034 1.1813 -1.4208 .8983 -.1558 2.2934 .6322 2.8579 -.6893 1.5954 -1.1813 1.1034 -1.4425 .8422 -.1785 2.2382 .6112 2.8010 -.7108 1.5390 -.8983 1.4208 -.8422 1.4425 .1054 2.5546 .8935 3.1191 -.4280 1.8567 -2.2934 .1558 -2.2382 .1785 -2.5546 -.1054 -.5042 1.8567 -1.8241 .5926 -2.8579 -.6322 -2.8010 -.6112 -3.1191 -.8935 -1.8567 .5042 -2.3869 -.1971 -1.5954 .6893 -1.5390 .7108 -1.8567 .4280 -.5926 1.8241 .1971 2.3869 -1.0901 1.1680 -1.2680 .7455 -.5171 2.6547 .7132 2.7769 -.6643 1.5704 -1.1680 1.0901 -1.4545 .8542 -.6252 2.6849 .5303 2.8819 -.8314 1.6597 -.7455 1.2680 -.8542 1.4545 -.2687 2.9288 .9450 3.0675 -.4288 1.8574 -2.6547 .5171 -2.6849 .6252 -2.9288 .2687 -.9340 2.2865 -2.2650 1.0335 -2.7769 -.7132 -2.8819 -.5303 -3.0675 -.9450 -2.2865 .9340 -2.4567 -.1272 -1.5704 .6643 -1.6597 .8314 -1.8574 .4288 -1.0335 2.2650 .1272 2.4567

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Level of Ethnic Identification (S/W)

ck1 ck2 ck3 xh1 xh2 xh3 s w

16 17 16 13 19 17 50 48

Tests of Between-Subjects Effects Dependent Variable: Attitude towards the Model Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 28.978a 2741.830 23.003 .070

df 11 1 5 1

Mean Square 2.634 2741.830 4.601 .070

F 2.508 2609.810 4.379 .067

Sig. .009 .000 .001 .797

4.796

5

.959

.913

.477

90.350 2919.521 119.329

86 98 97

1.051

a. R Squared = .243 (Adjusted R Squared = .146)

a Levene's Test of Equality of Error Variances

Dependent Variable: Attitude towards the Model F 2.654

df1

df2 11

86

Sig. .006

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXIII

Multiple Comparisons Dependent Variable: Attitude towards the Model

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) -.7331 -.2263 .3706 .7375 .4269 .7331 .5068 1.1037* 1.4706* 1.1600* .2263 -.5068 .5968 .9638 .6532 -.3706 -1.1037* -.5968 .3669 .0563 -.7375 -1.4706* -.9638 -.3669 -.3106 -.4269 -1.1600* -.6532 -.0563 .3106 -.7331 -.2263 .3706 .7375 .4269 .7331 .5068 1.1037* 1.4706* 1.1600* .2263 -.5068 .5968 .9638 .6532 -.3706 -1.1037* -.5968 .3669 .0563 -.7375 -1.4706* -.9638 -.3669 -.3106 -.4269 -1.1600* -.6532 -.0563 .3106

Std. Error .35702 .36239 .38272 .34779 .35702 .35702 .35702 .37764 .34219 .35157 .36239 .35702 .38272 .34779 .35702 .38272 .37764 .38272 .36893 .37764 .34779 .34219 .34779 .36893 .34219 .35702 .35157 .35702 .37764 .34219 .27711 .33373 .37394 .36313 .39075 .27711 .26705 .31586 .30298 .33559 .33373 .26705 .36654 .35551 .38368 .37394 .31586 .36654 .39349 .41912 .36313 .30298 .35551 .39349 .40950 .39075 .33559 .38368 .41912 .40950

Based on observed means. *. The mean difference is significant at the .05 level.

XXIV

Sig. .322 .989 .927 .287 .838 .322 .715 .049 .001 .017 .989 .715 .627 .072 .453 .927 .049 .627 .918 1.000 .287 .001 .072 .918 .944 .838 .017 .453 1.000 .944 .194 1.000 .998 .540 .993 .194 .661 .040 .001 .033 1.000 .661 .843 .148 .792 .998 .040 .843 .999 1.000 .540 .001 .148 .999 1.000 .993 .033 .792 1.000 1.000

95% Confidence Interval Lower Bound Upper Bound -1.7738 .3076 -1.2826 .8301 -.7450 1.4862 -.2763 1.7513 -.6138 1.4676 -.3076 1.7738 -.5339 1.5475 .0029 2.2045 .4731 2.4681 .1352 2.1848 -.8301 1.2826 -1.5475 .5339 -.5188 1.7124 -.0500 1.9775 -.3875 1.6939 -1.4862 .7450 -2.2045 -.0029 -1.7124 .5188 -.7085 1.4423 -1.0445 1.1571 -1.7513 .2763 -2.4681 -.4731 -1.9775 .0500 -1.4423 .7085 -1.3081 .6869 -1.4676 .6138 -2.1848 -.1352 -1.6939 .3875 -1.1571 1.0445 -.6869 1.3081 -1.6345 .1683 -1.2876 .8351 -.8386 1.5797 -.4082 1.8832 -.8163 1.6701 -.1683 1.6345 -.3590 1.3726 .0335 2.1738 .4955 2.4457 .0590 2.2610 -.8351 1.2876 -1.3726 .3590 -.5917 1.7853 -.1583 2.0858 -.5696 1.8760 -1.5797 .8386 -2.1738 -.0335 -1.7853 .5917 -.8913 1.6251 -1.2850 1.3977 -1.8832 .4082 -2.4457 -.4955 -2.0858 .1583 -1.6251 .8913 -1.6029 .9817 -1.6701 .8163 -2.2610 -.0590 -1.8760 .5696 -1.3977 1.2850 -.9817 1.6029

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Level of Ethnic Identification (S/W)

ck1 ck2 ck3 xh1 xh2 xh3 s w

16 17 15 12 19 17 48 48

Tests of Between-Subjects Effects Dependent Variable: Identification with the Model Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 93.083a 1459.900 79.412 .333

df 11 1 5 1

Mean Square 8.462 1459.900 15.882 .333

F 6.023 1039.048 11.304 .237

Sig. .000 .000 .000 .628

15.051

5

3.010

2.142

.068

118.023 1757.523 211.106

84 96 95

1.405

a. R Squared = .441 (Adjusted R Squared = .368)

a Levene's Test of Equality of Error Variances

Dependent Variable: Identification with the Model F 1.100

df1

df2 11

84

Sig. .371

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXV

Multiple Comparisons Dependent Variable: Identification with the Model

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) -.8743 -.3145 1.4983* 1.5054* .6928 .8743 .5598 2.3726* 2.3797* 1.5671* .3145 -.5598 1.8128* 1.8199* 1.0073 -1.4983* -2.3726* -1.8128* .0071 -.8055 -1.5054* -2.3797* -1.8199* -.0071 -.8126 -.6928 -1.5671* -1.0073 .8055 .8126 -.8743 -.3145 1.4983 1.5054* .6928 .8743 .5598 2.3726* 2.3797* 1.5671* .3145 -.5598 1.8128* 1.8199* 1.0073 -1.4983 -2.3726* -1.8128* .0071 -.8055 -1.5054* -2.3797* -1.8199* -.0071 -.8126 -.6928 -1.5671* -1.0073 .8055 .8126

Std. Error .41287 .42601 .45266 .40220 .41287 .41287 .41990 .44692 .39573 .40657 .42601 .41990 .45908 .40941 .41990 .45266 .44692 .45908 .43708 .44692 .40220 .39573 .40941 .43708 .39573 .41287 .40657 .41990 .44692 .39573 .42220 .41992 .48508 .43469 .42282 .42220 .40111 .46889 .41654 .40414 .41992 .40111 .46683 .41423 .40175 .48508 .46889 .46683 .48016 .46944 .43469 .41654 .41423 .48016 .41717 .42282 .40414 .40175 .46944 .41717

Based on observed means. *. The mean difference is significant at the .05 level.

XXVI

Sig. .288 .977 .017 .004 .550 .288 .766 .000 .000 .003 .977 .766 .002 .000 .168 .017 .000 .002 1.000 .470 .004 .000 .000 1.000 .322 .550 .003 .168 .470 .322 .514 1.000 .074 .022 .830 .514 .942 .001 .000 .007 1.000 .942 .012 .002 .237 .074 .001 .012 1.000 .794 .022 .000 .002 1.000 .603 .830 .007 .237 .794 .603

95% Confidence Interval Lower Bound Upper Bound -2.0784 .3299 -1.5570 .9280 .1781 2.8185 .3324 2.6784 -.5114 1.8970 -.3299 2.0784 -.6649 1.7844 1.0691 3.6761 1.2255 3.5338 .3813 2.7528 -.9280 1.5570 -1.7844 .6649 .4739 3.1518 .6258 3.0140 -.2174 2.2320 -2.8185 -.1781 -3.6761 -1.0691 -3.1518 -.4739 -1.2677 1.2818 -2.1090 .4979 -2.6784 -.3324 -3.5338 -1.2255 -3.0140 -.6258 -1.2818 1.2677 -1.9668 .3415 -1.8970 .5114 -2.7528 -.3813 -2.2320 .2174 -.4979 2.1090 -.3415 1.9668 -2.2146 .4660 -1.6541 1.0251 -.0820 3.0787 .1325 2.8783 -.6494 2.0350 -.4660 2.2146 -.7159 1.8354 .8376 3.9076 1.0684 3.6909 .2889 2.8452 -1.0251 1.6541 -1.8354 .7159 .2783 3.3474 .5095 3.1303 -.2704 2.2850 -3.0787 .0820 -3.9076 -.8376 -3.3474 -.2783 -1.5527 1.5668 -2.3419 .7308 -2.8783 -.1325 -3.6909 -1.0684 -3.1303 -.5095 -1.5668 1.5527 -2.1258 .5006 -2.0350 .6494 -2.8452 -.2889 -2.2850 .2704 -.7308 2.3419 -.5006 2.1258

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Level of Ethnic Identification (S/W)

ck1 ck2 ck3 xh1 xh2 xh3 s w

16 17 16 13 19 17 50 48

Tests of Between-Subjects Effects Dependent Variable: Targetedness Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 114.011a 1737.295 94.262 3.040

df 11 1 5 1

Mean Square 10.365 1737.295 18.852 3.040

F 6.637 1112.416 12.071 1.946

Sig. .000 .000 .000 .167

18.140

5

3.628

2.323

.050

134.309 2040.785 248.320

86 98 97

1.562

a. R Squared = .459 (Adjusted R Squared = .390)

a Levene's Test of Equality of Error Variances

Dependent Variable: Targetedness F 1.265

df1

df2 11

86

Sig. .258

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXVII

Multiple Comparisons Dependent Variable: Targetedness

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) -.2300 -.0719 2.1275* 2.0096* .5977 .2300 .1581 2.3575* 2.2395* .8276 .0719 -.1581 2.1994* 2.0814* .6696 -2.1275* -2.3575* -2.1994* -.1180 -1.5299* -2.0096* -2.2395* -2.0814* .1180 -1.4119* -.5977 -.8276 -.6696 1.5299* 1.4119* -.2300 -.0719 2.1275* 2.0096* .5977 .2300 .1581 2.3575* 2.2395* .8276 .0719 -.1581 2.1994* 2.0814* .6696 -2.1275* -2.3575* -2.1994* -.1180 -1.5299* -2.0096* -2.2395* -2.0814* .1180 -1.4119 -.5977 -.8276 -.6696 1.5299* 1.4119

Std. Error .43529 .44183 .46663 .42403 .43529 .43529 .43529 .46043 .41721 .42864 .44183 .43529 .46663 .42403 .43529 .46663 .46043 .46663 .44981 .46043 .42403 .41721 .42403 .44981 .41721 .43529 .42864 .43529 .46043 .41721 .44935 .36700 .38619 .42815 .41796 .44935 .46302 .47837 .51285 .50437 .36700 .46302 .40201 .44247 .43262 .38619 .47837 .40201 .45851 .44901 .42815 .51285 .44247 .45851 .48558 .41796 .50437 .43262 .44901 .48558

Based on observed means. *. The mean difference is significant at the .05 level.

XXVIII

Sig. .995 1.000 .000 .000 .743 .995 .999 .000 .000 .391 1.000 .999 .000 .000 .641 .000 .000 .000 1.000 .016 .000 .000 .000 1.000 .013 .743 .391 .641 .016 .013 1.000 1.000 .000 .001 .931 1.000 1.000 .001 .002 .828 1.000 1.000 .000 .001 .881 .000 .001 .000 1.000 .030 .001 .002 .001 1.000 .091 .931 .828 .881 .030 .091

95% Confidence Interval Lower Bound Upper Bound -1.4988 1.0389 -1.3598 1.2161 .7673 3.4878 .7735 3.2456 -.6712 1.8665 -1.0389 1.4988 -1.1108 1.4269 1.0154 3.6997 1.0234 3.4557 -.4218 2.0771 -1.2161 1.3598 -1.4269 1.1108 .8392 3.5596 .8454 3.3175 -.5993 1.9384 -3.4878 -.7673 -3.6997 -1.0154 -3.5596 -.8392 -1.4292 1.1932 -2.8720 -.1877 -3.2456 -.7735 -3.4557 -1.0234 -3.3175 -.8454 -1.1932 1.4292 -2.6280 -.1957 -1.8665 .6712 -2.0771 .4218 -1.9384 .5993 .1877 2.8720 .1957 2.6280 -1.6704 1.2104 -1.2396 1.0959 .8777 3.3774 .6525 3.3666 -.7355 1.9309 -1.2104 1.6704 -1.3197 1.6359 .8261 3.8889 .6231 3.8560 -.7690 2.4243 -1.0959 1.2396 -1.6359 1.3197 .9040 3.4949 .6827 3.4802 -.7059 2.0451 -3.3774 -.8777 -3.8889 -.8261 -3.4949 -.9040 -1.5760 1.3401 -2.9663 -.0934 -3.3666 -.6525 -3.8560 -.6231 -3.4802 -.6827 -1.3401 1.5760 -2.9405 .1167 -1.9309 .7355 -2.4243 .7690 -2.0451 .7059 .0934 2.9663 -.1167 2.9405

APPENDIX 4 HINDUSTANI GROUP ANOVA AND MULTI COMPARISONS

XXIX

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Strength of Ethnic Identification

ck1 ck2 ck3 xh1 xh2 xh3 s w

15 17 15 19 11 17 50 44

Tests of Between-Subjects Effects Dependent Variable: Attitude towards the Advertisement Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 35.153a 2345.022 18.637 1.569

df 11 1 5 1

Mean Square 3.196 2345.022 3.727 1.569

F 4.194 3077.277 4.891 2.059

Sig. .000 .000 .001 .155

15.568

5

3.114

4.086

.002

62.488 2509.949 97.641

82 94 93

.762

a. R Squared = .360 (Adjusted R Squared = .274)

a Levene's Test of Equality of Error Variances

Dependent Variable: Attitude towards the Advertisement F 2.036

df1

df2 11

82

Sig. .035

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXX

Multiple Comparisons Dependent Variable: Attitude towards the Advertisement

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) -.1391 -1.0013* -.7231 -1.0011 -1.1038* .1391 -.8623 -.5840 -.8620 -.9647* 1.0013* .8623 .2783 .0002 -.1024 .7231 .5840 -.2783 -.2780 -.3807 1.0011 .8620 -.0002 .2780 -.1027 1.1038* .9647* .1024 .3807 .1027 -.1391 -1.0013* -.7231 -1.0011 -1.1038* .1391 -.8623 -.5840 -.8620 -.9647 1.0013* .8623 .2783 .0002 -.1024 .7231 .5840 -.2783 -.2780 -.3807 1.0011 .8620 -.0002 .2780 -.1027 1.1038* .9647 .1024 .3807 .1027

Std. Error .30924 .31876 .30151 .34653 .30924 .30924 .30924 .29143 .33779 .29942 .31876 .30924 .30151 .34653 .30924 .30151 .29143 .30151 .33073 .29143 .34653 .33779 .34653 .33073 .33779 .30924 .29942 .30924 .29143 .33779 .27250 .29807 .32215 .34337 .31362 .27250 .30789 .33126 .35193 .32297 .29807 .30789 .35260 .37208 .34482 .32215 .33126 .35260 .39164 .36584 .34337 .35193 .37208 .39164 .38465 .31362 .32297 .34482 .36584 .38465

Based on observed means. *. The mean difference is significant at the .05 level.

XXXI

Sig. .998 .027 .169 .054 .008 .998 .070 .349 .121 .022 .027 .070 .940 1.000 .999 .169 .349 .940 .959 .781 .054 .121 1.000 .959 1.000 .008 .022 .999 .781 1.000 1.000 .035 .388 .131 .022 1.000 .127 .746 .307 .080 .035 .127 1.000 1.000 1.000 .388 .746 1.000 1.000 .996 .131 .307 1.000 1.000 1.000 .022 .080 1.000 .996 1.000

95% Confidence Interval Lower Bound Upper Bound -1.0415 .7634 -1.9315 -.0711 -1.6029 .1568 -2.0123 .0101 -2.0062 -.2013 -.7634 1.0415 -1.7647 .0401 -1.4345 .2665 -1.8478 .1237 -1.8385 -.0909 .0711 1.9315 -.0401 1.7647 -.6016 1.1582 -1.0110 1.0115 -1.0049 .8000 -.1568 1.6029 -.2665 1.4345 -1.1582 .6016 -1.2432 .6871 -1.2312 .4697 -.0101 2.0123 -.1237 1.8478 -1.0115 1.0110 -.6871 1.2432 -1.0884 .8831 .2013 2.0062 .0909 1.8385 -.8000 1.0049 -.4697 1.2312 -.8831 1.0884 -1.0055 .7274 -1.9596 -.0431 -1.7458 .2997 -2.1609 .1587 -2.1058 -.1017 -.7274 1.0055 -1.8453 .1208 -1.6303 .4623 -2.0376 .3135 -1.9905 .0611 .0431 1.9596 -.1208 1.8453 -.8369 1.3934 -1.2279 1.2283 -1.1988 .9939 -.2997 1.7458 -.4623 1.6303 -1.3934 .8369 -1.5490 .9929 -1.5324 .7710 -.1587 2.1609 -.3135 2.0376 -1.2283 1.2279 -.9929 1.5490 -1.3592 1.1539 .1017 2.1058 -.0611 1.9905 -.9939 1.1988 -.7710 1.5324 -1.1539 1.3592

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Strength of Ethnic Identification

ck1 ck2 ck3 xh1 xh2 xh3 s w

15 17 15 19 11 17 50 44

Tests of Between-Subjects Effects Dependent Variable: Attitude towards the Model Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 21.713a 2427.826 14.171 .026

df 11 1 5 1

Mean Square 1.974 2427.826 2.834 .026

F 2.979 3664.011 4.277 .039

Sig. .002 .000 .002 .844

8.556

5

1.711

2.582

.032

54.334 2577.616 76.048

82 94 93

.663

a. R Squared = .286 (Adjusted R Squared = .190)

a Levene's Test of Equality of Error Variances

Dependent Variable: Attitude towards the Model F 3.777

df1

df2 11

82

Sig. .000

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXXII

Multiple Comparisons Dependent Variable: Attitude towards the Model

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) .0184 -.6520 -.8233* -.8561 -.7451 -.0184 -.6704 -.8418* -.8745 -.7635 .6520 .6704 -.1713 -.2041 -.0931 .8233* .8418* .1713 -.0327 .0782 .8561 .8745 .2041 .0327 .1110 .7451 .7635 .0931 -.0782 -.1110 .0184 -.6520 -.8233 -.8561 -.7451 -.0184 -.6704 -.8418 -.8745 -.7635 .6520 .6704 -.1713 -.2041 -.0931 .8233 .8418 .1713 -.0327 .0782 .8561 .8745 .2041 .0327 .1110 .7451 .7635 .0931 -.0782 -.1110

Std. Error .28836 .29723 .28116 .32313 .28836 .28836 .28836 .27176 .31498 .27920 .29723 .28836 .28116 .32313 .28836 .28116 .27176 .28116 .30840 .27176 .32313 .31498 .32313 .30840 .31498 .28836 .27920 .28836 .27176 .31498 .30413 .24408 .29657 .29111 .28070 .30413 .28414 .33033 .32544 .31616 .24408 .28414 .27604 .27017 .25891 .29657 .33033 .27604 .31839 .30890 .29111 .32544 .27017 .31839 .30366 .28070 .31616 .25891 .30890 .30366

Based on observed means. *. The mean difference is significant at the .05 level.

XXXIII

Sig. 1.000 .252 .049 .097 .113 1.000 .196 .031 .072 .079 .252 .196 .990 .988 1.000 .049 .031 .990 1.000 1.000 .097 .072 .988 1.000 .999 .113 .079 1.000 1.000 .999 1.000 .174 .129 .109 .173 1.000 .325 .210 .173 .281 .174 .325 1.000 1.000 1.000 .129 .210 1.000 1.000 1.000 .109 .173 1.000 1.000 1.000 .173 .281 1.000 1.000 1.000

95% Confidence Interval Lower Bound Upper Bound -.8231 .8599 -1.5194 .2154 -1.6438 -.0029 -1.7990 .0869 -1.5866 .0964 -.8599 .8231 -1.5119 .1711 -1.6348 -.0487 -1.7937 .0447 -1.5783 .0512 -.2154 1.5194 -.1711 1.5119 -.9918 .6491 -1.1470 .7389 -.9346 .7484 .0029 1.6438 .0487 1.6348 -.6491 .9918 -.9327 .8673 -.7148 .8713 -.0869 1.7990 -.0447 1.7937 -.7389 1.1470 -.8673 .9327 -.8082 1.0301 -.0964 1.5866 -.0512 1.5783 -.7484 .9346 -.8713 .7148 -1.0301 .8082 -.9507 .9876 -1.4356 .1316 -1.7616 .1149 -1.8130 .1009 -1.6377 .1475 -.9876 .9507 -1.5829 .2421 -1.8824 .1989 -1.9257 .1767 -1.7649 .2379 -.1316 1.4356 -.2421 1.5829 -1.0487 .7060 -1.1080 .6998 -.9197 .7335 -.1149 1.7616 -.1989 1.8824 -.7060 1.0487 -1.0579 .9924 -.8943 1.0507 -.1009 1.8130 -.1767 1.9257 -.6998 1.1080 -.9924 1.0579 -.8764 1.0983 -.1475 1.6377 -.2379 1.7649 -.7335 .9197 -1.0507 .8943 -1.0983 .8764

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Strength of Ethnic Identification

ck1 ck2 ck3 xh1 xh2 xh3 s w

15 17 15 19 11 17 50 44

Tests of Between-Subjects Effects Dependent Variable: Identification with the Model Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 52.813a 1867.681 36.123 1.980

df 11 1 5 1

Mean Square 4.801 1867.681 7.225 1.980

F 4.167 1621.016 6.270 1.719

Sig. .000 .000 .000 .194

16.497

5

3.299

2.864

.020

94.478 2043.867 147.291

82 94 93

1.152

a. R Squared = .359 (Adjusted R Squared = .273)

a Levene's Test of Equality of Error Variances

Dependent Variable: Identification with the Model F 6.820

df1

df2 11

82

Sig. .000

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXXIV

Multiple Comparisons Dependent Variable: Identification with the Model

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) -.6306 -.2867 -1.0111 -1.7444* -1.5418* .6306 .3440 -.3805 -1.1138 -.9112 .2867 -.3440 -.7245 -1.4578* -1.2551* 1.0111 .3805 .7245 -.7333 -.5307 1.7444* 1.1138 1.4578* .7333 .2026 1.5418* .9112 1.2551* .5307 -.2026 -.6306 -.2867 -1.0111* -1.7444* -1.5418* .6306 .3440 -.3805 -1.1138* -.9112 .2867 -.3440 -.7245 -1.4578 -1.2551 1.0111* .3805 .7245 -.7333 -.5307 1.7444* 1.1138* 1.4578 .7333 .2026 1.5418* .9112 1.2551 .5307 -.2026

Std. Error .38024 .39195 .37074 .42609 .38024 .38024 .38024 .35835 .41535 .36817 .39195 .38024 .37074 .42609 .38024 .37074 .35835 .37074 .40667 .35835 .42609 .41535 .42609 .40667 .41535 .38024 .36817 .38024 .35835 .41535 .25554 .50765 .21203 .27111 .36657 .25554 .53476 .27060 .31901 .40329 .50765 .53476 .51539 .54237 .59586 .21203 .27060 .51539 .28535 .37722 .27111 .31901 .54237 .28535 .41333 .36657 .40329 .59586 .37722 .41333

Based on observed means. *. The mean difference is significant at the .05 level.

XXXV

Sig. .563 .977 .081 .001 .002 .563 .944 .895 .090 .144 .977 .944 .378 .012 .017 .081 .895 .378 .469 .677 .001 .090 .012 .469 .996 .002 .144 .017 .677 .996 .266 1.000 .001 .000 .006 .266 1.000 .939 .028 .387 1.000 1.000 .947 .194 .500 .001 .939 .947 .245 .942 .000 .028 .194 .245 1.000 .006 .387 .500 .942 1.000

95% Confidence Interval Lower Bound Upper Bound -1.7403 .4790 -1.4304 .8571 -2.0930 .0708 -2.9878 -.5010 -2.6514 -.4322 -.4790 1.7403 -.7657 1.4536 -1.4262 .6652 -2.3259 .0983 -1.9856 .1632 -.8571 1.4304 -1.4536 .7657 -1.8064 .3574 -2.7012 -.2143 -2.3648 -.1455 -.0708 2.0930 -.6652 1.4262 -.3574 1.8064 -1.9201 .4535 -1.5764 .5151 .5010 2.9878 -.0983 2.3259 .2143 2.7012 -.4535 1.9201 -1.0095 1.4147 .4322 2.6514 -.1632 1.9856 .1455 2.3648 -.5151 1.5764 -1.4147 1.0095 -1.4525 .1913 -2.0266 1.4533 -1.6818 -.3404 -2.6701 -.8187 -2.7519 -.3317 -.1913 1.4525 -1.4393 2.1272 -1.2396 .4786 -2.1521 -.0755 -2.2052 .3828 -1.4533 2.0266 -2.1272 1.4393 -2.4747 1.0258 -3.2620 .3465 -3.1779 .6676 .3404 1.6818 -.4786 1.2396 -1.0258 2.4747 -1.6844 .2178 -1.7615 .7001 .8187 2.6701 .0755 2.1521 -.3465 3.2620 -.2178 1.6844 -1.1312 1.5365 .3317 2.7519 -.3828 2.2052 -.6676 3.1779 -.7001 1.7615 -1.5365 1.1312

Univariate Analysis of Variance Between-Subjects Factors N Cultural Embeddedness

Strength of Ethnic Identification

ck1 ck2 ck3 xh1 xh2 xh3 s w

15 17 15 19 11 17 50 44

Tests of Between-Subjects Effects Dependent Variable: Targetedness Source Corrected Model Intercept Embeddedness Strength Embeddedness * Strength Error Total Corrected Total

Type III Sum of Squares 84.308a 1983.097 66.071 .295

df 11 1 5 1

Mean Square 7.664 1983.097 13.214 .295

F 10.127 2620.410 17.461 .390

Sig. .000 .000 .000 .534

19.022

5

3.804

5.027

.000

62.057 2145.630 146.365

82 94 93

.757

a. R Squared = .576 (Adjusted R Squared = .519)

a Levene's Test of Equality of Error Variances

Dependent Variable: Targetedness F .995

df1

df2 11

82

Sig. .458

Tests the null hypothesis that the error variance of the dependent variable is equal across groups. a. Design: Intercept+Embeddedness+Strength+ Embeddedness * Strength

XXXVI

Multiple Comparisons Dependent Variable: Targetedness

Tukey HSD

(I) Cultural Embeddedness ck1

ck2

ck3

xh1

xh2

xh3

Tamhane

ck1

ck2

ck3

xh1

xh2

xh3

(J) Cultural Embeddedness ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2 ck2 ck3 xh1 xh2 xh3 ck1 ck3 xh1 xh2 xh3 ck1 ck2 xh1 xh2 xh3 ck1 ck2 ck3 xh2 xh3 ck1 ck2 ck3 xh1 xh3 ck1 ck2 ck3 xh1 xh2

Mean Difference (I-J) -.2300 -.0719 2.1275* 2.0096* .5977 .2300 .1581 2.3575* 2.2395* .8276 .0719 -.1581 2.1994* 2.0814* .6696 -2.1275* -2.3575* -2.1994* -.1180 -1.5299* -2.0096* -2.2395* -2.0814* .1180 -1.4119* -.5977 -.8276 -.6696 1.5299* 1.4119* -.2300 -.0719 2.1275* 2.0096* .5977 .2300 .1581 2.3575* 2.2395* .8276 .0719 -.1581 2.1994* 2.0814* .6696 -2.1275* -2.3575* -2.1994* -.1180 -1.5299* -2.0096* -2.2395* -2.0814* .1180 -1.4119 -.5977 -.8276 -.6696 1.5299* 1.4119

Based on observed means. *. The mean difference is significant at the .05 level.

XXXVII

Std. Error .43529 .44183 .46663 .42403 .43529 .43529 .43529 .46043 .41721 .42864 .44183 .43529 .46663 .42403 .43529 .46663 .46043 .46663 .44981 .46043 .42403 .41721 .42403 .44981 .41721 .43529 .42864 .43529 .46043 .41721 .44935 .36700 .38619 .42815 .41796 .44935 .46302 .47837 .51285 .50437 .36700 .46302 .40201 .44247 .43262 .38619 .47837 .40201 .45851 .44901 .42815 .51285 .44247 .45851 .48558 .41796 .50437 .43262 .44901 .48558

Sig. .995 1.000 .000 .000 .743 .995 .999 .000 .000 .391 1.000 .999 .000 .000 .641 .000 .000 .000 1.000 .016 .000 .000 .000 1.000 .013 .743 .391 .641 .016 .013 1.000 1.000 .000 .001 .931 1.000 1.000 .001 .002 .828 1.000 1.000 .000 .001 .881 .000 .001 .000 1.000 .030 .001 .002 .001 1.000 .091 .931 .828 .881 .030 .091

95% Confidence Interval Lower Bound Upper Bound -1.4988 1.0389 -1.3598 1.2161 .7673 3.4878 .7735 3.2456 -.6712 1.8665 -1.0389 1.4988 -1.1108 1.4269 1.0154 3.6997 1.0234 3.4557 -.4218 2.0771 -1.2161 1.3598 -1.4269 1.1108 .8392 3.5596 .8454 3.3175 -.5993 1.9384 -3.4878 -.7673 -3.6997 -1.0154 -3.5596 -.8392 -1.4292 1.1932 -2.8720 -.1877 -3.2456 -.7735 -3.4557 -1.0234 -3.3175 -.8454 -1.1932 1.4292 -2.6280 -.1957 -1.8665 .6712 -2.0771 .4218 -1.9384 .5993 .1877 2.8720 .1957 2.6280 -1.6704 1.2104 -1.2396 1.0959 .8777 3.3774 .6525 3.3666 -.7355 1.9309 -1.2104 1.6704 -1.3197 1.6359 .8261 3.8889 .6231 3.8560 -.7690 2.4243 -1.0959 1.2396 -1.6359 1.3197 .9040 3.4949 .6827 3.4802 -.7059 2.0451 -3.3774 -.8777 -3.8889 -.8261 -3.4949 -.9040 -1.5760 1.3401 -2.9663 -.0934 -3.3666 -.6525 -3.8560 -.6231 -3.4802 -.6827 -1.3401 1.5760 -2.9405 .1167 -1.9309 .7355 -2.4243 .7690 -2.0451 .7059 .0934 2.9663 -.1167 2.9405