ABSTRACT
The popularity of the children’s film genre has grown and become a much more lucrative business the last few years. This research aims to investigate to what degree Word‐of‐mouth (hereafter WOM) may influence the decision process of children for watching a particular children’s film and the degree to which these children produce WOM after watching the film. After a theoretical review, five factors were distinguished that are of importance for the input WOM (WOM received before watching a particular children’s film). These factors are valence (positive or negative messages), source (from who they receive WOM) and how many (amount of WOM received). There are two other factors (control variables): type of film (mainstream or independent film) and with whom (option of with whom they can watch a particular children’s film) are also important to the input WOM. A questionnaire and a Conjoint analysis experiment were used to gather and analyze the information to be able to answer the main and sub questions of this research. Based on the results, this research demonstrates that WOM does indeed have an influential degree on the choice of children for watching a particular children’s film. The results of this research indicate that in general the factors valence and how many are important in the choice for the children when watching a particular children’s film, but without doubt the most important factor is source. Next to WOM, the control variables also have an influence in the decision process of the children to watch a particular children’s film. Furthermore, the sender/receiver characteristics (opinion leadership, susceptibility to normative influence, susceptibility to informational influence, preference for heavily advertised films, gender and age) play a small role in the production of output WOM (amount of WOM produced after watching a particular children’s film). There are little differences in what type of children produce the most WOM among their friends. The children aged 8 are the ones that produce the most WOM, thus sharing their last viewing experience with most of their friends. Key words: children’s film; input WOM; output WOM; valence; source; independent film; mainstream film; opinion leadership; susceptibility to normative and informational influence
ACKNOWLEDGEMENTS
I signed up for the Pathé Unlimited Pass (PUC) at Pathé cinemas, since its introduction around 2005. For only € 19,‐ per month I have unlimited access to regular showings at all Pathé cinemas throughout the Netherlands. This was the reason I started going to the movies more often. By doing so I started to realize which genres were more appealing to me. One of these genres in particular that captured me the most was the children’s genre. While searching for a topic for my master thesis, I decided to better understand this genre, so I chose to investigate this specific film genre. I want to know what factors influence children in their decision process to watch a particular children’s film. A factor that might influence them could be WOM they receive from family and friends. By combining these two aspects I found my research topic. Therefore, this research determined the degree to which children’s WOM influences children’s film.
1
INTRODUCTION
Each year hundreds of new films are released by Hollywood and independent filmmakers (Liu, 2006). They spend millions each year promoting their blockbuster films. Apart from this film promotion, there is one method that they have continued to rely on to draw big audiences ever since the dawn of the film industry: WOM (McLain, 2008; Zigu, 2011).
In order to try and avoid buying tickets to bad films, consumers rely on secondary cues ‐the opinion of critics and WOM reviews from friends who have already seen the film‐ to give them an indication of its quality (Haupert, 2006). Potential consumers become more aware of WOM and so do their opinions and decisions (Babutsidze, 2011). Films that people appreciate and enjoy instantly become popular, as people create a buzz amongst their close ones. People will recommend family members or friends to watch a particular film if they themselves find it interesting. However, even if a film has the biggest actors and the most extensive marketing, if people find it bad they may generate negative WOM, thereby destroying the hopes of the film’s success (Zigu, 2011). This is the case for any genre in the film industry, no matter how small or big. WOM can also create a certain reputation for that film without anyone ever watching it. Input WOM is spread by a sender perceived to be more neutral. Input WOM is often more convincing than any other form of advertising. WOM cultivates the reputation and can persuade filmgoers to watch a particular film.
Children’s film has become a much more lucrative enterprise than it was in the half century of Disney hegemony from the 1930s to the 1980s. As a consequence, there is now much more competition and there are now significantly more players in the field (Booker, 2010).
To further elaborate on the reason why the genre children’s film was chosen, the following facts will be discussed. Ten of the top 25 films released during 2010 were rated G and PG (MPAA, 2011). Ticket sales for G and PG rated films have fluctuated the past couple of years, but have been on a rise again the last few years (for details see Appendix I). Another important factor is that WOM networks among children can be of importance to innovation activities. This has two reasons. Firstly, children possess the need to gain knowledge and use this knowledge in the role of consumers, so they may be involved in the product development process itself. Secondly, children may also strongly influence each other in decisions to adopt innovations (Kratzer and Lettl, 2009). Moreover, children’s WOM is relevant for this genre, since it strongly influences final decision. Children aged 6 to 12 have 72% of a lot or some influence in the film category (Del Vecchio, 1997) (for details see Appendix II).
1.3 STRUCTURE OF THE PAPER
2
THEORETICAL REVIEW
This chapter’s main focus will be on the theoretical review. A definition of WOM will be
given firstly to help gain insight into the term. The different forms of WOM, the degree of interpersonal influence of WOM and sender/receiver characteristics will be assessed afterwards. This chapter ends with the conceptual model.
2.1 DEFINITION OF WOM
There have been many definitions of WOM in literature. Arndt defines WOM as oral, persontoperson communication between a recipient and a communicator whom the recipient perceives as noncommercial, regarding a brand, product or service (Arndt, 1967b). This definition consists of three essential elements (Kirby and Marsden, 2006).
Firstly, WOM is interpersonal communication. This element sets WOM apart from mass communication and other impersonal channels available to consumers. Furthermore, the code of this type of communication is language. Other, less tangible forms do not qualify for the label ‘WOM’ in themselves. An actor promoting his or her new film or the entire cast speaking out about making the film does not automatically fit into the principle of WOM. These practices might accompany WOM, but they are not essential to it, because they are not seen as a close source. This is the reason why it is often considered as advertising. Therefore, this research focuses on interpersonal messages ‐one‐to‐one communication‐ and neglects –one‐to‐many communication‐ for example Facebook and Hyves. Secondly, the producers of WOM are not motivated commercially, even though the content of WOM communication is mostly commercial, or at least it is perceived that way. The producers of WOM are people that do not work for a specific company, they just talk about products or services at their own will. Thirdly, the content of WOM communication is commercial when viewed from a marketing perspective. These commercial messages may have commercial consequences. WOM is viewed as non‐commercial when viewed from another consumer’s perspective. In both cases commercial entities, products, product categories, brands and marketers or even own advertising is often embedded in the message. This restriction in the meaning of the term underscores that WOM is a technical term appropriated for marketing, consumer behaviour and mass media.
2.2 FACTORS INFLUENCING THE IMPACT OF WOM ON FINAL DESICION The following section will highlight the factors influencing the impact of WOM on final decision. Timing (consumer’s perspective): when is a message sent? Valence (content): what message is sent? Source (sender): who is sending the message? Receiver: who is receiving the message? 2.2.1 TWO TYPES BASED ON TIMING
The following sections discuss the timing viewed from the consumer’s perspective as displayed in figure 1. FIGURE 1 Consumer timing perspective
Consumers use WOM for two main reasons; risk reduction and social need. WOM can be differentiated on the basis of which direction the communication flows. Someone who is then considered to be the transmitter of the information does initial recommendations of a product, service or enterprise. WOM might be spread either before or after, in this case, an experience.
It can be of financial, performance, physical, psychological and social risk (Jacoby and Kaplan, 1972). Input WOM (receiver) minimizes such a risk in a way that it is seen as a strategy for risk reduction. In the film industry, this is when filmgoers are advised to go and watch a particular film. Confidence grows stronger, the closer the advice comes to one’s social connection. Information that comes from family members or an individual’s circle of friends is considered more reliable than impersonal sources. Input WOM among consumers is particularly important in the selection process of a product or service (Anton, 2009).
2.2.2 TWO TYPES BASED ON VALENCE In the subsequent sections attention will be paid to WOM valence from the consumer’s reception perspective, as portrayed in figure 2. FIGURE 2 Consumer reception perspective
The valence of WOM might be either positive, neutral or negative (Anderson, 1998). Not many studies have addressed neutral statements, because it is perceived as less informative and these statements are not as important. Neutral statements may not contain a lot of informational insight and many times individuals do not know how to react. These statements have also little or no commercial relevance and they do not greatly influence behaviour. These are the reasons why most studies only describe WOM valence as either positive or negative (Hoyer and Macinnis, 2009). In this research the focus will be on WOM valence viewed from a positive and negative perspective. Positive WOM consists of when positive testimonials and endorsements desired by the company are passed along among individuals in their target group (Buttle, 1998). On the other hand, negative WOM includes behaviour such as product denigration, relating unpleasant experiences, rumour and private complaining (Anderson, 1998). The reason that WOM valence matters is very straightforward; positive WOM enhances an individual’s view on the expected quality, whereas negative WOM will reduce it (Liu, 2006).
This is also known as positive WOM bias. It has to be said that most individuals who have extremely negative WOM share it with fewer people. On the other hand, people with less harsh comments about a product or service, share this more openly with other individuals. This works both ways, as people who are not completely satisfied, but still have a rather pleasant experience with a product or service do not pass this on to as many people. On average, less than very satisfied customers speak to significantly fewer people about their experience than the most satisfied customers (Naylor and Kleiser, 2000). However, more than half of dissatisfied consumers engage in negative WOM. Furthermore, dissatisfied consumers talk to three times more people about their bad experiences than satisfied consumers do about their pleasant experiences (Hoyer and Macinnis, 2009). Burzynski and Bayer (1977) researched the effect of positive and negative WOM on motion picture appreciation. Filmgoers, who were exposed to negative WOM prior to viewing the film subsequently, also expressed more negative evaluations of the film than filmgoers who were exposed to positive WOM. Negative WOM can have such an impact on WOM produced (post) that in some cases filmgoers who are exposed to this negative WOM immediately redeem their tickets (Burzynski and Bayer, 1977). People are eight times as likely to receive positive (favourable) rather than negative (unfavourable) WOM (Arndt, 1967a). However, these results suggest that negative WOM has more effectiveness in shaping others’ behaviour. Sweeney et al. (2005) also supported similar results. They suggest that in consumer evaluation, negative information plays a much greater role than positive information. Moreover, people pay a lot more attention and hold on to negative WOM than they do to positive WOM. Negative WOM may be diagnostic due to the fact that it has more significance because it seems to tell us how information differs from one another. Most of the information people hear about a product or service is positive, so negative information may receive more attention because it is surprising, unusual and different. Additionally, negative WOM may also prompt consumers to attribute problems to the product or service in question, not to the consumer who uses it (Hoyer and Macinnis, 2009).
However, there are mixed findings about the degree to which individuals talk when they are dissatisfied about, in this case, a film. Some studies state that dissatisfied consumers talk very little or not at all about a film, while other studies conclude that dissatisfied consumers will engage more often in interaction to voice their dissatisfaction. 2.2.3 TWO TYPES OF SOURCE BASED ON TYPE OF SENDER From the consumer’s perspective, this section of the paper assesses the transmitter of the input WOM. There are different methods to research the impact of WOM often based on who is sending ‐source and source credibility‐ the message. For example, the study by Gilly et al. (1998) used theoretical constructs such as source expertise, demographic homophily, perceptual homophily, seeker expertise and opinion leadership of source to conduct their research. Another way to research source and source credibility is by applying the theory of the strength of a tie.
The introduction of the concept the strength of a tie proceeded on an intuitive notion in defining the concept in terms of four elements. The strength of a tie is characterized by a (probably linear) combination of the amount of time, the emotional intensity, the intimacy (mutual confiding) and the reciprocal services (Granovetter, 1973). Strength is the emotional or band connection individuals have with one another. A consumer’s social relations with other relevant actors typically include a spectrum of ties that range from strong primary to weak secondary (Brown and Reingen, 1987). More intimate connections between close friends and family members are defined as strong ties; the source is someone who personally knows the decision‐maker. On the other hand, weak ties exist between acquaintances or business associates; someone who is just an acquaintance or does not know the decision‐maker at all (Papakyriazis and Boudourides, 2001; Duhan et al., 1997).
Furthermore, strong ties were not only being perceived as more influential, but they were also far more numerous as sources of information than weak ties (Brown and Reingen, 1987). Weak ties might have a greater impact on the rate of information dissemination than strong ties when network size is small, the amount of weak tie contacts are numerous or the effect of advertising is weak. Both types of source based on the type of sender are relevant and stronger than advertising (Goldenberg et al., 2001).
As a result, both strong and weak ties act as important sources in the dispersion of interpersonal communication through WOM for the film industry. This market dispersion is very important. Children that have watched a film can spread information to family members, close friends or individuals outside their social vicinity. Although it is mentioned above that WOM received from strong ties is perceived as more influential, weak ties are also just as important, because they act as an activator for the flow of information.
2.3 SENDER/RECEIVER CHARACTERISTICS THAT INFLUENCE WOM IMPACT AND OUTPUT WOM
Following the factors that influence the impact of WOM on final decision in the previous sections, these upcoming passages will assess the sender/receiver characteristics that influence WOM impact and output WOM: opinion leadership, consumer susceptibility to interpersonal influence, gender and age. 2.3.1 OPINION LEADERSHIP Opinion leadership has a strong influence on the output WOM. Opinion leaders tend to be, in many ways, similar to those they influence. They are more sociable, innovative and more interested in the topic of the product or service in question. Moreover, they typically receive and transmit more information about a given topic. An opinion leader may be expressing a need to help others make a more informed decision or may simply wish to share a message, such as a particularly funny or controversial advertisement (Haywood, 1989).
He or she is just a channel of information and advice to other people (Brooks, 1957). When opinion leaders possess many direct contacts and are also mutually interconnected, they are considered close to all other group members. Children with a high degree of centrality ‐that is, with many direct contacts to others in social networks‐ can be identified as opinion leaders. However, it is suggested that opinion leadership among children cannot be identified on the basis of betweenness centrality, meaning that opinion leaders are only influential in very local terms. Opinion leaders can spread innovations, but many opinion leaders are necessary in order to achieve a broader reach (Kratzer and Lettl, 2009). 2.3.2 CONSUMER SUSCEPTIBILITY TO INTERPERSONAL INFLUENCE The degree to which WOM influences the decision of children to watch a particular film is highly dependent on their susceptibility to interpersonal influence. Some children may be highly susceptible to social influence and in that case WOM might have a strong effect. On the other hand, incoming WOM might have not much effect on children that are not susceptible to social influence. Individuals with low self‐esteem tend to be more readily influenced than others (Janis, 1954). There are two types of social influence that can be distinguished namely, informational and normative influence (Deutsch and Gerard, 1955).
Informational influence is defined as the acceptance of information obtained from another as evidence about reality (Deutsch and Gerard, 1955). Informational influence is based on the desire to make informed decisions. When faced with uncertainty, an individual will seek information. Of the many available sources, the most likely to be accepted are those that are viewed as credible. Those with presumed expertise or significant others are considered as referents with high credibility (Bearden and Etzel, 1982). Informational influence may happen in two different ways. Individuals may actively be searching for information from opinion leaders or from a group with the appropriate expertise. Otherwise, the individuals draw a conclusion by observing the behaviour of significant others (Park and Lessig, 1977).
The individual wants to stand out among others. These individuals want to show other individuals who they are and do so, for example, by talking about a particular film. This way they can distinguish themselves from others. It is also reflected in the acceptance of positions expressed by others. The association can either attempt to resemble the reference group as much as possible or flows from an attachment or liking for the group. The individual is responsive to the reference group out of a feeling for it, not because of a desire to be associated with it (Bearden and Etzel, 1982). Utilitarian influence reflects in attempts to comply with the wishes of others to achieve rewards or to avoid punishment. Individuals are afraid to make the wrong choices in fear that they will not be accepted within or that they will be kicked out of the group. If an individual feels that certain types of behaviour will result in rewards or punishment from others and if these outcomes are viewed as important, he or she will find it useful to meet the expectations of these significant others (Bearden and Etzel, 1982).
This research focuses on susceptibility to interpersonal influence. Consumer susceptibility to interpersonal influence is defined as follows: the need to identify or enhance one’s image with significant others through the acquisition and use of products and brands, the willingness to conform to the expectations of others regarding purchase decisions and/or the tendency to learn about products and services by observing others and/or seeking information from others (Bearden et al., 1989). McGuire (1968) concluded that consumer susceptibility to interpersonal influence is a general trait that varies across individuals and that an individual’s relative degree of being influenceable in one situation tends to have a significantly positive relationship to his or her degree of being influenceable in a range of other social situations. 2.3.3 GENDER AND AGE
In particular young women have been found to be quite skilled at spreading WOM (Ranjbarian et al. 2011). For example, Bush et al. (2004) found that females were more suitable than males to say positive things about, recommend products or services, encourage friends or relatives to buy certain products or to make use of certain services endorsed by their favourite athlete.
2.4.1 TYPE OF FILM
Children’s film can be distinguished with regard to their style of production into mainstream and independent film, just like adult films. There are different lines of argumentation about the definition of independent films: they are either conceptualized as experimental films or as ambitious films and they are often screened at festivals and less frequently in cinemas. In the special case of children’s films, independent film is defined in opposition to mainstream films as ambitious films, which tackle difficult topics and are not produced for entrainment purposes only (Völcker, 2005). This notion already implies that mainstream films are the dominant species in the film industry, aimed at the mass market, whereas independent films are in general focused on a niche market (Gemser et al., 2007). This research will use films produced into the mainstream category, but also independent films (mainly local films). 2.4.2 WITH WHOM
This section focuses on the control variable with whom (coordinated consumption effect) children can watch a particular children’s film. The degree to which consumers are influenced by the intended behaviours of others is referred to as the coordinated consumption effect. In this case, the more individuals intend to go and watch a given film, the more likely these consumers are to find somebody else to accompany them. The social pressures to join in are also much greater. The coordinated consumption effect is most likely to occur at the film’s release. Characteristics are that the particular film has been advertised heavily or immediately after launch when the advertising effects still remain strong (Broekhuizen et al., 2011).
2.5 CONCEPTUAL MODEL
3
RESEARCH METHODOLOGY
This chapter describes the research methodology. Firstly, the methodology for measuring WOM will be explained. Secondly, the data collection will be presented. This chapter will end with the measurement of the variables. 3.1 METHODOLOGY FOR MEASURING WOM Traditional attempts to measure WOM are based on two principal techniques: inference and survey (Dellarocas, 2004). The questionnaire method will be used to collect relevant information regarding the influence of children’s WOM on children’s film. The reason for choosing a questionnaire is because it has three specific objectives. Firstly, the questionnaire translates the information needed into specific questions. Secondly, it also urges the respondent to become involved and lastly, it also minimizes response error (Malhotra, 2004). Moreover, surveys remain the most popular method to study WOM, largely because individuals can be asked directly about their communication habits; the error then lies in the self‐reporting of behaviour (Dellarocas, 2004). Furthermore, another reason for choosing the questionnaire method is, because in reality it is very difficult ‐if not impossible‐ to exactly observe who, when and over which films the children talk about 24 hours a day and 7 days a week. If in reality this research had been conducted based solely on observation, it would have become tremendously difficult to detect and analyze all the different communication flows.
Regarding cognitive and social development, John (1999) classifies three consumer socialization stages: the perceptual stage (3‐7 years), the analytical stage (7‐11 years) and the reflective stage (11‐16 years). The stage of development of a child will in fact influence their performance on each of the tasks involved in answering a question. Children first have to understand the question and determine the intended meaning. In the second step they have to retrieve information from memory and use this to come to a tentative answer. They then have to formulate the answer, either by choosing the appropriate response category or by actively verbalizing their thoughts (Borgers et al., 2000).
For this research it was decided, based on the above description, to target children from seven to twelve years old to fill out the questionnaire. This is because these children find themselves in the analytical and reflective stage and children’s film is aimed mostly at children up to twelve years old. Furthermore, the cognitive growth of children for filling out the questionnaire was taken into account and this is why all the questions were phrased simply, making them easier to answer.
Consequently, for this research the respondents were confronted with 16 films (profiles) which allow to estimate the main effects for each factor level of a full fractional factorial design.
There are three methods to conduct a Conjoint analysis. In the first method, respondents are asked to assign a preference score to each profile. This type of method is typical when a Likert‐type scale is used or when the respondents are asked to assign a number from one to a hundred to indicate preference. In the second method, respondents are asked to assign a rank to each profile ranging from one to the total number of profiles. In the third method, respondents are asked to sort the profiles in terms of preference (SPSS, 2011). The first ‐preference score method‐ was used for this research, because it is much easier for the children to take part in by giving a score between 1 (unhappy face) and 10 (happy face) for each film. The other two methods are more complex for the children to rank or sort each of the films (profiles) according to their preference. This can become more difficult for the children, compared to only assigning a score to a film in terms of how gladly they would want to watch each of the 16 different films.
3.2 DATA COLLECTION
The questionnaire had to be adapted to the cognitive limitations for the chosen age group of this research. Standard research questionnaires for adults could not be used. The reason for this being is that their phase language is still developing and reading skills are just being acquired. Understanding language is very important for data quality. The biggest difference between questionnaires for adults and children is the use of negatively phrased questions (Borgers et al., 2000). As a result, for this research, negatively phrased questions are avoided, because the children will not easily understand those questions.
3.2.1 PRETEST
On request, the questionnaire is available in Dutch to the interest of the reader. The English version of the questionnaire is enclosed in Appendix III.
Before the research could be conducted at the selected primary schools, the questionnaire was tested under some children to see if they understood the questions and were able to answer them appropriately. On the Bastiaan square in the city of Delft the questionnaire was tested on some children aged seven to twelve years. After the questionnaire was tested, several adjustments were made to the questionnaire to make it more understandable for the chosen age group of this research.
3.2.2 SCHOOL SELECTION AND CLASS PRESENTATION
Data for this research was gathered from schoolchildren attending randomly selected primary schools throughout the Netherlands. This was done to ensure a more representative and diverse sample. This approach is similar to the one that was used in the research of Kratzer and Lettl (2009). The primary schools located in the Netherlands were contacted and information about the research was given. After obtaining permission from several of the primary schools to conduct this research, the data collection started. To make it more appealing to the children to completely fill out the questionnaire, several cinema tickets were raffled off under the participating classrooms. This research was conducted at three different primary schools in the Netherlands: O.B.S. De Pendinghe situated in Groningen, R.K. Cornelis Musiusschool in Delft and G.B.S. De Mikkelhorst in Haren. The total number of respondents for this research totalled 194 children in 13 classrooms. Children from grade 4 (7 and 8 years), grade 5 (8 and 9 years), grade 6 (9 and 10 years), grade 7 (10 and 11 years) and 8 (11 and 12 years) took part in this research. Grade 4 totalled only one classroom, while grade 5/6 totalled seven classrooms and five classrooms for grade 7/8. However, the net sample included 152 valid and acceptable samples; 78.35% of the total amount of respondents. The incomplete questionnaires were not used and deleted from the sample.
In the beginning of the presentation it was explained to the children that if they did not understand a question, they were free to raise their hands, so that the question could be clarified in order for him or her to be able to fully understand it and answer it correctly.
For the second part of the research ‐the Conjoint analysis‐ the children were asked to give a score from one (lowest) to ten (highest) to a set of 16 films (profiles) according to their preference. The 16 films each described different combinations of the five chosen attributes. One by one, all the 16 films were displayed with the aid of the
Microsoft Office program PowerPoint® in the classroom using a SMARTBoard®.
Appendix IV gives the examples for the first two films. The children needed to give a score between one (unhappy face) and ten (happy face) to each of the 16 films on the form they were given. A copy of this score method can be found in Appendix V.
The information obtained from the questionnaires and the Conjoint analysis was treated confidentially. The closed questions and the Conjoint analysis were analyzed through the statistical program SPSS (Statistical Package for the Social Sciences). The open questions were analyzed through the Microsoft Office program Excel®. Figure 5
FIGURE 5 Data sources to test the conceptual model 3.3 MEASUREMENT OF THE VARIABLES The chosen variables for this research were taken from previous researchers that have examined similar or semi‐similar variables in their respective studies. In some cases the questionnaire items were adapted to the purpose of this research, so that it was understandable and suitable for the children to respond.
3.3.1 QUESTIONNAIRE
The variable opinion leadership (questions 1, 2 and 3) was measured by using the scale measure from the research of Kratzer and Letll (2009). The preference for heavily advertised films variable indicating less versus heavily advertised films (questions 4, 5 and 6) was adapted from Basuroy et al. (2006) and Boatwright et al. (2007). The variables for susceptible to normative influence (questions 7, 8 and 9) and susceptible to informational influence (questions 10, 11 and 12) were measured by using the questionnaire items from the research of Bearden et al. (1989) and Mangleburg et al. (2004). The variables evaluation of last experience and amount of WOM produced (questions 13, 14 and 15) was measured by adapting the scale used by Broekhuizen et al. (2011). Gender (question 16) and age (question 17) were included as background variables. Adaptations of Likert‐type scales to better suit children have included the use of unhappy, neutral, or happy faces and a thermometer (to represent a continuous scale) (Reynolds and Kamphaus, 2003). Therefore, questions 1 through 12 were valued by continuously using the five point Likert scales measurement system and were divided between always (1) and never (5).
3.3.2 CONJOINT ANALYSIS
The Conjoint analysis has five attributes (factors), which are shown in table 1. These attributes are: valence, source, how many, type of film and with whom. The valence attribute has two options, either positive or negative. The second attribute is source and it has four options. This attribute specifies where the children obtain information from; family members, school friends, friends in their neighbourhood, or friends that live far away. Under the how many attribute there are also four options namely, that they might have received recommendations from one individual, two individuals, three individuals, or from four or more individuals. The attribute type of film has only two options namely, independent film or mainstream film. The last attribute is with whom they go and watch the film. This attribute also has four options. The children can go alone, with their parents, with one or two friends or with three or more friends. In Appendix VI, the 16 different films used for this research have been outlined.
4
DATA ANALYSIS AND RESULTS
4.2 OPINION LEADERSHIP
4.4 SUSCEPTIBILITY TO NORMATIVE INFLUENCE
Table 7 illustrates the mean and standard deviation of the items that were used to measure susceptibility to normative influence.
4.6 EVALUATION OF LAST FILM EXPERIENCE AND THE AMOUNT OF WOM PRODUCED
4.6.1 LAST VIEWED FILM
4.6.3 AMOUNT OF WOM PRODUCED
4.7 CORRELATION BETWEEN THE VARIABLES Correlation between the variables described above is presented in table 15. The Pearson correlation test showed that there are 10 significant correlations (linear relationships) between the variables (p<0.05). There is a significant correlation between opinion leadership on the one hand, susceptibility to normative influence, susceptibility to informational influence, the amount of WOM produced and age on the other. This suggests that opinion leaders are more susceptible to normative and informational influence. Opinion leaders continuously need to listen to a lot of information (high amount of WOM) coming, often, from friends and this could greatly influence a respondent. Furthermore, older respondents are often seen as opinion leaders.
A positive significant correlation is also present between the preference for heavily advertised films on the one hand, susceptibility to normative influence and susceptibility to informational influence on the other. This indicates that the respondents, that have a preference for heavily advertised (mainstream) films are also more susceptible to normative and informational influence. The reason for this might be that they want to satisfy their needs and are afraid to make wrong film choices or they do not want to hear bad stories about a particular film from other individuals.
There is a significant correlation between susceptibility to normative influence on the one hand, susceptibility to informational influence and the amount of WOM produced on the other. Thus, the respondents that are more susceptible to normative influence are also more susceptible to informational influence. It can be suggested that the more the respondents are susceptible to normative influence, the more WOM they produce.
4.8 CONJOINT ANALYSIS
To analyze the second part of this research, the Conjoint analysis experiment, a similar approach was used as the example given in the PDF‐file of SPSS (2011). Information from relevant articles and books were also used to explain certain terms related to the Conjoint analysis more thoroughly.
4.8.1 UTILITY SCORES
A utility score, or part‐worth, is derived from the responses of each individual, which measures the unique subjective judgment of preference (Chan, 2005). The 16 films vary in their performance on the attributes and by analyzing the preference for each of the 16 films, it is possible to assess the importance of the attributes. The importance scores can be used to determine the utility scores. A higher utility score means that the factor levels have larger positive values, whereas a lower utility score means that the factor levels have larger negative values (SPSS, 2011).
Furthermore, these utilities can be added together to give a total utility per film. There are 16 films, as it is outlined in appendix V. For example, the total utility of film 1 with valence positive, source friends that live far away, how many four or more individuals, type of film mainstream film and with whom with one or two friends would give the following:
Table 17 displays the estimated total utility and the actual utility (mean) for all the 16 films (Appendix V). Film 13 has the highest total utility and the highest mean. This means that film 13 has the greatest preference among respondents. On the other hand, film 3 has the lowest total utility and the lowest mean and thus indicates that respondents ‐if they were given the opportunity to watch this film‐ will not likely choose to watch it.
Table 18 shows the correlation between the estimated and the actual
4.8.2 RELATIVE FACTOR IMPORTANCE SCORES FOR THE WHOLE GROUP Relative importance is how much difference each of the factors (attributes) could make in the total utility of the product or service. Relative factor importance scores thus represent the categories’ influence on the respondents’ overall judgments. The relative factor importance score is equal to the difference between the highest and lowest standardized utility scores (parts‐worth) for a given factor, divided by the sum of the differences for all factors and multiplied by 100% (Burke and Cooper, 2009). The relative factor importance scores represent percentages and the property that they sum up to 100 (SPSS, 2011).
4.8.3 RELATIVE FACTOR IMPORTANCE SCORES FOR THE SENDER/RECEIVER CHARACTERISTICS
The following sections will give the relative factor importance scores for the following variables: opinion leadership, preference for heavily advertised films, susceptibility to normative influence, susceptibility to informational influence, gender and age. Using the median split, the variables were divided into two groups namely opinion leadership (low versus high), preference for heavily advertised films (low versus high), susceptibility to normative influence (low versus high), susceptibility to informational influence (low versus high), gender (boys versus girls) and age (7‐9 years versus 10‐12 years). The differences in importance ratings are checked between the groups.
4.8.3.1 OPINION LEADERSHIP
In table 20 the importance scores for opinion leadership are shown. The factor with whom has the highest overall preference score for the groups with low and high opinion leadership, whereas the factor how many has the lowest score in determining overall preference in both groups. For the group with low opinion leadership, WOM, which has a combined effect of 51.31%, is more important than the combined effect of the control variables regarding the decision process for watching a particular children’s film. The group of high opinion leadership found the combined effect of the control variables (50.78%) to be more important than WOM in the decision process to watch a particular children’s film.
4.8.3.2 PREFERENCE FOR HEAVILY ADVERTISED FILMS The result of the importance scores regarding preference for heavily advertised films is displayed in table 21. The factor how many has the lowest overall preference score for the groups with low and high preference for heavily advertised films. Although those with a preference for heavily advertised films do not seem to find the factor type of film more important, it can be said that they rely less on WOM. This may hint that strong advertising acts as a replacement for WOM. The factor with whom has the highest score in determining overall preference for respondents in both groups. The factor source plays a significant role in overall preference within both groups, but not as significant as the factor with whom. The group with low preference for heavily advertised films, which has a combined effect of 51.33%, finds WOM to be more important compared to the group with a high preference for heavily advertised films in the decision process to watch a particular children’s film.
4.8.3.4 SUSCEPTIBILITY TO INFORMATIONAL INFLUENCE
4.9 CONCEPTUAL MODEL RESULTS In this paragraph the results of the different relationships of the conceptual model will be briefly summarized. 4.9.1 QUESTIONNAIRE DATA 4.9.1.1 Output WOM Using the questionnaire data there is a significant mean difference in degree of amount of WOM produced across age. Respondents aged 8 (with an average of 12.94) talked much more with their friends about their last film experience. As a consequence, they produce the most WOM compared to respondents of age 11 (with an average of 3.83). This age group talks the least about their last film experience with their friends. There is an indication that the more positive the respondents are with their last viewed film, the more WOM they produce. Precaution should be taken when generalizing this indication, because it was not significant. Moreover, opinion leadership (0.37), susceptibility to normative influence (0.38) and susceptibility to informational influence (0.23) are significantly correlated with the amount of WOM produced.
4.9.2 CONJOINT ANALYSIS
5
CONCLUSION
In this chapter the results discussed in the previous chapters will take on a more abstract level. The main research question and the sub questions will be answered first, followed by the theoretical and managerial implications. This paper will end with the limitations of this research and suggestions for possible future research.
5.1 RESULTS
This research is the first to specifically address children’s WOM, since there is not much known about the degree to which children spread WOM, the valence (positive or negative messages) or the impact on film choice. Therefore, this research has tried to shed some light on the influences of children leading up to a particular choice for a specific children’s film.
The following two sub questions were formulated in order to be able to answer the main research question.
Especially those that have a preference for heavily advertised films, attribute great importance to with whom they go to see a film (see table 21).
2. What factors influence the production of output WOM?
The sender/receiver characteristics (opinion leadership, susceptibility to normative influence, susceptibility to informational influence, preference for heavily advertised films, gender and age) play a small role in the production of output WOM. There are little differences in what type of children produce the most WOM among their friends. The differences in traits cannot explain the amount of WOM produced. The only significant mean difference is in the amount of WOM produced across age. The children aged 8 are the ones that produce the most WOM, thus sharing their last viewing experience with most of their friends. There is no significant mean difference in the amount of WOM produced across gender. Even though there is also no significant mean difference in the amount of WOM produced across evaluation of last film experience, there is a indication that if the children find their last film experience to be very good, the more WOM they will produce. Most output WOM is positive. This is also referred to as positive WOM bias. This is exactly the reason why the more positive they are about the films, the more WOM they produce. The correlation table (table 15) shows that the difference in opinion leadership, susceptibility to normative and susceptibility to informational influence also explain the production of WOM.
This research is aimed at answering the following main research question:
“To what degree does WOM influence the choice of children for watching a children’s film and how do these children engage in WOM after watching the film?”
The children aged 8 are the ones that share most of their last film experience with their friends. This indicates that they produce most WOM after watching a particular children’s film. Furthermore, the children that have a preference for heavily advertised films are also more susceptible to normative and informational influence and produce more WOM. Socially susceptible individuals are the ones that have a greater preference for watching heavily advertised films, because they want to be accepted and do not want to make a wrong choice. Heavily advertised (mainstream) films are generally accepted on a more broader scale. 5.2 THEORETICAL IMPLICATIONS
The main theoretical contribution of this research to WOM literature is that WOM received and produced by children has an influence in the decision process of the children to watch a particular children’s film. This research shows how WOM works for children ‐the degree of WOM influence on the actual behaviour of a child‐ to be precise. This is one of the first researches to investigate children’s WOM in this product category and shows that children’s WOM is positively relevant for the children’s film genre. From a promotional activity point‐of‐view, WOM is considered to be one of the most influential methods that moviegoers rely on in the film industry. This is in line with previous research (e.g., Haupert, 2006; McLain, 2008; Zigu, 2011) and what is against previous research (e.g., Bush et al. 2004; Ranjbarian et al., 2011) is that the sender/receiver characteristics ‐opinion leadership, susceptibility to normative influence, susceptibility to informational influence, preference for heavily advertised films, gender and age‐ play very little to no role in generating output WOM. This indicates that there is little difference in what kind of children produce the most WOM among their friends. An explanation could be that the children find it difficult to make sense of those characteristics. For example it does not really matter if the children are or not susceptible to normative or informational influence, although there is little difference in relative factor importance scores. It might be that the chosen age group is more susceptible to normative and informational influence.
Remarkably, the factor how many has a negative influence on the children’s decision to watch a particular children’s film. The more individuals recommend a particular children’s film to a child, the less he or she wants to watch that film. This can be explained by the fact that a child may have the tendency to disagree with the WOM sender. The children might not want to watch a particular children’s film if more individuals recommend it to them for watching, because they might think that they are the last ones to watch that film. They might want to watch a particular children’s film only if nobody or fewer individuals recommend it to them. The children may thus display signs of reactance, to the extent that too many recommendations irritate them (Hoffmann and Broekhuizen, 2009). Another explanation could be that recommendation received from a close friend whose past recommendations have been right on target may be more than enough to get to the child to watch a new film even if it is in a genre he or she generally dislikes (Linoff and Berry, 2011).
This research shows that the factor type of film ‐mainstream films‐ (see table 16)
is not very important. This research gives an explanation, driven by the fact that WOM received from friends may be seen as more reliable as a source for film information, rather than mainstream film advertisement on television, radio, magazines or the internet (heavy advertised films). Therefore, WOM messages are perceived as being more personally relevant than commercial advertising messages (Nickels, 1984).
Based on some previous empirical studies (e.g., Van Eck et al., 2011), opinion leaders tend to be less susceptible to normative and informational influence. In contrast, this research shows that opinion leaders tend to be more susceptible to normative and informational influence based on the correlation, as seen in table 15. This can be clarified by the fact that opinion leaders by definition communicate more frequently with others regarding their product or service domain of expertise. Furthermore, to determine their position they need to assess social information frequently. They are as such constrained by the norms and information of the social system in which they operate. Opinion leaders probably do not rely entirely on norms and information of others to make decisions about the products or services in their area of expertise (Clark and Goldsmith, 2005).
