• No results found

THE IMPACT OF SOCIAL ENVIRONMENTS AND MOVIE TYPES ON THE EFFCT AND THE CHARACTERISTICS OF SOCIAL INTERACTIONS IN THE MOVIE MARKET

N/A
N/A
Protected

Academic year: 2021

Share "THE IMPACT OF SOCIAL ENVIRONMENTS AND MOVIE TYPES ON THE EFFCT AND THE CHARACTERISTICS OF SOCIAL INTERACTIONS IN THE MOVIE MARKET"

Copied!
87
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

THE IMPACT OF SOCIAL ENVIRONMENTS AND MOVIE

TYPES ON THE EFFCT AND THE CHARACTERISTICS OF

SOCIAL INTERACTIONS IN THE MOVIE MARKET

By Fang Zeng

Master Thesis

Submitted to the Faculty of Economics in Fulfilment of the Requirements for the Degree of Master of Science in International Economics and Business

(2)

Abstract

(3)

TABLE OF CONTENTS

Page ABSTRACT………... 2 LIST OF TABLES……… 5 CHAPTER 1: INTRODUCTION………. 6 1.1. Introduction to Research………. 6

1.2. Problem Statement and Research Questions………... 8

1.3. Research Structure………... 9

CHAPTER 2: THEORY AND HYPOTHESES………. 11

2.1. Literature Review………. 11

Social Interaction………... 11

Social Environment……….. 14

Motion Pictures……… 16

2.2. Conceptual Framework and Research Hypotheses……….. 17

2.2.1. Social Environments Influence Social Interactions……….... 17

Cultural Norms and Level of Constraint……… 18

Institutions and Level of Constraint………... 21

Constraints on Social Interactions……….. 25

2.2.2. Type of Movie influences Social Interactions………. 26

Supply of Information and Type of Movie………. 27

Level of Information Supply and Social Interactions………. 28

CHAPTER 3: RESEARCH METHOD………. 31

3.1. Introduction to Research Method………. 31

3.2. Selection of Country for Social Environment……….. 31

3.3. Identification of Type of Movie……….... 32

3.4. Target Population……….. 33

3.5. Sample Frame and Size………. 33

3.6. Fieldwork: Questionnaire……….. 34

3.7. Data Collection……….. 36

3.8. Research Variables and Measurements……….. 39

3.9. Operationalization and Data Analysis……… 43

3.9.1. Sample Characteristics………. 43

3.9.2. Exploratory Factor Analysis………. 44

3.9.3. Comparison of Means between Subgroups……….. 46

3.9.4. Descriptive Statistics and Correlations………. 49

(4)

CHAPTER 4: RESULTS AND DISCUSSION……… 55

4.1. Results and Evaluation of Research Hypotheses……… 55

4.2. Discussion………... 73

CHAPTER 5: CONCLUSION AND LIMITATIONS……….. 77

5.1. Conclusion……… 77

5.2. Limitations and Further Research………... 79

REFERENCES……….. 81

APPENDIX A. Mainstream and Art House Movies in the Sample……… 88

B. Survey Questionnaire………. 90

C. Codebook of Questionnaire Items……….. 92

(5)

LIST OF TABLES

TABLE Page

2.1. Cultural Dimension Scores for Eight Countries 20

2.2. Key Figures of Motion Picture Industry in the Netherlands and China in 2004 24

2.3. Effects of Constraints on Social Interactions 25

3.1. Sample Demographics 44

3.2. Exploratory Factor Analysis for Motivations of Movie Attendance 50

3.3. Compare Sample Means of Subgroups 52

3.4. Compare Sample Medians of Subgroups 53

4.1. Results of Linear Regression for Frequent WOM Activity and Explanatory Variables 55

4.2. Results of Linear Regression for WOM Inflow Activity and Explanatory Variables 58

4.3. Results of Linear Regression for Level of WOM Influence and Explanatory Variables 60

4.4. Results of Linear Regression for WOM Outflow Activity and Explanatory Variables 62

4.5. Results of Linear Regression for Detailed WOM Activity and Explanatory Variables 54

4.6. Results of Multinomial Logistic Regression for WOM General Discussion and Explanatory Variables 66

4.7. Results of Multinomial Logistic Regression for Effect of CC Behaviour and Explanatory Variables in the Specific Case 68

4.8. Results of Multinomial Logistic Regression for Effect of CC Behaviour and Explanatory Variables in the General Case 69

4.9. Results of Multinomial Logistic Regression for Characteristics of CC Behaviour and Explanatory Variables in the Specific Case 71

(6)

CHAPTER 1: INTRODUCTION

1.1. Introduction to Research

Based on the results of many studies, the globalization of economic and business activities has been one of the most important trends in the world economy due to the widespread availability and use of technology, the expansion of global finance, the development of multinational firms, the decentralization of economic activities, as well as the reductions in trade barriers (Amin & Thrift, 1994; Dicken, 1992; Dicken 1994; Dunning, 2000). In the meanwhile, as some researchers discussed, the society, where those economic and business activities take place, is transforming from a market-based concept to a network concept, signalled by a conflux of technological, economic, socio-political, and cultural developments (Ozcan, 2004). “Everything is becoming electronically connected to everything else: products, people, companies, countries, everything” (Davis & Meyer, 1998) and “Communication is the economy” (Kelly, 1998). Regarding this global perspective of the world economy and network social reality in the Internet era, recent study in economics has begun to integrate sociological views into the modelling of consumer behaviour in terms of the individual choice and the final consumption (Durlauf, 2001). In line with this new approach, many researches have focused on how social environments and interactions between consumers influence individual decision-making processes. Consequently, an understanding of consumer behaviour, in terms of consumer to consumer interactions, is essential to clarify the challenges faced by firms.Especially, for multinational firms, they target consumers across countries. In this case, consumers might interact differently due to the constraints from social environments, such as cultural norms.

(7)

consumers to consider the other’s decision while making their own decisions. With this respect, the choices of individuals and firms are interdependent (Ozcan, 2004). Firms should take consumers’ social behaviour into account to understand the market structure and dynamics, to forecast financial performance of products, and to assess the impact of new technologies and innovations. Thus, they add market value for their products (Prahalad & Ramaswamy, 2004). On the other hand, consumer’s social behaviour is influenced by social environments (Webster et. al., 2001). How consumers interact with others is influenced by the culture, institutions etc. of a society. As a result, the integration of social behaviour into the decision-making process should be considered given a social environment. In view of the above two perspectives, some researchers aimed to relate some variables, like opinions, needs, values, attitudes and personal traits to consumers’ decision-making processes, to model the effect of social interactions in the consumption life cycle, in order to understand what drives product consumption (Austin, 1986, 1989; Eliashberg et al., 2005, and Hidalgo et al., 2006) and the dynamics of consumption. Meanwhile, some were interested in the difference between social interactions regarding different social environments (Granovetter, 1973, 1985; Feld, 1981; Fisher, 1982; Webster et. al., 2001). However, most of recent studies only focused on one aspect of the integration; there were rare studies on consumers’ social behaviour across societies. Therefore, we examine the effects and the characteristics of consumers’ social interactions under the consideration of social environments in this research.

However, it is hard to set up a research on consumer behaviours and consumer interactions at an aggregate level for the future competition in international markets given different social environments. Regarding this, we choose the motion picture industry as our research field. The reasons are associated to the trend of the world economy and the transformation of the society to a much closer network where consumers exchange always more information to each other.

(8)

many firms and individuals from different societies are involved in it and motion pictures are a key driver of the market for entertainment products (Eliashberg et al., 2005). For instance, in 2004 alone, spending on movie tickets was around $9 billion in the U.S. and almost $11 billion in the world (Standard & Poor’s 2004). In addition, as Eliashberg noted, the United States has been the dominant player in the movie market, and the entertainment industry is the leading export industry in the United States. In 2004, the market shares of US films are from 48.3% (in France) to 85.1% (in Iceland) across Western European countries (European Cinema Yearbook, 2005). Meantime, in China, Hollywood imports account for 30% to 50% of all box office revenues, even for 60% to 80% in some cities (Access Asia Limited, 2004).Therefore, movie’s production, distribution and exhibition are involved in an international business environment, which provides a research field for understanding consumer behaviours in different countries.

Secondly, movie consumption is closely related to the social interactions due to the important role of communication in terms of motion picture success (Hennig-Thurau et al., 2001; Hidalgo et al., 2006). Thus, we could integrate sociological views into the modelling of consumer behaviour and dynamics of consumption. Meanwhile, due to visible product characteristics, movies provide a research field where we are able to develop and test the theories about consumer interactions, and then to understand how they influence the choices made by consumers.

The third reason is related to the abundance and public availability of rich data. This makes this industry more appealing for our research.

1.2. Problem Statement and Research Questions

(9)

interact is not only a matter of individual choice. Considering the difference between societies, with whom and how much the moviegoers interact could be constrained by all kinds of physical, technical, social and cultural aspects of the environments where the interactions take place (Homans, 1950; Feld, 1981; Fischer, 1982). Meanwhile, the influence of ‘type of movie’ on social interactions should also be taken into account since moviegoers are differently motivated for different types of movie (Bourdieu, 1984). The objective of this research is to observe and compare the interactions between moviegoers given different social environments, and different types of movie, in order to examine the effect and the characteristics of the interactions across societies, and to discover the differences related to the level of constraints in a social environment and the movie type. Therefore, our specific research questions are addressed as follows:

Are there differences between the effect and the characteristics of social interactions in the movie market concerning the social environments? Whether are the differences of social interactions related to type of movie? Which factors might also influence social interactions given a social environment or a type of movie?

1.3. Research Structure

(10)
(11)

CHAPTER 2: THEORY AND HYPOTHESES

2.1. Literature Review

We first present a review of the previous researches, which comprises the theoretical basis for this research. In order to observe and compare the effect and the characteristics of moviegoers’ interactions in different social environments given the type of movie, we focus on those studies on social interactions, social environments and motion pictures.

Social Interaction

Typically, economists assume that individual behaviour and choice are not directly influenced by opinions, attitudes and actions of others; even if they believe that these indirect effects are greatly important for the individual market behaviour. The reason is there are few techniques which could be used for analyzing social influences on a behaviour. However, as Becker and Murphy (2003) explained, those activities and consumption taking place publicly are most subject to strong social influences from others. Also they took movie consumption as a real example to attest the importance of social interactions.

Social interaction: According to Rummel (1976), social interaction is any behaviour affecting or considering each other’s subjective experiences or intentions. Since social interaction requires a mutual orientation, both parties should be aware of each other’s actions and react to each other in a dynamic and changing process.

(12)

available to one another. Thus, the social interactions among moviegoers concern the transmission of information about movie performance, which provides the knowledge on perceived quality of a movie to a consumer. On the other hand, preference interactions consider individuals choose actions simultaneously. The utility each individual obtains is based on the actions of the others. In this regard, Hidalgo, Castro and Rodriguez-Sickert (2006) assumed that one moviegoer’ final attendance at the cinema is associated with the choice the others made since consumers usually attend in group, that is, the coordinated consumption (CC).

Moreover, as Eliashberg (2005) summarized, moviegoers are heavily influenced by others’ opinions and choices. Movie attendance has a strong social component. Many researchers relate factors like opinions, needs, values, attitudes to consumers’ choices (Austin, 1986/1989; Becker et al.; 1985).

Therefore, the transmission of information about observed value of a movie and the coordinated movie consumption are associated with individual decisions to attend movies. On one hand, moviegoers’ choices are interdependent as the choice-related information, opinions, attitudes transmitted between them (Ozcan, 2004). In terms of this kind of consumer to consumer interaction about information, many researchers have focused on word-of-mouth (WOM) (Whyte, 1954; Katz & Lazarsfeld, 1955; Arndt, 1967; Anderson, 1998; Chevalier & Mayzlin, 2003; Ozcan, 2004), since it is a widely accepted notion in consumer behaviour, and has great influence on shaping consumers’ attitudes and behaviours (Brown & Reingen, 1987). On the other hand, moviegoers’ choices are also interdependent as the others’ movie attendance, such as CC explained by Hidalgo et al. (2006). Social interactions between moviegoers could be examined in terms of WOM and CC in the movie market (Hidalgo et al., 2006), which provides the opportunity to explore the factors influencing movie attendance.

(13)

communication from one consumer to another consumer (Anderson, 1998). It is reflected in the discussion and expression of relevant opinions and attitudes on movie performance, concerning positive or negative consumption experience with movies (Singh, 1990; Anderson, 1998; Eliashberg et al., 2005).

However, there is little consensus about the definition, the mechanism and the motivation of WOM. As early as Katz and Lazarsfeld (1955) found, WOM communication is the most important source of influence in demand. Concerning the movie market, Assael (1987) found interpersonal communication, like WOM, is critical to a movie’s success since consumers trust their family, friends, or acquaintances more than the other commercial source of information, like advertisement on TV and trailers. Also, Austin (1989) stressed the importance of extent and direction of WOM in the movie market. In that case, we can expect the success of a movie depends on how much moviegoers communicate with each other and how the communication is related to moviegoers’ behaviour, negatively or positively. More recently, Hennig-Thurau et al. (2001) found that WOM can also influence moviegoers’ expectations on a movie’s quality, like advertising and critic reviews. Chevalier and Mayzlin (2003) also suggested WOM is particularly crucial for the success of entertainment goods because these goods, like movies, are quite often consumed collectively, and also involved in daily conversation. We consider WOM as a social phenomenon in this study.

(14)

if more than two moviegoers who have not seen the movie meet, they are also more likely to go together, which indicates a positive effect on attendance likelihood. At present, there are few researches having done in this field concerning motion pictures. Since this consuming behaviour arises from and is constrained by moviegoers’ social relations with others, we also consider CC as a social phenomenon in this study.

Social Environment

In terms of WOM activity and CC behaviour, we focus on communication between consumers. However, as Newcomb, Turner and Converse (1965) stated, people communicate most frequently with those who they feel like, if they can. That means with whom and how much people choose to interact is not simply based on their own feeling, or sentiments (Webster et al., 2001). Some external conditions are also associated with people’s interactions with the others. Granovetter (1973) indicated the significance of how people are socially connected for communication. Moreover, he stated that people do not behave or decide independently outside a social environment; instead their interactions are embedded in concrete, ongoing environments (Granovetter, 1985). Many sociologists believe that the effect and the characteristics of social interactions are somehow restricted (Homans, 1950; Blau, 1964, 1986; Feld, 1981; Fisher, 1982). As Blau (1964, 1986) suggested, individuals make choices under external conditions that could restrict or work for their purposive actions and interactions. Therefore, we take social environment into account, which is also called social context.

(15)

organizations, governance and social progress etc.

With regard to the motion picture industry, different social environments indicate different levels of constraints, which affect moviegoers’ behaviours and support or limit the interactions between them.

Constraints from social environment: this concept is employed in many fields of a society. In general, it refers to those social factors which affect the implementation of particular actions. In the motion picture industry, constrains from a social environment are recognized in terms of cultural norms and institutions. This is because motion picture is produced, distributed and exhibited by different types of organizations and individuals along the value chain (Eliashberg et al., 2005). In this end, it is consumed by individuals with different cultural and social backgrounds.

We recognize the constraints from cultural norms based on Hofstede’s (1993, 2003) cultural dimensions, such as Power Distance, Individualism versus Collectivism and Long-term versus Short-term Orientation. For the constraints from the institutions, we define the institutions in the movie market first. From one perspective, institutions are bundles of regulations, polices authorized by government agencies, like the Ministry and administrators, which support or limit the cinema activities (DiMaggio & Powell, 1983; Scott, 1987; Borum & Westenholz, 1995; Karnøe, 1995). For instance, the European Union’s MEDIA Programme supported the smaller film productions, and the next MEDIA II and MEDIA PLUS aimed at strengthening the competitiveness of the European audiovisual industry with a series support measures1 (2001-2005). In this case, the constraints from institutions are at a relatively low level since the

1

(16)

producers, distributors and exhibitors are supported by different organizations and the movie market is more open due to the regulations and policies. From another perspective, institutions are families and organizations in view of the social contagion and constitute (DiMaggio & Powell, 1983). Given a social environment, the structure of families, the characteristics of industrial organizations, as well as their market and social behaviours determine the level of constraints from the society. For example, China has relatively small family size and fewer domestic film producers, which limit the interactions within a family and between families (number of contacts), also limit moviegoers’ access to new products. In this case, the constraints from institutions are at a relatively high level.

Motion Picture

Considering motion pictures, they are differentiated from each other in terms of content, style, genre (Bordwell & Thompson, 2001; Zuckerman & Kim 2003), production and/or marketing budget (Geer, 1998) and the degree of ‘artistic’ versus commercial qualities (Bagella & Becchetti, 1999; Hennig-Thurau et al. 2001; Baumann, 2002) etc. Thus, we should consider whether types of movies are associated with the influence of WOM activity and CC behaviour on attendance likelihood. Because moviegoers do have different preferences for a movie type (Austin & Gordon, 1987). As Hennig-Thurau et al. (2001) assumed, a moviegoer is more attracted by certain movies than the others. In view of the movie type preference, moviegoers’ social interactions should be considered in line with the type of movie.

(17)

with higher degree of artistic versus commercial qualities. On the other hand, mainstream movies are major (Zuckerman and Kim, 2003) and large (Krider and Weinberg, 1998) movies, and quite often defined as feature films (Hirschman & Pieros, 1985). This type of movies normally has larger production and marketing budgets and uses more new technologies. Relative to art house movies, they are normally with lower degree of artistic versus commercial qualities and have the participation of movie stars or presence of special effects (Geer, 1998; Bagella & Becchetti, 1999). In view of the ambiguous definitions for both art house and mainstream movies, in the research, we make the distinction by examining the ‘market role’ of the cinemas where certain movies are released, as Zuckerman and Kim (2003), Gemser et al (2005) did in their studies. Since the market role of a cinema is associated with its specialization, many cinemas around the world can be classified into two groups: art house and mainstream. An art house cinema specializes in those movies made by independent, low-budget filmmakers without the resources to use specific effects, studio sets, stars etc. A mainstream cinema is interested in promoting those movies with high studio costs, using many new technologies, with the participation of movie stars etc (Eidsvik, 1991).

2.2. Conceptual Framework and Research Hypotheses

In order to answer the research questions and on the basis of those previous studies, we create a research model in terms of the impact of social environments and movie types. In this model, we explain the relationship between the social interaction and a social environment by the level of social constraints; and explain the relationship between that and a movie type by the level of information supply.

2.2.1. Social Environments Influence Social Interactions

(18)

external condition that affects people’s social behaviour in this research. The different social environments are defined as different countries, which impose different levels of cultural and institutional constraints on WOM activities and CC behaviours.

Cultural Norms and Level of Constraint

As one aspect of the constraints, cultural norms are rules of procedures that individuals use to assure that their behaviours are reasonable (DiMaggio & Powell, 1991). These norms originate from public opinion, educational systems, professions, ideologies etc. (Scott, 1987) and work as unwritten rules of proper social conduct to which people in this society adhere (Handelman & Arnold, 1999). With this respect, cultural norms might moderate moviegoer behaviour, concerning with whom and how much to interact. In view of the difference between cultures across countries, the constraints from such social environments are supposed to be at the different levels. Thus, the extent and the characteristic with which moviegoers exchange information on movies and attend cinemas might vary from country to country due to the level of constraints. Therefore, we categorize countries into the low or high constraint group in line with Hofstede’s cultural dimension scores. Due to the focus on information exchange (WOM) and way of participation (CC), we only pay attention to three characteristics advanced by Hofstede, which are Power Distance, Individualism and Long-term Orientation.

(19)

indicative of more coordinated movie consumption between individuals (e.g. with more people attending the cinema together). In this culture, less powerful members of institutions (like children in a family) accept the power inequality. Thus, individual behaviours could be constrained by more powerful members in the institution (like parents in a family). Children normally attend the cinema together with their parents instead of with other children from neighbourhood. Then a child attends the cinema together with maximum two persons. In line with the culture dimension scores summarized by Hofstede in Table 2.1, for instance, the Netherlands has 38L, a relatively low Power Distance score and China has 80H, a relatively high one. They represent a low and a high constraint case respectively.

Individualism measures the degree to which people prefer to act as individuals than as members of groups. With high Individualism scores, cultures tend to accept and endorse individual initiative (Hofstede, 1980). In this case, individuals more commonly express their own ideas, opinions (Dawar, Parker and Price, 1996). Therefore, high Individualism scores indicate moviegoers are more frequently involved in information exchange activities (e.g. more frequent WOM communication) with less constraint from the cultural norms in the society. On the other hand, as McCarty and Shrum (1994) suggested, collectivism implies cooperation and consideration of the group rather than the individual. In this regard, individuals in low Individualism cultures (i.e. high collectivism) tend to participate in groups. Individual behaviour is more constrained by the group consideration. Based on the scores in Table 2.1, taking the same example, the Netherlands has 80H and China has 20L, which implies more information exchange in the Netherlands and more cooperation behaviour in China.

(20)

individuals are very important. Therefore, the opinions and experiences of others are valued and exchanged in details between individuals (e.g. detailed WOM conversations). Moreover, those with Long-term Orientation cite trust and reliability as sources of power. Individuals are constrained by the perspective for the future relationships. Thus, there are more CC behaviours between individuals since an individual trust all members in the group and accept their behaviours.On the opposite, countries with low Long-term Orientation scores (i.e. short-term orientation) are less constrained by persistence and emphasize rules as a source of information and credibility. In this case, individuals respect for traditions and tend to value impersonal information (e.g. film reviews on magazines). Thus, in such cultures, moviegoers exchange less personal experiences and opinions with others (Hofstede, 2003). As illustrated in Table 2.1, for the Netherlands, the score is 44M, which is quite low relative to 118H for China.

Table 2.1: Culture Dimension Scores for Eight Countries

H = top third, M = medium third, L = bottom third (among 53 countries and region for the first four dimensions; among 23 countries for the fifth)

Country Power Distance Individualism Long-term Orientation

Germany 35L 67H 31M France 68H 71H 30L The Netherlands 38L 80H 44M Hong Kong 68H 25L 96H Indonesia 78H 14L 25L China 80H 20L 118H Japan 54M 46M 80H USA 40L 91H 29L

Source: Geert Hofstede Resource Page (www.geert-hofstede.com 2002, 2003)

(21)

Collectivism and Long-term Orientation, those countries are in the group with the higher level of constraints. The individual behaviour in these cultures is constrained by the powerful members in the group, the agreements within the group, as well as the long-term relationship. China is the example in this case.

Institutions and Level of Constraint

As another aspect of the constraints, institutions are cognitive, normative, and regulative structures and activities that give stability and meaning to social behaviour (Scott, 1995). In view of historical, cultural, political and social elements, institutions are built differently across countries and then provide different levels of stability and meanings to social behaviour. As a result, how to categorize countries into the low or high constraint group could also be considered in terms of three carriers of institutions. They are families, industrial organizations, and government policies on film (Scott, 1995). All of them are associated with moviegoers’ interactions directly or indirectly.

First, as a movie consumption unit, the family should be considered in terms of the average size across countries since a big family tends to have a greater extent (with more family members) of CC behaviour between family members. For instance, more children attend the cinema with their parents. In this case, the more the kids are in a family, the greater the extent of CC this family might have. With regard to different average family sizes in countries, especially due to some policies, such as Family Planning Policy2 in China, moviegoers in those countries with a relatively small average family size should have less extent (with less accompanists) of CC behaviours. Meanwhile, in view of the strong relation within a family (Assael, 1987), trust and reliability between family members lead to more information exchange within a family. Therefore, individuals in countries with a relatively small average family size are more constrained by number of contacts within the family. In Table 2.2,

2

(22)

we still take the Netherlands and China as the examples3. On average, Dutch families are bigger than Chinese ones. In particular, the average family size in the Netherlands has increased marginally in recent years (Van Nimwegen, et. al, 2003) while it has decreased in China due to the policy.

Second, in terms of the individual access to movies, industrial organizations are indirectly related to WOM activity and CC behaviour. Along the value chain for motion pictures, producers, distributors and exhibitors are treated as the main industrial organizations in our case. The more producers, distributors and exhibitors are involved in a national movie market, the more movies individuals could consume, which means a better access to movies. Consequently, moviegoers could exchange more information on movie performance and have more choices when they attend the cinema with others. However, the level of involvement4 of those industrial organizations varies across countries. That influences how many and what kind of movies the individual moviegoers could consume. In Table 2.2, we find that the Netherlands and China have almost the same amount of film producers (214 vs. 200). However, in Chinese movie market, there are much more distributors (2000 vs. 22), cinemas and screens (i.e. exhibitors, 6343/7500 vs. 170/553) than those in Dutch market. If we take the market size into account, the Netherlands will have a higher real number of distributors and exhibitors than that in China. Since we target moviegoers in different countries and any individual could be a moviegoer, population is an indicator of the market size. Then we find that only 29,485 moviegoers share one screen in the Netherlands while 171,000 inhabitants have to share one screen in China (in Table 2.2). China provides a much bigger movie market than the Netherlands. Therefore, the real involvement of Dutch industrial organizations is greater than that in China. Especially, the new releases of films (absolute number, in Table 2.2) in the Netherlands in 2004 are almost two times (307 vs. 160) than those in

3

The 2004 key figures of motion picture industry are considered as the indicators of three carriers of institutions.

4

(23)

China, which implies the effectiveness of Dutch film producers, distributors, and exhibitors. Consequently, the Netherlands provides a better access to movies for consumers.

Third, government policies on film also indirectly affect moviegoers’ social interactions concerning the access to movies. With regard to the development of motion picture industry, countries are in different stages. Therefore, government film policies could aim to increase the output and improve the quality of films; expand the market for domestic films; promote domestic movies abroad and foreign movies at home; fund the film production and encourage film investment; such as the policies in most European countries due to the dominant market share of Hollywood films in the European movie market5. Even the policies from European Union, like MEDIA Programme, MEDIA II and MEDIA PLUS, also aim to strengthen the competitiveness of producers, distributors and exhibitors by some support measures. Meanwhile, government policies could also aim to lessen the government involvement and intervention in the motion picture industry in order to decentralise the power to film producers, distributors and exhibitor, for instance the Reform Policies in those transferring countries. Like in China, the monopoly of one state-owned company (China Film Corporation) in the movie market was broken due to the decentralization of administration. Relatively, these policies are stricter and more concern the whole industry than those in developed European countries. In line with the data of two example countries in Table 2.2, the Netherlands has much more new foreign films released (276 in 2004, which is 89.9% of total new releases) than those released (40 in 2004, only 25% of the total new releases) in China. Especially, every year only 40 foreign films are imported into Chinese movie market because of the Film Import Policy (Ba, 2005). Besides that, the Dual-Administration of SARFT (the State Administration of Radio Film and Television) and the Ministry of Culture limits the contents of production. Therefore, moviegoers in China have the worse

5

(24)

access to movies. The diversification, the output, and the quality of movies are influenced by the policies. In this case, moviegoer’s behaviour is constrained by the intervention of the government.

Table 2.2: Key Figures of Motion Picture industry in the Netherlands and China in 2004

2004 Key Figures The Netherlands China

Population(millions) 16.3 1,296.5

Average family size (parents with children) 4 3 Gross National Income per capita (EUR) 25,289 1,029

New releases of films (yearly) 307 160 Number of domestic film producers 214 200*

Number of distributors 22 2000* Number of screens/cinemas 553/170 7500/6343 Number of inhabitants per screen 29,485 171,000 New releases of foreign films (yearly) 276 40 Foreign films/releases of films Ratio (%) 89.9 25

Source: World Bank; Netherlands Film Fund; Population Trends and Family Policies in the Netherlands (Van Nimwegen, et. al, 2003); 2005 Chinese Motion Picture Industry Market Analysis Report; 2005 Chinese Motion Picture Industry Investment Analysis Report (Ba); The Fifth National Population Census 2001, National Bureau of Statistics of China (www.stats.gov.cn)

*: In China, some figures are approximate numbers.

Concerning the family, the involvement of industrial organizations and the government policies on film, countries could be categorized into the low or high constraint group. Average family size in a country could be constrained by political, economic and social elements. The more the constraints from a society, the more responsibility parents should take within a family, and then the smaller the family size could be6. In this regard, countries with a small family size are in the high constraint group, vice versa. Then, for industrial organizations, their involvement affects the supply side of the movie market, which interacts with moviegoer’s market and social behaviour in terms of the access to movies. In those countries where industrial

6

(25)

organizations are not freely be involved in the value chain for films, the level of constraints is relatively high. In the end, for government film policies, those countries with relatively strict film policies more focus on authorizing industrial organizations, stimulating the development of the whole industry and protecting domestic producers. Consequently, the access to various and high quality movies is constrained.

Constraints on Social Interactions

In order to understand how the level of constraints in a social environment influences a moviegoer’s WOM activity and CC behaviour, we use Table 2.3 to summarize all cases analysed above in terms of three culture dimensions and three institution carriers.

Table 2.3: Effects of Constraints on Social Interactions

Effect of Low Constraint Effect of High Constraint Constraints WOM CC WOM CC Power Distance + + – – Individualism + – – + Long-term Orientation – – + + Family Size + + – – Involvement of Industrial Organisations + + _ –

Degree of Government Interference + + – –

In Summary Positive Positive Negative Negative

(26)

Being derived from the reasoning in section 2.1.1 and 2.1.2, we propose the following two sets of hypotheses:

H1: Different social environments (i.e. nation) influence WOM activities of individual moviegoers.

H1a: Individual moviegoers’ involvement in frequent WOM activities is negatively related to the level of constraints they experience in their social environment (i.e. nation).

H1b: Individual moviegoers’ communication with people (number of contacts and level of influence) is negatively related to the level of constraints they experience in their social environment (i.e. nation).

H1c: Individual moviegoers’ involvement in detailed WOM conversations is negatively related to the level of constraints they experience in their social environment (i.e. nation).

H2: Different social environments (i.e. Nation) influence CC behaviours of individual moviegoers.

H2a: Individual moviegoers’ movie attendance concerning the number of accompanists is negatively related to the level of constraints they experience in their social environment (i.e. nation).

H2b: Individual moviegoers’ movie attendance concerning the type of accompanist is related to the level of constraints they experience in their social environment (i.e. nation).

2.2.2. Type of Movie Influences Social Interactions

(27)

individuals have with movies might be influenced by the type of movie. In another word, moviegoers attend the cinema by different reasons in terms of the type of movie, which results in a certain level of involvement with a type of movie. As some researchers (Rothschild, 1979; Zaichkowsky, 1985) concluded that when consumption of a product or a service, is considered to be important to an individual’s ego, self esteem, or needs and interests, then a high involvement state is likely to exist, vice versa. Since information searching and processing are the consequences of the involvement (Rothschild, 1984), we expect moviegoers for different types of movies are likely involved in different levels of information searching and processing activities. Therefore, we take the type of movie into our research model and examine its effects on the social interactions.

Supply of Information and Type of Movie

In line with the definition of art house and mainstream movies in the previous part, we characterize two types of movies by the level of information supply.

On one hand, art house movies have a relatively high degree of artistic quality with a relatively low marketing budget. Since fewer resources are used to “dress up” the product, moviegoers get relatively less information on movie performance before they see the movie. For instance, there are fewer advertising campaigns on the movie. Moreover, due to the high degree of artistic quality, this type of movie requires the consumers have sufficient cultural competences to decode and understand it (Barnett & Allen, 2000). Not all moviegoers are able to decode the movie in a right way. Therefore, fewer comments on movie performance are given by those who have seen the movie. In this case, we categorize art house movies into the low level of information supply group.

(28)

produced by new technologies and promoted by large advertising campaigns. As a result, individual moviegoers get relatively more information on movie performance before they see the movie. In view of the high degree of commercial quality, individual’s everyday experience is the sufficient ground for understanding and appreciation (Shrum, 1991). Thus, more comments on movie performance are available so that we could categorize mainstream movies into the high level of information supply group.

Level of Information Supply and Social Interactions

As explained, the cinema attendance for art house movies is basically related to a moviegoer’s self-concept, personal values and cultural capital (Bourdieu, 1984; Barnett & Allen, 2000). Meanwhile, this type of movie provides a low level of information supply. In order to gain the competences for understanding the movie, consumers are more likely to be involved in information searching and processing activities, such as more frequent communications with others and discussions on movies, as well as the information exchange with more people. Moreover, moviegoers for art house movies are normally motivated by their own interests; the lack of information stimulates their searching for further information to fulfil themselves. Furthermore, in view of the important role of self esteem and personal value for them, moviegoers for art house movies attend the cinema with less people in one time than those for mainstream movies. This is because they are less willing to give up the things important for themselves.

(29)

family and social activities, which limit their searching for further information to have fun together with family members and friends. In addition, moviegoers for mainstream movies normally attend the cinema with more people in one time than those for art house movies due to their entertainment purpose and less requirement on cultural competences.

Based on the understanding on the relationship between the level of information supply, type of movie and social interactions, we propose another four hypotheses as follows:

H3: Different types of movies influence WOM activities of individual moviegoers. H3a: Individual moviegoers’ involvement in frequent WOM activities in negatively related to the level of information supply in the type of movie they choose.

H3b: Individual moviegoers’ communicate with more people (number of contacts and level of influence) is negatively related to the level of information supply in the type of movie they choose.

H3c: Individual moviegoers’ involvement in detailed WOM conversations is negatively related to the level of information supply in the type of movie they choose.

H4: Different types of movies influence CC behaviours of individual moviegoers.

H4a: Individual moviegoers’ movie attendance concerning the number of accompanists is positively related to the level of information supply in the type of movie they choose.

H4b: Individual moviegoers’ movie attendance concerning the type of accompanist is related to the level of information supply in the type movie they choose.

(30)

Figure 2.1: Conceptual Framework

Social Environment

(at the country level) Cultural Norms

Institutions

Type of Movie

(at the industry level) Art-house Movie Mainstream Movie Level of Constraint Low Constraint High Constraint Supply of Information Low Supply High Supply Social Interactions Word-of-Mouth Activities Coordinated Consumption

As shown in Figure 2.1, for social environments, cultural norms and institutions are examined at the country level, as a source of constraints from social environments. Therefore, countries are categorised into two groups by the level of constraint. That is, countries with low constraint and those with high constraint. In view of the effect of those constraints on consumers’ behaviour, we suggest moviegoers have different social interactions in terms of WOM activities and CC behaviour across counties.

(31)

CHAPTER 3: RESEARCH METHOD

3.1. Introduction to Research Method

In this chapter, we will introduce how this research is done. Considering our focus on the comparison for the effect and the characteristics of social interactions, given by different social environments and different types of movie, the research starts with the selection of the country for the social environment, and the identification of type of movie. Then, we define the target population of this research and explain how the sample frame and the sample size are identified. Moreover, we show how the survey is designed and conducted. Furthermore, the research variables employed in our model are explained in details. In the end, data analysis demonstrates how we test hypotheses.

3.2. Selection of Country for Social Environment

(32)

institutional constraints in these two countries could be categorised into the low and the high constraint group respectively. Moreover, the market shares in both countries are dominated by U.S. films, which provide a common ground for them to compare. In the end, the last reason for the selection is about the convenience of data collection. It is possible to make the research survey in these two countries.

3.3. Identification of Type of Movie

To identify the type of a movie, we use the approach similar to Zuckerman and Kim (2003), as well as Gemser et al (2005), to examine the ‘market identity’ of the cinemas. A cinema is identified as an art house if it is specialised in showing movies with relatively low budgets, with higher degree of artistic qualities, or for minority taste (e.g. Romance and Cigarettes, Transamerica, The White Masai7). On the contrary, a cinema is a mainstream movie cinema if it mainly shows movies with relatively high budgets, with higher degree of commercial qualities, or for majority taste (e.g. Poseidon, Pirates of the Caribbean, The Da Vinci Code). Therefore, based on the identification of a cinema, we categorise the movies concerned in this research. In the Netherlands, Cinema Image in Groningen8 is one art house selected in this case since it mainly shows foreign movies, or movies with relatively higher artistic qualities. Concerning mainstream cinemas, Cine City in Hoogezand and De Bios in Drachtern are selected to collect mainstream movie data. These two cinemas normally show movies for majority taste. In China, most cinemas at the city centre or close to a shopping centre show movies with relatively higher commercial qualities since they target moviegoers seeking entertainment. However, those cinemas close to universities mainly show movies for minority taste, and target moviegoers seeking high arts instead of popular culture. For those high educated moviegoers, cultural competences required for understanding art house movies are satisfied (Austin, 1986).

7

The detailed list of the films in the sample is presented in Appendix A.

8

(33)

As a result, two cinemas at the city centre of Nanjing9 and another two beside the universities are selected respectively for collecting mainstream and art house movie data.

3.4. Target Population

The target population of our research is people who are attending the cinemas during the time period data is collected. People who want to see a movie, but still do not make their decisions to go to the cinema are not included in our moviegoer population. The reason for choosing people at the cinema is that they have consumed the goods (movies). Thus, we are able to examine the positive influence of the WOM communication and the CC behaviours in terms of the movie attendance. In this way, we integrate the social interactions of moviegoers into the modelling of their market behaviours.

3.5. Sample Frame and Size

In order to observe and compare WOM activities and CC behaviours across countries, the sample is formed by moviegoers visiting three cinemas in Groningen, Hoogezand and Drachten (Netherlands), as well as those visiting four selected cinemas in Nanjing (China). The selection of these cities is based on the city size10, the availability of the field survey and the convenience11. Therefore, our sample is non-probability sampling12, which allows field workers to choose whoever has seen a movie or is going to see a movie at the cinema. The sample size is considered so that we know at least how many respondents should be included into the sample. A

9

Nanjing is a middle-scale city, the capital of Jiangsu province in China.

10

Such as Nanjing, Groningen is also a middle-size city in the north of the Netherlands, the capital of Groningen province.

11

Hoogezand and Drachten are relatively small concerning the city size. However, we only got the permission to make the field survey in two cinemas in these two cities. Especially, they are not far away from Groningen (within 40 minutes by bus).

12

(34)

sufficient sample size for the research can be calculated depending on the value of the population proportionP, the standard error of the proportion

p

σ and the confidence level required (Anderson, Sweeney and Williams, 2002, pp.267). In our population,

is the proportion of the observed and expected moviegoers at the cinema who are involved in WOM activity and have CC behaviour with regard to gender, age and education. Since

P

p

σ measures the deviation of the sample proportion from the population proportion, we should expect a low level of the error, which is supposed to be within the interval range of ±5% in our sample, then

p

σ is ± 0.05. Furthermore,

is unknown, and the estimate

P p is also unknown, thus we take such that (1– ) as large as possible. In this way, we calculate the sample size in terms of the worse case scenario, that is, equals 0.5. As a result, such sample size is sufficient for all other cases in order to get enough respondents at the cinema for the research. Therefore, based on the expression

P P P

P

13

in Anderson’s textbook, our sample should include at least 400 respondents while 90% confidence level is required in our research. That says there is a 90% probability that the sample proportion could estimate the population proportion with a less than 5% standard error; thus Z is 2 for calculation. Considering the calculated sample size and the group comparison between countries, we require a sample of at least 400 subjects in each country.

3.6. Fieldwork: Questionnaire Survey

The subjects of this study are human beings who often do not behave in their usual manner when they realise being watched. In order to discover how consumers usually behave, we conduct questionnaire survey in the fieldwork (cinemas) to collect

13

For infinite population, the expression for computing the standard deviation of Pis:

p σ = z

n p p(1− )

; then we transform it into:

(35)

raw data. According to Reinard (1998), questionnaire is one survey form in which individuals respond to written items to report their attitudes and understanding of things, including their behaviours. Our questionnaire is designed for those who are attending the cinema for a movie. All questions are formulated in four sections.

The first section (question 1 to question 3) contains general information about the respondents: their age, gender and education. By these questions, we are interested in finding out whether there is a particular trend for both young and old moviegoers or both male and female respondents, as well as whether the level of education makes the respondents have different consuming behaviours.

The second section (question 4 to question 9) focuses on an individual’s opinions about WOM activity and CC behaviour for a particular movie. Based on these questions, we distinguish the type of a movie by the title (question 4). Also we could know the time people making the choice for a particular movie (question 5), the effect and the characteristics of CC behaviours (question 6 and 7), as well as the effect of WOM activities concerning recommendations from others (question 8 and 9).By the question “with how many people attending the cinema” (question 6), we measure the effect of coordinated movie consumption. Meanwhile, the question “with whom people attending the cinema” (question 7) observes the characteristics that people consume the movie together in one time. Since social relations normally include the relations with partner, with friends, with family, and with colleagues, this question is asked in line with these dimensions. Furthermore, by question 8 and 9, we measure WOM communication in terms of the number of contacts.

(36)

case. They show our interests on whether CC exists in general and with whom people interact, as well as how many people are coordinated in terms of CC (question 11 and 12). In the meanwhile, we pay attention to people’s recommendation to others, (question 13), and the influence of people’s general discussion on movies (question 14). In particular, we formulate one question that provides the frequency of visiting the cinema in one year (question 10) as the indicator of movie consumption in this research.

The fourth section (question 15 to question 28) aims to detect how moviegoers are motivated and how much they are involved in WOM activity and CC behaviour, as well as from which source of information they know about the quality of a movie. For instance, question 17, 18 and 19 are a set of questions, which could provide us the most general source of information about movie quality. Meanwhile, according to the measures and the antecedents of WOM summarised by Walker (2001), some questions concern the depth and the frequency of WOM activities (question 15 and 28), as well as the examination on the role of movie quality (question 16). In addition, we are interested in people’s motivation for movie attendance. We want to know whether there is a trend in the movie market for both the consumers motivated by personal needs and interests (question 20, 21, 22, 26, and 27) and those motivated by entertainment and social experience (question 23, 24 and 25). Especially, we want to find out whether moviegoers with different motivations are involved in the WOM activities and CC behaviours differently while seeing different types of movies.

(37)

3.7. Data Collection

In order to maintain the data validation, all questions were pre-tested before fielding the main study both in the Netherlands and in China. In the pilot study, in total 150 questionnaires were distributed in two countries and each country got 75 filled-in questionnaires back. Based on those answers, we made two changes. First, we divided 28 questions into four sections, especially separating the specific and the general cases of movie attendance. In this way, we avoid the misunderstanding on similar questions in different cases. Second, we changed the measurement of question 14 (i.e. general discussion on movies) from a number to categories since we found many answers were unclear, such as “discuss with many people”, instead of a number. Therefore, we made four categories like Less than 5, 5 to 9, 10 to 15 and More than 15. The final version contains four sections presented in Appendix B.

The questionnaires were distributed in two countries in the same period (i.e. a period of a new release) rather than on the same day since we consider there are different release schedules in two different countries. In general, the first show of a movie is on Thursday in the Netherlands. However, in China, the first show is generally on Friday, Saturday or some special festivals (e.g. Christmas Eve and Valentine Day). In this case, we just focus on the start of a new release period for some movies. We started the survey in the Netherlands on 7th June, 2006, then, the field worker went to the selected cinemas a few times a week, which was in line with the schedule of shows at the cinema and the convenience of field workers. This questionnaire survey took five weeks to its collection to date. In China, the survey also took around five weeks from its distribution on 4th June, 2006 to its collection to date. In order to make our sample be a good representative of the target population, the field workers tried distributing the questionnaires to the moviegoers for different movies14.

14

(38)

With the requirement of 400 respondents in each country, we planned to collect at least 200 sufficient questionnaires15 for each type of movies in each country. In total, we have four subgroups. Hereby, we have equal amount of respondents for each subgroup. However, generally speaking, more people visit mainstream movies than those at the art house because it is much easier for average moviegoers to reach and understand mainstream movies, in view of the low budget, minority taste and high artistic quality features of art house movies. With respect to the proportions of visitors to two types of movies in the real world, we consider whether a proportional sampling is necessary. Since we are interested in comparing moviegoers’ WOM activity and CC behaviour in each subgroup, rather than observing the average WOM activity and CC behaviour of the whole population, it is more important for us to get equal amount of respondents for each group, than taking proportional sampling into our case.

In total, we got 900 questionnaires back, including 19 questionnaires with missing values. By excluding those incomplete ones, 881 questionnaires constructed our sample in two countries. The bias from the exclusion of some useful information was accepted in this study since this kind of bias is inevitably present in surveys of human populations16. 452 copies were collected in the Netherlands where 215 copies were collected at the art house and 237 copies were collected at the mainstream cinemas. In China, 429 copies are collected, which include 211 copies for art house movies and 218 copies for mainstream movies. This portion difference might diminish statistical power of the estimators, such as average means and standard errors. However, in order to avoid the bias from the exclusion of useful information, we took all respondents into account.

15

In this case, a sufficient questionnaire means there is no missing value in this questionnaire. In order to get exactly 200 sufficient questionnaires, we distribute more than 200 copies so that we assure the correspondence rate of questionnaires is 100 percent.

16

(39)

3.8. Research Variables and Measurements

Based on the conceptual frame work, we expect the effects of WOM communication and CC behaviour, as well as the characteristics of CC behaviour are related to the social environment and the type of movie. In order to test the hypotheses of the study, we define two kinds of social interactions as the dependent variable in our model. Since we measure WOM activity and CC behaviour from different aspects, in different cases, and with different types of data, ten dependent variables of interest are defined separately. On the other hand, considering the influence of social environments and types of movies, we use them as two independent variables to explain the variation in the dependent variables. In addition, we also employ some control variables to increase the power of analysis.

Dependent Variables

Firstly, for assessing WOM activity, six items are used in line with Walker’s study (2001). He summarised and examined three aspects of the WOM communication that was suggested by Churchill (1979). The first aspect is enthusiasm, which is measured as the frequency (how often WOM takes place), and the number of contacts, including both people contacting with and contacted by a consumer (Anderson, 1998; Brown and Reingen, 1987; Reingen and Kernan, 1986). Therefore, in our research, WOM117

scales the frequency of people’s discussion on movies; and WOM2

3

4

5

gives the number of people who recommend the movie to a moviegoer. Here, the moviegoer is treated as the receiver of the information. Meanwhile, WOM measures the effect of the recommendation from those people, which is about how much a moviegoer could be influenced by the recommendation. Moreover, WOM records the number of people to whom a moviegoer intends to recommend the movie. In this case, the moviegoer is the finder of the information. Furthermore, WOM estimates one general case of the WOM communication, which indicates the number of people with whom a moviegoer generally discusses movies.Then, the second aspect is about how much is said in the

17

(40)

WOM communication (Bone, 1992). We define WOM6 to scale the depth of WOM

communication, indicating how detailed the conversation on movies might be. The third aspect is praise, which means the favourableness of the WOM communication (Arndt, 1968; Burzynski and Bayer, 1977; Herr et. al., 1991; Singh 1990). With this respect, people would like to give information and prefer to give positive comments. There is no clear measure on praise; instead we only ask the questions with favourableness of WOM or some questions with neutral attitude on WOM, such as question 13 and 14. Furthermore, we target those who are attending the cinema. These people are more likely told about positive experience, information or evaluation regarding the movie since they have made the purchasing decision.

Concerning the measurements, the responses to WOM1, WOM3 and WOM6 are based on a 5-point Likert Scale, anchored by Totally Disagree/No Influence At All (1) and Totally Agree/Very Strong Influence (5) 18(Blumberg, et. al., 2005). Being different from Walker’s study (2001), we use 5-point scale instead of 7-point scale in view of the advantages of ease of administration, ease of coding and analysis, and ease of understanding. As Levine and Baden (1978) suggested, a short scale gives as much information as a long scale. In that study, they converted 7-point scales to 5-point scales and compared the values obtained with the 5-point scales; they found the compare scales were not significantly different. The responses to WOM2, WOM4 and WOM5 are based on numbers.

Secondly, in order to measure CC behaviour, four items are employed to assess the effect and the characteristics of CC in a specific and a general case respectively. E_CC1 measures the effect of CC behaviour for a certain movie by the number of people attending the cinema together in one time. P_CC1 indicates the characteristics of CC behaviour for a movie, which is assessed by the type of accompanist, like alone, partner, friends, family and colleague. Then, another two items, E_CC2 and P_CC2

scale the same variables for a general case. Data for each item is categorical and

18

(41)

coded by four (0 for Alone to 4 for Colleague) or six numbers (0 for Alone to 6 for more than 5 people).

Independent Variables

Two independent variables of interest are defined. One is the NATION and another is the Movie Type (M_TYPE). Both are categorical and coded using a dummy variable. In terms of NATION, the dummy variable has a value of 0 if the data are collected in the Netherlands and a value of 1 if collected in China. Concerning Movie Type, the dummy variable has a value of 0 if the respondent attends the cinema for an art house movie and a value of 1 if for a mainstream movie.

Control Variables

Next to nation and movie type, we also use control variables. They do not relate to the independent variables but explain significance variance in the dependent variables, help remove predictable variance from the error term and help prevent omitted variable bias (Tabachnick & Fidell, 1983). These control variables are the general information about the respondents, the time of decision making, the frequency and the motivation of movie attendance, as well as the belief in information from WOM communication and the attitude on movie quality.

(42)

level professional school, high level professional school and university19. Five categories are coded by five numbers (0 for primary school to 4 for university).

The time of decision making (D_TIME) assesses whether longer decision making process is associated with more WOM inflow (other’s recommendation through conversations). As Gatignon and Robertson suggested (1986), consumers under high time pressures are less likely to be involved in WOM conversation. We expect shorter decision making process is related to less WOM conversations. The data for this variable is categorical and coded by five numbers (0 for last month to 4 for now, at the cinema).

The frequency of movie attendance (FREQ_ATT) is included to examine whether frequent visiting leads more frequent and detailed WOM communication, and whether moviegoers are more likely to exchange information on movies if they frequently visit the cinema. As Dichter (1966) observed, one’s frequent and/or intensive involvement with a product creates a willingness to engage in WOM about the product. This variable is scaled by the times of visiting cinemas within one year.

The motivation of movie attendance examines whether moviegoers with different motivations engage in WOM activity and CC behaviour at different levels and with different characteristics. Based on Kelly and Freysinger (2000)’s study, individuals could be motivated by seeking entertainment and social experience when they consume cultural products; otherwise, they could also be motivated by seeking fulfilment of personal values. Also in Ba’s survey of Chinese motion picture industry (2005), he found moviegoers go to the cinema for entertainment, for social activities, for artistic enjoyment and for the knowledge on individuals and societies. Therefore, three items are designed for entertainment and social purposes in this study. They are

19

(43)

for participating social activities (FOR_SOC), for spending free time (FOR_TIME) and for pure entertainment (FOR_ENT). Another five items are generated for individual fulfilment. They consist of for cultivating own interests (FOR_INT), for the performance of an actor/actress (FOR_ACT), for the work of a director (FOR_DIR), for developing some ideas (FOR_ID) and for being an expert on movies (FOR_EXP). We expect moviegoers with the fulfilment purpose engage in more frequent and detailed WOM communication and are more likely to be involved in WOM conversations. They search and exchange information on movies to develop their own ideas. For those people with the entertainment purpose, we expect a higher level of CC behaviour since they go to the cinema for situations (i.e. for social life, for free time, for fun). The responses to each item are based on 5-point Likert scale, set by Totally Disagree (1) and Totally Agree (5).

The belief in information from WOM (WOM_INFO) and the attitude on movie quality (M_QUA) control extraneous variation in WOM activities. Moviegoers are more likely to engage in frequent and detailed WOM conversations with more people if they believe the information on movies from WOM and concentrate on movie quality. We measure these two control variables also by a 5-point Likert Scale, set by Totally Disagree (1) and Totally Agree (5).

3.9. Operationalization and Data Analysis

3.9.1. Sample Characteristics

(44)

moviegoers are high educated (high level professional school 20.9%, university 47.9%). These characteristics of participants in our sample are summarised in Table 3.1 below.

Table 3.1: Sample Demographics

N = 881 Gender Male Female 360 (40.9%) 521 (59.1%) Age <18 18-35 35-50 >50 100 (11.3%) 561 (63.7%) 153 (17.4%) 67 (7.6%)

Education Primary School

High School

Middle Level Professional School High Level Professional School

University 8 (0.9%) 161 (18.3%) 106 (12.0%) 184 (20.9%) 422 (47.9%) 3.9.2. Exploratory Factor Analysis

Next to those demographic variables, we also have the motivation of movie attendance as the control variable in this study. Eight items are used to measure two kinds of motivations in terms of entertainment and fulfilment purposes. In order to assess the unidimensionality20 of the multi-item scales for the motivation, we use the factor analysis. Items with factor loadings of greater than 0.50 on the factor with they are hypothesized to load are considered adequate indicators of that factor. Based on the assumptions of factor analysis, the correlations among items should be significant, which is indicated by the MSA (Measure of Sampling Adequacy) value. The minimum acceptable MSA level is 0.5 for each item (Hair et. al, 1998, pp.122). Both the overall (for all items) and the individual (for each item) MSA levels indicate sufficient correlations among items for the factor analysis.

The principle component factor analysis (in SPSS for Data Reduction) was performed

20

Referenties

GERELATEERDE DOCUMENTEN

Zo zal bij hoge mastery-approach doelen, een hoge mate van extraversie een positieve invloed hebben op de relatie tussen mastery-approach doelen met voice-gedrag en een lage mate

Gezien het feit dat er onder Dutch GAAP door de meeste ondernemingen uit het sample de goodwill direct ten laste van het eigen vermogen of het resultaat wordt gebracht

The aim of this research is to identify the perceptions of child counsellors (CCs) and members of the Yabonga board on the definition of Orphans and (other) Vulnerable

These systems are highly organised and host the antenna complexes that transfer absorbed light energy to the reaction centre (RC) surrounding them, where the redox reactions

H1: Regardless of the valence, a review written by a professional critic has a stronger effect on moviegoers intention to see a movie in the cinema than a OCR written by an

This study was about finding out the effect that a prior notification about product placement would have on consumer’s brand awareness, brand attitude and

International journal of quality and reliability management, 14(9). Out of the crisis. Cambridge: Cambridge University Press. DOE see SOUTH AFRICA. Department of

Since Chinese businessmen generally actively participate in this system, studying the economic activities of Chinese tax farmers in the tax farming system will help us understand