Marijn de Boer June 12

(1)

Marijn de Boer

(2)

Modeling the Added Value of the 3D Effect in Feature Films 2

“Modeling the Added Value of the 3D Effect in Feature Films”

Author

Marijn de Boer

Schuitendiep 1001

9712KD Groningen

T 0617250301

E m.de.boer.31@student.rug.nl

S 1933655

University of Groningen

Faculty of Economics and Business

Master Thesis Business Administration

Specialization: Marketing Management & Marketing Research

Supervisor: Dr. J.E. Wieringa

Second Supervisor: Alec Minnema

Company: Wolff Groningen

(3)

Management Summary

The purpose of this research is to explain which variables influence appreciation for the 3D effect in 3D movies. This is also the main research question of this paper. Due to the availability of new techniques, the phenomenon of 3D cinema has made a major comeback in the cinema industry as more and more 3D movies are released every year. However, both box-office numbers as well as professional opinions are divided with regard to 3D movies. Therefore, Dutch moviegoers are questioned on their behavior on, and perception towards, 3D cinema. This data, combined with movie data, is the input of a predictive model explaining the relationship between various consumer and movie characteristics on the appreciation for the 3D effect in a 3D movie.

Based on literature, eight different main hypotheses are formed which reflect the expected effect of the different consumer and movie characteristics. The experience a moviegoer has with 3D cinema together with three perceptional variables are the four consumer characteristics which are expected to influence the appreciation for the 3D effect in a 3D movie. The three perceptional variables are the moviegoers’ perception on (1) the general relative advantage of 3D cinema, (2) the quality of the movie and (3) the price surcharge. Also, four movie characteristics are expected to influence appreciation for the 3D effect; (1) the genre, (2) the budget, (3) the duration and (4) the IMDb-rating of the movie.

In order to collect the necessary data to test the hypotheses, a questionnaire was developed and spread through the internet. Here, consumers were asked to rate the 3D effect of the 3D movie they have seen and also had to answer question regarding their perceptions and behavior. Movie data was collected through various sources. With this data, two models were developed. A model which aggregates all the data (OLS) and a model which makes a distinction between different classes based on various demographic data (LCR). The latter identified six different classes which all show distinctive characteristics; (1) Male Mass, (2) Young Male Pathé Avoiders, (3) Frequent Pathé Visitors, (4) Old Female 3D Fans, (5) 3D Dislikers and (6) Young Female Pathé Avoiders.

(4)

Modeling the Added Value of the 3D Effect in Feature Films 4 The moviegoers’ perception on the general relative advantage of the 3D effect in 3D movies and on the comfortableness of the 3D-glasses positively affects their 3D ratings. Also, when they perceive the price surcharge which has to be paid as justified, this also positively affects their 3D ratings. Finally, for three classes, the 3D rating is lower when the moviegoer has seen more 3D movies. However, the 3D rating of the other three classes goes up when they have seen more 3D movies.

This last finding indicates that saturation with 3D cinema is not yet an issue for everyone. However, the overall average 3D rating given by the sample is 53.89 (on a scale from 0-100). Therefore, in order to improve appreciation for the 3D effects in 3D movies at Wolff Cinema Group, several recommendations are given. First, when a 3D movie is animated and/or a Family/Adventure, this should be programmed in 3D more often relative to other 3D movies. The contrary holds for Drama/Music movies. Secondly, it is wise that Wolff Cinema Group elaborates more on the reason for the price surcharge in order to gain more comprehension regarding this surcharge and hence, increasing 3D appreciation. Thirdly, it is suggested to experiment with different glasses, as its comfortableness might affect 3D appreciation.

(5)

Preface

This master thesis is the concluding piece of my Master degree in Business Administration, with a specialization in both Marketing Management and Marketing Research. The foundation of this thesis was based on my personal love for movies and interest in consumer research. Therefore, working on this thesis was both educational and fun. I would like to thank Ardi Sanwikrama and Wolff Groningen for giving me the opportunity to write my thesis at their beautiful cinema. The tour through the cinema, the interesting interview with Douwe Vos, insights in their moviedata and their consumerdata have proven very helpful for the formation of my thesis. I would like to thank my first supervisor Jaap Wieringa, for his feedback and tips, especially regarding the research methods I have used. Furthermore, I would like to thank Alec Minnema for his feedback on my first concept, which helped me to improve my thesis into a finished product. I hope that the findings prove to be useful for Wolff Cinema Group and finally, that you as a reader will enjoy reading this thesis.

(6)

5.1 Conclusions ... 42 5.1.1 Hypothesis 1 ... 42 5.1.2 Hypothesis 2 ... 43 5.1.3 Hypothesis 3 ... 43 5.1.4 Hypothesis 4 ... 44 5.1.5 Hypothesis 5 ... 45 5.1.6 Hypothesis 6-7 ... 46 5.1.7 Hypothesis 8 ... 47 5.1.8 Reflection ... 47 5.2 Practical recommendations ... 49 5.3 Research recommendations ... 50 5.3.1 Limitations ... 51

5.3.2 Directions for further research ... 51

Literature ... 53

Appendix ... 59

Appendix A: Survey (in Dutch) ... 59

Appendix B: Complete movie list ... 61

Appendix C: Distribution of cinematic behavior ... 62

Appendix D: Distribution of moviegoers´ perception ... 62

Appendix E: Log-Likelihood Values ... 63

Appendix F: Correlation matrix ... 63

(8)

1 Introduction

1.1 Background of the problem

3D movies are increasingly being released in the cinemas. In the year 2009, the year of the release of the 3D blockbuster ‘Avatar’, 13 movies were released in cinemas in 3D, in 2010 this number was 36 and in 2011 the count is already at 48 movies which are released in 3D in cinemas (Sony Professional Education and Knowledge, 2011). Not only new movies are produced in 3D, also classic movies like Lion King, Titanic and the Star Wars saga are being re-released in 3D. Despite the increase in the number of movies which can be seen in 3D, many 3D movies have disappointing box office numbers. For example, the box office of the 3D version of Pirates of the Caribbean: On Strangers Tides was only 38% of the total opening box office in the US, leading to missed income for the distributer (Greenfield, 2011). According to the Economist (2011), not only Pirates of the Caribbean suffered from a low 3D attendance. Also Hollywood blockbusters like Kung Fu Panda 2, Green Lantern and Harry Potter and the Deadly Hallows part 2 generated more money from 2D screens. Franich (2011) from Entertainment Weekly also demonstrated the preference for 2D. He posted a poll in his article on the latest Harry Potter movie asking readers which version of the movie they are planning to see; over 75% answered ‘ in 2D’, with not even 10% answering ‘in 3D’.

In most theaters, 3D movies are shown in 3D, while the 2D version of the movie is only shown in selected theaters. However, the disadvantages of a 3D movie are increasingly mentioned among moviegoers. According to Richard Greenfield (2011), researcher at BTIG Research, the three most commonly mentioned complaints of 3D movies are (1) uncomfortable glasses, (2) expensive price surcharges and (3) an oversaturated market. A question which could be answered is whether the positioning of 3D movies should not be reversed; movies available in 3D are shown in 2D in every theater, and in 3D only in selected theaters. According to Greenfield, a movie like Pirates of the Caribbean: On Strangers Tides could have benefit from this situation, as more moviegoers were then more able to see the 2D version.

(9)

Modeling the Added Value of the 3D Effect in Feature Films 9 people visited the cinema. Of these 28 million moviegoers, more than 4.2 million went to see a 3D movie (NVF, 2010). A film production company specialized in 3D filming, also states that the cost increase of filming in 3D over 2D is not that high. They calculated an average cost increase of a 3D film over a 2D version of 19%. Because filming in 3D over 2D is a fixed cost increase, this percentage decreases as the budget of the movie grows (Safonova, 2011).

Besides contradicting box office numbers for 3D and 2D movies, the opinions of key figures in the movie industry are also divided. For instance, Christopher Nolan, director of amongst others Inception and The Dark Knight declared that he is not a huge fan of 3D. What he dislikes the most about 3D movies is the dimness of the projected image (Boucher, 2010). This opinion is shared with the famous American film critic Roger Ebert and film editor Walter Murch, who beliefs that 3D does not work with our brain and it never will (Ebert, 2011). In contrast, Martin Scorsese, director of amongst others Gangs of New York and The Departed, has a more positive standpoint of 3D cinema. He states that he has always liked 3D and that 3D cinema is quit logical as in real life we see in, and are, 3D (Kermode, 2010). Peter Jackson, director of the Lord of The Rings movies, has one explanation for negative opinions about 3D movies; when an audience has seen a bad 3D movie, they come to realize that they have paid extra for a movie that was as bad as a bad 2D movie, and therefore are a bit reluctant to see another 3D movie (Eisenberg, 2011).

Besides the issue whether one likes or dislikes 3D cinema, there is also a medical issue concerning 3D movies. Bos (2011) argues that it is irresponsible to participate in public traffic after watching a 3D movie, as it affects ones vestibular system. Martin Banks, a professor of optometry at the University of California at Berkeley states that due to a phenomenon called “vergence-accomodation conflict”, 3D movies can cause headaches and other discomforts. Also, Michael Rosenberg, an associate professor at the Feinberg School of Medicine, believes that watching 3D can amplify minor eye problems (Grifantini, 2010).

(10)

1.2 Wolff Cinema Group

This thesis is written for Wolff Cinema Group and more particular for Wolff Groningen. Next to Pathé and JT Bioscopen, Wolff Cinema Group has the third largest market share of the cinema exploiters in the Netherlands (5.9%) (NVF, 2010). Wolff Groningen is one of ten cinemas which Wolff Cinema Group exploits across the Netherlands. In total, they exploit 50 screens. Wolff Groningen has 10 screens which makes it the fifth largest multiplex in the Netherlands (NVF, 2010). The cinema is located at the site of Euroborg, the stadium of the soccer team FC Groningen. It is easily accessible by car, therefore attracting both young and old audiences from the city as well as moviegoers outside of Groningen.

Like all cinemas, Wolff Cinema Group faced the decision whether to digitalize their screens or not. In 2010, the EYE Film Instituut Nederland, NVB and NVF together developed the 38 million euro project Cinema Digitaal. The purpose of this project is to act as a collective for individual Dutch cinema exploiters, thereby helping them with financing the conversion to digital screens (Cinema Digitaal, 2011). By July 2011, 50% of all screens in the Netherlands were digital (Cineserver, 2011).

Because of the rise in popularity in 3D cinema, Wolff Cinema Group also decided to step in and fully digitalize their cinemas and by August 2011, all screens at Wolff Groningen were digitalized. Douwe Vos, manager and programmer at Wolff Groningen, states that the rise of popularity of 3D cinema was a big reason in deciding to convert to digital screens. According to Vos, the conversion to digital screens was inevitable on the long term, but they decided to digitalize on such a short term because of 3D cinema. Of the ten digital screens, five of them are now equipped with the necessary equipment in order to be able to show 3D movies. Vos states that this 50/50 ratio is quite common among cinemas. However, they are planning to make one smaller screen also 3D compatible. This way, 3D movies which have decreased in popularity still can be showed without having to use up a larger screen where other movies might attract more audience on that specific timeslot (D. Vos, personal communication, November 18, 2011).

1.3 Research objective

(11)

Modeling the Added Value of the 3D Effect in Feature Films 11 were mentioned before were particular to the US. In China and Russia, the revenue from 3D showings of Pirates of the Caribbean: On Strangers Tides were respectively 85 and 71 percent (Young, 2011). In the Netherlands, 3D showings of this movie were 87 percent of the total revenue. For Kung Fu Panda 2, a movie which in the US generated more money with the 2D version, this percentage is even higher; 89 percent (MACCS International, 2011). However, there are some explanations for these high percentages for 3D showings. Firstly, 3D movies generate more revenue due to the €2.50 price surcharge. Secondly, 3D versions get programmed more often than the 2D version. This makes it more difficult for moviegoers who want to see a 3D movie in 2D to find an appropriate timeslot. So, one can state that the 3D box office numbers in the Netherlands are slightly biased. Maybe, if 2D versions were programmed more often, the difference in 2D and 3D box office would be less big.

When combining this programming issue with the increase in released 3D movies and contradicting opinions it comes to show that this topic requires consumer research. For theater managers it is relatively easy to adapt the exhibition capacity which is allocated to a new movie, either by deleting the movie from their program or by shifting it to a different size screening room (Sawhney and Eliashberg, 1996). Therefore, the outcomes of this research can lead to recommendations on the programming of 3D movies. In order to find out the opinion of the Dutch moviegoers about 3D cinema, they are questioned about what they like or dislike about different attributes of 3D cinema. Ultimately, the appreciation for the 3D effect in a 3D movie is the key to the moviegoers’ opinion on 3D cinema. The main research question in this thesis therefore is:

‘What influences the level of appreciation for the 3D effect in a 3D movie?’

1.4 Theoretical relevance

(12)

Modeling the Added Value of the 3D Effect in Feature Films 12 these movies. And because consumers differ per country, a new insight in Dutch moviegoers is also a contribution to existing literature. Why more research on the movie consumers is important is also indicated by Wierenga (2006), who stresses that after all, they are the ultimate destination of the value chain of a motion picture. Finally, Eliashberg et al. (2006) also state that an understanding of audience behavior is very important in order to make the challenges faced by producers, distributors and exhibitors clearer. Therefore, this study provides both a relevant contribution to existing motion picture literature as well as a practical relevance for theater managers.

1.5 Thesis structure

(13)

2 Theoretical framework

The main research question revolves around the factors which determine the level of appreciation a consumer has for the 3D effect in a 3D movie. In other words, what factors form the basis of the consumers’ attitude towards 3D movies? Hoyer and MacInnis define an attitude as a relatively global and enduring evaluation of an object, issue, person or action (2007). It is this evaluation which is important for cinema exploiters to understand. This is because attitudes influence behavior, i.e. they can predict peoples’ overt actions, especially when attitudes are strong (Ajzen, 2001).

3D cinema can be considered an innovation. The attitudes of a moviegoer towards this innovation can be formed by different determinants. Firstly the features of the attitude object (the movie) naturally affect the attitude the moviegoer has. Secondly, the characteristics of the moviegoer can also affect one’s attitude towards an innovation. Personal characteristics are a very commonly used as a predictor of the diffusion of an innovation (Ostlund, 1974). In this study, ones’ personal characteristics are referred to as consumer characteristics (§2.1) and the features of the attitude object are referred to as movie characteristics (§2.2). Both determinants are theoretically founded in this chapter.

2.1 Consumer characteristics

The first consumer characteristic which is taken into account is the experience a consumer has with 3D cinema. Knowledge & experience is one of several factors which affect a consumers’ attitude (Hoyer and MacInnis, 2007). This indicates that the experience a moviegoer has with 3D cinema can affect its attitude towards it. It might be that one’s desire to see a 3D movie has been saturated as a result of the number of previously watched 3D movies. Greenfield (2011) also states that the 3D attendance is suffering because the novelty effect wears off. This saturation effect can also be found in innovation diffusion models (e.g. Bass, 1969; Rogers, 1995). Therefore, the possibility of saturation playing a role in the formation of attitudes towards 3D cinema is covered in the paragraph on 3D cinema experience.

As much as Ostlund (1974) acknowledges the fact that personal characteristics are important variables in innovation diffusion studies, he argues that perceptual variables can be even more effective variables. Therefore, the next three consumer characteristics which are conceptualized are perceptual variables.

(14)

Modeling the Added Value of the 3D Effect in Feature Films 14 that the addition of novel attributes (like the 3D effect to a 2D movie) is likely to improve product evaluation, since consumers interpret these novel attributes as additional benefits (Mukherjee and Hoyer, 2001). It is possible that the moviegoers’ general perception of the additional benefit of an extra dimension also affects their appreciation for the 3D effects in different 3D movies. Therefore, the moviegoers’ general perception on the general relative advantage of 3D cinema is also considered in this study.

The third consumer characteristic is the moviegoers’ perception on the quality of the movie. As the 3D effect is an addition to the base product (the movie itself), it seems reasonable that the evaluation of this 3D effect is related to the evaluation of the movie itself. Therefore, it might be that the moviegoers’ perception of the quality of the 3D effect in a movie is affected by their perception of the quality of the movie the 3D effect is a part of.

The last consumer characteristic is the moviegoers’ perception on the price premium which has to be paid for a 3D movie. The price premium for 3D movies in the Netherlands is fixed for all theaters. The premium is € 2.50 including and € 1.50 excluding 3D glasses. According to Greenfield, the expensive price premium is the biggest concern of 3D movies in the United States (2011). There, the average price premium for 3D movies is 46%. This price premium is a disadvantage of 3D cinema compared to 2D cinema for moviegoers. It might be possible that the consumers’ appreciation for the 3D effect is affected by its attitude towards this disadvantage. Therefore, the perception towards the 3D surcharge is the third consumer characteristic which is theoretically founded and hypothesized.

These four consumer characteristics are discussed in more detail in the paragraphs below.

2.1.1 Experience with 3D cinema

The 3D movie is a good thing for variety seekers. Because of the recent popularity of 3D cinema, the variety in the supply of movies has increased. For instance, the latest Harry Potter movie was available in 2D, 3D, IMAX, IMAX 3D and in most cases, also both available in Dutch and English (MACCS, 2011). Shen and Wyer (2010) show that when people feel bored with a decision, they are likely to choose an alternative. Also, McAlister (1982) finds that a consumers’ choice is dependent on their past choices. These findings are indications that the moviegoers’ past behavior are likely to affect their future choices concerning 2D or 3D versions.

(15)

Modeling the Added Value of the 3D Effect in Feature Films 15 from 1922 (SPEAK, 2011), but it can still be perceived as such. In line, Rogers (1995) states that ones’ perceived newness to a certain idea determines his or her reaction to it. And due to the large increase in 3D movie releases, it might have struck moviegoers as something new. When applying the three basic forms of innovation defined by Jacobs (2007) on a 3D movie, one can state that a 3D movie is a product innovation. One can also argue that it is a process innovation as new technologies are used during the filming process. According to the typology of Garcia and Calantone (2002), the 3D movie can be described as an incremental innovation, as it is only a new feature/improvement to the existing 2D technology in the existing movie market. In other words, a novel attribute to a movie is added.

Box office numbers in the United States suggest that moviegoers are increasingly resisting this innovation. Ram and Sheth (1989) find that generally, the resistance to an innovation is a result of the innovation failing to meet consumer needs, wants and preferences. Rogers’ model for diffusion of innovations demonstrates how an innovation is being communicated throughout time among members of a social system (1995). Over time, an innovation can lead to adoption or rejection. According to Wood and Swait (2002), almost every decision of a consumer concerning an innovation, whether it involves something really new or simply trying something else, is affected by the consumers’ need for cognition and their need for change. On the same level, Venkatraman and Price (1990) make a distinction between cognitive and sensory innovators. They found that cognitive innovators are more inclined to evaluate every single attribute of a product, where sensory innovators are more inclined to buy something they perceive as new without evaluating specific innovation attributes. These findings show that the adoption of innovations is very dependent on the personal characteristics of the consumers. Visitors of 3D movies (or version of a movie) can be either sensory innovators, who still perceive 3D cinema as something new, without caring about the attributes of a 3D movie, or cognitive innovators, who also perceive 3D cinema as something new and are very consciously choosing for a 3D film, taking the attributes of the movie in account.

Rogers (1995) showed that it is also possible to reject an innovation which has been previously been adopted. This so called disenchantment discontinuance can also be recognized in the decreasing box-office numbers for 3D-movies; an innovation which has been adopted, and perhaps now is being rejected by the social system. The famous Bass model (1969) demonstrates a similar effect; after a certain period of exponential growth, a period of decay follows. In other words, over time, the market is slowly saturating. Perhaps the market for 3D movies is also in the process of saturation. When combining this presumption with the finding that a consumers’ choice is dependent on their past choices (McAlister, 1982), the first hypotheses can be formulated:

Hypothesis 1: The number of previous seen 3D movies will negatively affect appreciation for the 3D

(16)

2.1.2 General relative advantage perception

Rogers and Shoemaker (1971) define the relative advantage of an innovation as follows: “the degree to which an innovation is perceived as being better than the idea it supersedes”. In the case of 3D cinema, it relates to the degree to which moviegoers perceive a 3D movie/version as better, or worse, than a 2D movie/version. Tornatzky and Klein (1982) performed a meta-analysis on the direct relationship of innovation attributes to the adoption of the innovation. They found that relative advantage was one of the only two significant characteristics which positively related to adoption. Ostlund (1973) also found that relative advantage is a particular important attribute in the diffusion of innovations. And according to Holak and Lehman (1990), relative advantage is often cited as the most significant variable when considering the rate of adoption of an innovation. Fornell et al. (1996) incorporated a similar variable to relative advantage, namely perceived value, in their model explaining The American Customer Satisfaction Index (ACSI). They showed that perceived value indeed is positively related to satisfaction.

More recent studies also incorporate relative advantage, which demonstrates the sustainability of this variable. For instance, Flight et al. (2011) used relative advantage as an innovation characteristic in their study where they investigate consumer perceptions of several innovation characteristics in different cultures for technology-based consumer durables. In another industry, internet banking, relative advantage is also used as an innovation attribute (Gounaris and Koritos, 2008). Another recent example is a study by Rijsdijk and Hultink (2007). They found that relative advantage is fully mediating the effect of product intelligence on consumer satisfaction. This indicates that consumers appreciate products because of the relative advantage they perceive in them and not for their intelligence itself (Rijsdijk and Hultink, 2007). The appreciation for the 3D effect in a 3D movie therefore is expected to be affected by the moviegoers’ general perception on the relative advantage of 3D cinema. This ultimately leads to the second hypothesis:

Hypothesis 2: A more positive perception on the general relative advantage of 3D cinema will have a

positive effect on appreciation for the 3D effect

2.1.3 Movie quality perception

(17)

Modeling the Added Value of the 3D Effect in Feature Films 17 In a study on the innovation evaluation process, Olshavsky and Spreng (1996) found that in some cases, the degree of satisfaction with an innovation was assessed relative to their perception on the performance of the present product. Several adoption diffusion models also use the comparison between the current and the new product as a decision rule (Mahajan et al, 1990). This implies that the evaluation of the innovation is influenced by the evaluation of the current (or base) product. In the case of 3D cinema this implies that it is possible that the moviegoers’ evaluation of the 3D effect in a movie is affected by their evaluation of that particular movie. Olshavsky and Spreng (1996) found that when consumers were very satisfied with the present product, they reject the innovative concept. However, in the case of 3D cinema, the old (the movie) and the new (the 3D effect) are experienced simultaneously. Therefore, it is expected that these two factors are positively related, which leads to the third hypothesis:

Hypothesis 3: A positive perception on the quality of the movie will have a positive effect on

appreciation for the 3D effect in that movie and vice versa

2.1.4 Price premium perception

In marketing literature it is commonly assumed that new features and product innovations in general tend to decrease price sensitivity toward the enhanced product (Nowlis and Simonson, 1996). When applying this principle to 3D cinema, one might expect that moviegoers are less price sensitive towards a 3D movie. However, Greenfield (2011) believes that pricing is the biggest concern of 3D movies. The price of a ticket for a 3D movie is a form of a portioned price, as the price of the actual movie and the surcharge for the 3D version are quoted separately, even though users have to pay for both components in order to see the 3D movie (Hamilton and Srivastava, 2008).

(18)

Modeling the Added Value of the 3D Effect in Feature Films 18 recall the total costs around 8% lower than the actual costs, when dealing with a portioned price (Lee and Han, 2002).

The perception of a price of a product or service can also affect the evaluation of the given product or service. Voss et al. (1998) find that when price and performance are inconsistent, pre-purchase price perceptions have a negative effect on satisfaction. Dodds et al. (1991) also find that consumers are likely to use price as a cue to form an expectation of the performance, especially in the case of performance uncertainty. The quality of the 3D effect in a 3D movie is not the same for every movie, which leads to a form of uncertainty about its performance. Because of the relative high price of a 3D movie (due to the surcharge), expectations of its performance might be higher than for a 2D movie. Hence, the price surcharge which might be overlooked or taken into account might also affect a moviegoers’ appreciation for the 3D effect in a 3D film after the film was seen. The construct of price perception towards the price surcharge is split up into two sub-constructs; the perception towards (1) the height and (2) the justification of the surcharge. Therefore, two separate hypotheses are tested:

Hypothesis 4a: Perceiving the 3D-price surcharge as inexpensive will have a positive effect on

appreciation for the 3D effect

Hypothesis 4b: Perceiving the 3D-price surcharge as justified will have a positive effect on

appreciation for the 3D effect

2.2 Movie characteristics

Many studies have already been performed on the effects of various movie attributes on word of mouth and/or box office returns (e.g. Basuroy, Chatterjee and Ravid 2003; Elberse and Eliashberg 2003; Ginshberg and Weyers 1999; Litman 1983; Moon, Bergey and Iacobucci, 2010; Prag and Cassavant 1994; Terry, Butler and De’Armond, 2005; Wallace, Seigerman and Holbrook; 1993 Yong, 2006). But little research has been done in the field of 3D cinema. Hence, the effects of 3D movie attributes on the appreciation for the 3D effect are unknown. In this study, four movie attributes are suggested to be of influence on a moviegoers’ appreciation for the 3D effect in a 3D movie.

(19)

Modeling the Added Value of the 3D Effect in Feature Films 19 with a good 3D effect. The last movie attribute which is taken into account is the duration of the 3D film. Wallace et al. (1993) also included this attribute in their study on the factors affecting box-office returns. It might be that due to the acclaimed medical issues concerning 3D cinema (Bos, 2011; Grifantini, 2010), the duration of a 3D movie also affects the attitude towards the 3D effect. These four movie characteristics are discussed in more detail in the paragraphs below.

2.2.1 Movie genre

The genre of a movie is a common movie attribute which is used in various academic literature which focuses on the motion picture industry (e.g. Elberse and Eliashberg, 2003; Moon et al, 2010; Yong, 2006). It is possible that the genre of a 3D movie also is an important determinant for the appreciation for the 3D effect in a movie. Also because the technique of creating a 3D animation (e.g. Kung Fu Panda 2), differs from the process of creating a live action film (e.g. Pirates of the Caribbean) (Preimesberger, 2011).

Moon et al. (2010) identify six major genre categories; thriller, romance, action, drama, comedy and animation. 3D movies (or versions) are present in all of these genres. Examples are Shark Night 2, Step Up 3D, Drive Angry, Dolphin Tale, Gulliver’s Travels and Toy Story 3 respectively (IMDb, 2011). The genre of a movie has been found to be an important factor for consumers in choosing which film to see (Vézina, 1997). Moynihan and Berger even found a relationship between the economic condition of a country and the popularity of certain genres (2010). Elberse and Eliashberg (2003) opt that the genre of a movie is, together with its MPAA rating, one of two key attributes which determines a movie‘s potential audience. Gazley et al. (2011) also find that consumers’ preference for movies differences across genres. Sochay (1994) find that, at least for America, the taste in genres of moviegoers is constantly changing. These findings imply that different genres attract different consumers, who each have different opinions about 3D cinema. Because the genre of a movie is important for the consumers’ movie choice behavior and it has an effect on the type of audience which is attracted to a certain movie, it can be suggested that the genre of a 3D movie is also of influence on the appreciation for the 3D effect in the movie.

(20)

Hypothesis 5a: Appreciation for the 3D effect in a 3D movie will differ across different genres

Hypothesis 5b: An animated 3D movie will result in more appreciation for its 3D effect compared to a

filmed 3D movie

2.2.2 Movie budget

Ravid (1999) shows that big movie budgets are correlated with higher revenue. Basuroy et al. (2003) show that a big movie budgets reduce the impact of negative reviews, which suggest that even though there are negative acclaims, consumers still have positive expectations of a movie due to its big budget. Litman (1983) argues that big budgets reflect higher quality. Therefore, it might be possible that the higher the budget of the 3D movie, the better the 3D effect in this movie and hence, the higher the appreciation for this 3D effect.

The budget might also reflect upon an issue which surrounds 3D cinema, namely the production method. There are two ways of creating a 3D movie. First, a 3D movie can be shot with an advanced 3D camera which instantly creates an image for each eye. Secondly, the monocular image shot with a normal camera can be transformed by rendering a second stereoscopic image sequence in post-production. This is also known as 2D/3D conversion (Kunter et al., 2009). So far, only a few movie titles converted from 2D to 3D, such as Nightmare before Christmas, G-Force, and parts of Harry Potter (Smolic et al., 2011). The film Clash of the Titans (2010) was also created using 2D/3D conversion. According to Hughes (2011), applying such conversion costs around $12 million, which can be worth the investment when the price surcharge on the movie tickets equate to approximately $3-5 million extra per week. However, a lot of movie critics stated that the 3D conversion in Clash of the Titans was awful (Green, 2010). Vos also states that after viewing the movie, Wolff Cinema Group decided to exhibit only the 2D version, due to the bad 3D quality (D. Vos, personal communication, November 18, 2011). When comparing the budget of Clash of the Titans ($125,000,000) and the 3D hit Avatar (2009) ($237,000,000) it shows that Avatar’s budget is almost twice as large (IMDb, 2011). A same comparison can be made using their ratings; Avatar earned 83 out of 100 points based on 35 critic reviews, Clash of the Titans earned only 39 out of 100 points based on 37 critic reviews (Metacritic, 2011). This difference in budget and critic score can also be related to the quality of the 3D effect and hence, can influence the moviegoers’ attitude towards this effect.

(21)

Modeling the Added Value of the 3D Effect in Feature Films 21 Holbrook, 1982). Therefore, it seems logical that the level of realism also has an effect on the moviegoers’ attitude towards the 3D movie. Even more so because the statement by Hirschman and Holbrook was based on the moviegoers’ experience with 2D movies from the early eighties. Therefore, it might be that this effect is even higher for contemporary 3D movies. However, the level of realism is a subjective determinant and therefore hard to measure. Therefore, in this study it is assumed that the budget also affects ones appreciation for the 3D effect through the level of realism. Combined with the findings which suggest that budget also influence the quality of the movie, the sixth hypothesis can be formulated:

Hypothesis 6: The height of the budget of the 3D movie will positively affect appreciation for the 3D

effect

2.2.3 Duration

A regression model from Wallace et al. (1993) found that an extra hour in the length of a movie contributes extra revenue. This suggest that the longer the movie, the better. However, their study focused on rental income. There are not many studies which research the relationship between the duration of a movie and its evaluation. It could be that the positive effect found by Wallace et al. (1993) also relates to the 3D movies; the longer the movie, the more appreciation for the 3D movie. However, as Greenfield (2010) points out, having to wear glasses in the cinema is one of the most heard concerns of 3D cinema, as wearing glasses can become uncomfortable and tiresome. This indicates that the longer the movie, the more tiresome it becomes to watch the 3D movie. Carrier (2011) also found other negative effects of watching 3D movies (2011). His findings suggest that, compared to 2D movie viewers, 3D movie viewers are almost 3 times more likely to experience a headache and over 4 times more likely to experience eyestrain (Carrier, 2011). Martin Banks, currently researching the effects of 3D on the visual system, thinks these discomforts are caused by a conflict between the vergence and the accommodation of the eyes (Grifantini, 2010). Smolic et al. (2011) state that generally, the 3D-viewing experience is comfortable, but only if salient objects stay close to the screen, otherwise the visuals will come outside the stereoscopic comfort zone.

(22)

Hypothesis 7: The duration of the 3D movie will negatively affect appreciation for the 3D effect 2.2.4 Perceived quality

The last movie characteristic which is considered is the overall perceived quality of the movie. In paragraph 2.1.3 it is already suggested that the moviegoers’ individual perception on the quality of a 3D movie reflects in their appreciation for the 3D effect in that movie. It might be that this effect also is present when using the score of the movie which is based on thousands of ratings. This score can be considered as a statement on the quality of the movie. However, it is common marketing practice that an overstatement (a high rating) results in more favorable ratings, and understatement (a low rating) results in less favorable ratings (Olshavsky and Miller, 1972). In other words, when the larger audience perceives a 3D movie as a good movie, this reflects in the appreciation of the individual moviegoer for the 3D effect. This forms the last hypotheses:

Hypothesis 8:An aggregated positive perception on the quality of the movie will have a positive effect on the appreciation for the 3D effect in that movie and vice versa

2.3 Conceptual Model

(23)

Modeling the Added Value of the 3D Effect in Feature Films 23 Figure 1: Conceptual Model

Movie characteristics Consumer characteristics Number of previously watched 3D movies Appreciation for a 3D movie Perceived general relative advantage of 3D cinema Perception on the height of the price

surcharge Perceived quality of

the movie

Genre of the movie

Budget of the movie

Duration of the movie

Aggregated perceived quality of the movie

(24)

3 Research design

3.1 Plan of analysis

In order to test the hypotheses, two predictive models are developed. Both models capture the effects from the different consumer and movie characteristics on appreciation for the 3D effect. Once optimized, the models are able to explain which specific characteristics have an effect on 3D appreciation. Also, the strength and the direction of the effects are defined. As a result, appreciation for the 3D effect in future 3D movies can be predicted, which is exactly the purpose of predictive models (Leeflang et al. 2002). The estimation of the models is done in the form of multiple regression analysis, as the relationship between a single dependent variable and several independent variables is analyzed (Hair et al. 2005).

The first model is developed using Ordinary Least Squares (OLS) regression. OLS is a commonly used method for estimating a regression line. This method determines the best-fitting line by minimizing the square of the distances of all the points from the line and thereby minimizing the errors (Malhotra, 2007). It is a powerful and flexible procedure for analyzing associative relationships between a single dependent variable and one or more independent variables (Malhotra, 2007). In the process of optimizing this model, the importance of the individual dependent variables is considered. This is done using a technique called Weighted Least Squares (WLS), which allows for treating observations unequally. Finally, this model is tested and, if necessary, treated for multicollinearity. This is because multicollinearity makes the parameter estimates unreliable (Leeflang et al. 2002).

Because there is a possibility that the parameters which are estimated in the first model differ across unobserved, or latent, subgroups, the second model is developed using Latent Class Regression (LCR). Aside from the variables used in the OLS estimation which influence the dependent variable (predictors), variables which influence the latent variable (covariates) are also used in the process of estimating this model (Vermunt and Magidson, 2005). These covariates are used to predict class membership. Once estimated, the profile of different classes can be described and the associated parameter estimates can be analyzed.

(25)

Modeling the Added Value of the 3D Effect in Feature Films 25 (1) the genre of the movie, (2) whether the movie is animated or not, (3) the budget of the movie, (4) the duration of the movie and (5) the overall perception on the quality of the movie.

3.2 Data and variables

In order to collect all necessary data to develop the model, two steps are taken. Firstly, a survey has been set up in order to gain the necessary consumer data. This data is supplemented with the necessary movie data which is collected from secondary sources.

3.2.1 Consumer data

The survey is designed using the software of Qualtrics. After activation, the software provides a hyperlink. It allows everyone who has received the link to fill out the questionnaire. The link is sent to the email-database of Wolff Groningen, it is posted on the Facebook page of Wolff Groningen and Wolff Bioscopen and it is posted on the intranet page of the Research Lab of the Faculty of Economics and Business. As an incentive, Wolff Groningen gives away a Wolff movie package to one of the respondents. This incentive is highlighted in the emails and posts which are directed towards potential respondents and mentioned in the introduction of the survey. The introduction of the survey also explains the nature and the content of the survey. After the introduction, basic demographic data is asked (sex, age and place of residence) in order to be able to form a good image of the survey sample and to use for the latent class regression. Next, the respondents are asked the questions necessary to collect the data for the independent variables.

The first variable is the number of times they have seen a 3D movie (H1). The second variable is perceived general relative advantage (H2). The studies which were used in the meta-analysis from Tornatsky and Klein (1982) used various methods of measuring the relative advantage of an innovation. However, the most adequate measures came from studies where the adopters had to rate the relative advantage themselves (e.g. Hayward, Allen & Masterson, 1976; Ettlie and Vellenga, 1979). Therefore, for the determination of the respondents’ perceived general relative advantage of 3D cinema, the respondents are asked directly to which extend they think that in general, the 3D effect is of relative advantage compared to 2D cinema. The next questions related to the independent variables are about the perception of the price premium. The respondent has to indicate their perception on the height (H4a) and the justification (H4b). The next part of the survey is set up in order to collect the data for H3 and the dependent variable.

(26)

Modeling the Added Value of the 3D Effect in Feature Films 26 the list they have seen in 3D. When they have found a 3D movie which they have seen, they are firstly asked to rate the movie on a scale of 0-100 (H3). Secondly, they are asked to rate the 3D effect in this same movie, also on a scale of 0-100. This data serves as the dependent variable in the model. It might be that the score that is given by the respondent is not fully representative for the actual appreciation the moviegoer had while watching the 3D movie because consumers in general pay more attention to negative attribute performance when evaluating a product after use (Mittal, Ross & Baldasare, 1998). Therefore, it is possible that the negative aspects of the 3D movie (e.g. uncomfortable glasses) are more dominant when forming an appreciation rating. However, as all consumers share this behavior, it does not influence the data in a way.

Aside from the questions which collect the necessary data, three additional questions are asked. The respondents are asked how many times and where they visit the cinema, which can help to answer whether the survey sample is incidental or hardcore moviegoers and whether they are representative for the Dutch moviegoers. Lastly, they are asked about their opinion on the fact that glasses have to be worn throughout a 3D movie. This is to check whether the acclaims of Greenfield (2011), that having to wear these is one of the biggest concerns of American moviegoers, also hold for the Dutch moviegoers. At the end of the survey, the respondents are asked to leave behind their email-address for a chance to win the Wolff movie package. An overview of the survey (in Dutch) is presented in appendix A (only one of the 48 movies in the list of the main part of the survey is shown).

3.2.2 Movie data

(27)

(28)

4 Results

4.1 General overview

During a period of three weeks, the questionnaire was active. After this period, 236 surveys were completed and registered. After deleting the invalid and incomplete surveys, 218 correct surveys remained. Most of the deleted surveys were only filled until the part were the respondents actually had to rate the movies and the accompanying 3D effect. Other respondents whose survey was invalid did not comprehend the goal of the last part, as they only rated the movie or the 3D effect. In the paragraphs below, general findings are presented of the corrected sample.

4.1.1 Sample profile

Figure 2 shows that most of the 218 respondents are between the age of 20 and 27. The youngest respondent is 16 and the oldest respondent is 53. According to the NVB (2010) the age category 16 to 23 years old contains the most and most frequent cinema visitors in the Netherlands. Based on this information, the sample used for this research is quite representative for the Dutch moviegoers. The division between males and females is approximately half and half in most categories. The only exception is the 28-31 category, as this category consists mostly of male respondents exists.

The next properties of the sample which can be outlined concern the cinema they visit. As it is possible to visit multiple cinemas, respondents were asked which cinema they mainly visit. Figure 3 shows that most respondents mainly visit Wolff Bioscopen and Pathé Bioscopen. A small portion of the sample mainly visits JT Bioscopen and only one respondent mainly visits Utopolis. Among the respondents whose answer was ‘other, namely...’ Cinema and Tivoli was mentioned most often. These two cinemas are part of Leeuwarden Bioscopen, which was also mentioned. Other answers include ForumImages and CineSneek.

(29)

Modeling the Added Value of the 3D Effect in Feature Films 29 In summary, further statements and conclusions made in this paper are based on data retrieved from, in general, younger males and females which mainly visit Wolff Bioscopen and Pathé Bioscopen.

4.1.2 Cinematic behavior

After the questions regarding general demographic data, respondents were asked a couple of questions regarding their behavior of- and perception towards cinema visits. The table below gives an overview of the viewing behavior of the respondents in the sample. They all have seen at least one 3D movie, which was also a condition for partaking in this research. On average, this sample visits the cinema 18 times per year. Since 2009, when 3D cinema became widely popular again, the respondents in this sample on average have seen five 3D movies. The highest number of 3D movies which was seen is 30. For a more elaborate figure on the distribution of the viewing behavior see appendix C which presents a histogram of the answers on both questions.

Table 2 gives an overview of the perceptual subjects which were questioned, including the scale on which the answer could be given, the minimum, maximum and the mean of the answers which were given. A more elaborate figure of the perceptions of the respondents can be seen in appendix D. Here, for all four perceptional variables a histogram showing the distribution of the data is presented.

When considering the data in this table and the distributions which are shown, a couple remarks can be made. Frist of all, the respondents are very mixed when it comes to the general added value of the 3D effect. There are respondents which are very positive and very negative. Most respondents gave a 6 and 7 on a scale of 1-9. However, the considerably large share of respondents which gave a 5 or lower brought the mean score to 5.02. Respondents are notable more negative when it comes to having to

Number of… Min Max Mean annual cinema visits 1 150 18.10 3D movies seen 1 30 5.61

Table 1: viewing behavior

(30)

Modeling the Added Value of the 3D Effect in Feature Films 30 wear glasses. In this issue, more than half of the respondents gave a rating of 4 or lower, leading to an average of 4.08. When looking at the issue of the price surcharge which has to be paid for a 3D movie, on average, the respondents are mixed with a slight tendency towards negativity when considering the justification of the surcharge. When considering the height of this surcharge, the negativity is more salient. Almost all respondents rated the height of the surcharge a five or higher, leading to an average of 6.32 which indicates that the respondents perceive the surcharge to be quite expensive.

Perception on: Scale Min Max Mean

general added value of 3D cinema 1(no added value) – 9(added value) 1 9 5.02 having to wear glasses 1(unpleasant) – 9(pleasant) 1 9 4.08 the justification of the surcharge 1(unfair) – 9(fair) 1 9 4.36 the height of the surcharge 1(cheap) – 9(expensive) 2 9 6.32

4.1.3 Movie charactaristics

The 48 movies which were presented to the respondents, on average, were rated 28.23 times. However, eight movies were rated less than nine times. These movies are Yogi Bear (6 times), Animals United (3 times), Justin Bieber: Never Say Never (5 times), Cats & Dogs 2 (1 time), Alpha & Omega (3 times), StreetDance 3D (5 times), Arthur Christmas (6 times) and Glee: The 3D concert movie (4 times). These low numbers of ratings are considered to be insufficient in order to be reliable. The table 3 gives a summary of the remaining 40 movies which are used for further analysis. It shows large differences between the smallest and biggest budget ($228,000,000), between the shortest and longest movie (74 minutes), IMDb rating (4.6 points), the number of ratings (128 ratings), movie rating (41.28 points) and most importantly, 3D effect ratings (40.91 points). These dynamics in the movie data are modeled in the next chapter.

Table 2: perceptions

Table 3: Movie characteristics summary

Variables Statistics summary

Genres Family/Adventure (37.5%) Action/Adventure (17.5%) Horror/Thriller (17.5%) Sci-fi/Fantasy (22.5%) Drama/Music (5%)

Animation 40 % of the movies

Min Max Mean

Budget (*$1,000,000) 9 (Nova Zembla) 237 (Rapunzel) 113.28

Duration (minutes) 88 (Piranha) 162 (Avatar) 106.55

IMDb rating (0-10) 3.9 (Shark Night) 8.5 (Toy Story) 6.6

Number of ratings 9 (Shark Night) 137 (Avatar) 33.05

(31)

4.2 Model 1: Ordinary Least Squares

As mentioned in the previous chapter, first a model is created which aggregates all the data and thus estimates outcomes which hold for the entire sample. This is done using the Ordinary Least Squares (OLS) method. The model building process will follow the three major parts of the classic approach; (1) specification, (2) estimation and (3) validation (Leeflang et al. 2002).

4.2.1 Specification

The first step in the process of creating a model is to specify the elements in that particular model (Leeflang et al. 2002). As mentioned in the plan of analysis, this model tries to explain the role the characteristics of consumers and movies have on the appreciation for the 3D effect in a 3D movie. Therefore, the model has the following form:

∑

Where

3DAPRc,m = Appreciation for 3D effect in movie m of consumer c;

n3Dc = Total number of 3D movies consumer c has seen;

PGRAc = Perception of consumer c on the General Relative Advantage of 3D cinema;

PQMc,m = Perception of consumer c on the Quality of the Movie they have rated;

PHSc = Perception of consumer c on the Height of the price Surcharge;

PJSc = Perception of consumer c on the Justification of the price Surcharge;

GNR1m = Dummy variable for genre g in movie m: 1 if Family/Adventure; 0 otherwise;

GNR2m = Dummy variable for genre g in movie m: 1 if Action/Adventure; 0 otherwise;

GNR3m = Dummy variable for genre g in movie m: 1 if Horror/Thriller; 0 otherwise;

GNR4m = Dummy variable for genre g in movie m: 1 if Sci-fi/Fantasy; 0 otherwise;

ANIm = Variable which indicates whether movie m is animated or not;

BGTm = Budget of movie m in US dollars (*1,000,000);

RNTm = Runtime of movie m in minutes;

IMDbm = The rating of movie m on the Internet Movie Database;

PWGc = Perception of consumer c on having to Wear Glasses;

(32)

4.2.2 Estimation

The next step in the creation of a model is to obtain estimates for the parameters which were presented in model above. As mentioned before, the model is estimated using OLS regression1. Before the model is estimated the variable concerning the consumers’ perception on the height of the price surcharge (PHSc)is transformed. As the questioning of this variable is opposite from the questioning regarding the justification of the surcharge (PJSc), PHSc is equalized with respect to PJSc. This means that a value of 1 becomes 9, 2 becomes 8, and so forth. Hence, interpreting the results regarding these variables becomes more convenient.

In order to obtain estimates for every genre, dummy variables have to be created. In this research, five different genres are tested. However, only four dummy variables are used in the estimation process. This is because, in the case of five variables, only four are independent. Information about this fifth omitted variable can be derived from information on the other four categories. This variable is known as the reference category (Malhotra, 2007). In this case, the reference category is Drama/Music.

In order to make the model more robust, weighting can be applied. In weighting, each case or respondent is assigned a weight to reflect its importance relative to other cases or respondents (Malhotra, 2007). In the case of this research, it is expected that ratings for the more recently released movies are more accurate. Therefore, the number of months since the release of the movie is considered as a weight variable. The largest number of months in the data set is 23 (Avatar). In the model, the rating for the 3D effect in this movie should be weighted less than a rating for the 3D effect in The Darkest Hour (1 month since release). Therefore, the number of months is transformed by 1/(number of months), which results in the more recently released movies having a higher value. Next, the values of the weight variables are calculated using a log-likelihood maximization. It shows that the log-likelihood function is maximized when the power is -.100 (see appendix E). Finally, the weight values are saved and used in the further estimation of the model.

Aside from the independent variables which are needed to test the hypotheses, one additional variable is added, namely PWG. This represents the perception of the consumer towards having to wear glasses. This variable is added to check if this variable, other than the variables resulting from the hypotheses, plays a role in explaining the appreciation for the 3D effect in a movie.

The outcomes of the first estimation of the model are presented in the table 4. In the following paragraph, the outcomes of this model are inspected for multicollinearity and normality and if necessary, the model is re-estimated.

1_{A model with unique intercepts for each movie (OLSDV) was also estimated. However, this model was not}

(33)

4.2.3 Validation and re-estimation

Inspecting for multicollinearity is important as it makes the parameter estimates unreliable (Leeflang et al. 2002). In order to check for multicollinearity, the Variance Inflation Factor (VIF) has been calculated. When this value exceeds 10, this can become a problem which should be treated (Mason and Perreault, 1991). In this model, the largest VIF value belongs to GNR1 (14.472). No other variable

exceeds the value of 10. However, multicollinearity occurs in pair. In order to find out which variable highly correlates with GNR1, a Pearson correlation table is produced using the variables with the highest VIF values (see appendix F). This shows that the Family/Adventure dummy (GNR1)

significantly correlates very highly (.794) with the variable which indicates whether a movie is animated or not (ANI). When looking at the movie list in appendix B, it shows that with an exception of one movie (Avatar) all animated movies are Family/Adventure, which explains the high level of correlation between GNR1 and ANI. One solution for the multicollinearity in this model is to simply remove the ANI variable. However, then the effect of an animated Sci-fi/Fantasy (GNR4) can no longer be explained. Therefore, first a different solution is applied, namely creating new predictors (Leeflang et al. 2002). The fi/Fantasy dummy is split up in to two different variables; Animated Sci-fi/Fantasy and Unanimated Sci-Sci-fi/Fantasy. For the Family/Adventure genre this is not necessary, as the only unanimated Family/Adventure movie is already excluded from analysis due to insufficient ratings (Cats & Dogs 2), meaning all 3D Family/Adventure movies in this model are animated. The model is estimated again removing the ANI variable and the replacing the GNR4 with the two new Sci-fi/Fantasy variables (GNR4A and GNR4UNA).

Table 4: Estimate outcomes

F-value P-value R2 Adjusted R2

97.222 .000 .510 .505 Parameter estimate Std. Error Stand. Beta t- value p-value VIF Constant -16.004 7.878 -2.032 .042

Total number of 3D movies seen(N3D) .076 .078 .020 .965 .335 1.191

Perception on General Relative Advantage of 3D cinema(PGRA) 4.641 .304 .362 15.272 .000 1.501 Perception on the Quality of the Movie(PQM) .457 .030 .331 15.295 .000 1.248 Perception on the Height of the Surcharge(PHS) .176 .379 .010 .465 .642 1.173 Perception on the Justification of the Surcharge(PJS) 1.526 .268 .128 5.696 .000 1.341

Genre: Family/Adventure(GNR1) -8.011 4.392 -.134 -1.824 .068 14.472

Genre: Action/Adventure(GNR2) -.859 2.714 -.013 -.317 .752 4.178

Genre: Horror/Thriller(GNR3) 3.233 2.999 .032 1.078 .281 2.351

Genre: Sci-fi/Fantasy(GNR4) 5.660 2.910 .099 1.945 .052 6.905

Animation(ANI) 15.669 2.819 .280 5.558 .000 6.776

Budget of the movie(BGT) .012 .012 .033 1.017 .310 2.864

Runtime of the movie(RNT) .061 .059 .049 1.023 .307 6.182

IMDb score of the movie(IMDb) -1.937 .810 -.067 -2.392 .017 2.090

Perception on having to Wear Glasses(PWG) 1.727 .319 .129 5.413 .000 1.523

(34)

Modeling the Added Value of the 3D Effect in Feature Films 34 The output of the re-estimation shows that now the highest VIF value is 7.415 (see appendix G). This value is still not very low but it is definitely lower than the threshold value of 10 (Mason and Perreault, 1991), which indicates that multicollinearity is no longer a serious issue for this model. Next, the error terms of the model are inspected on normality. When non-normality is detected, like multicollinearity, this makes the parameter estimates unreliable (Leeflang et al. 2002). Testing for normality is done by using the Kolmogorov-Smirnov and the Shapiro-Wilk tests and a Q-Q plot and histogram. These two visuals are presented in figure 4 and show good signs of normality. However, the Kolmogorov-Smirnov and the Shapiro-Wilk are significant (.000) which means that the assumption that the error terms are normally distributed cannot be made. In order to remedy the non-normality, the dependent variable has to be transformed. For this model, an appropriate method for transformation is to square the dependent variable2.

When the dependent variable is squared, the Shapiro-Wilk test is still significant, with a p-value of .008. However the Kolmogorov-Smirnov is now insignificant with a p-value of .200. Also, the Q-Q plot and the histogram of the new residuals, which are presented in figure 5, respectively show a better fit on the line and a less skewed distribution. Therefore, it can be assumed that the residuals have a normal distribution.

2_{A Square Root-, Inverse- and Log Transformation proved to be insufficient in treating the non-normality issue} Figure 5: Q-Q plot and Histogram of residuals after transformation

(35)

Modeling the Added Value of the 3D Effect in Feature Films 35 The VIF values in the re-estimated model have been unchanged as the independent variables remained unchanged in the process of re-estimation. Therefore, now both multicollinearity and non-normality are no longer an issue and the estimated parameters of the new model can be interpreted. These are presented in the table 5.

The ANOVA analysis in table 5 provides the statistical test for the overall model fit. The p-value of .000 indicates that the model as a whole is significant. Another goodness-of-fit measure is the R-square, or coefficient of determination. The adjusted R-square is adjusted for the number of independent variables and the sample size to account for diminishing returns (Malhotra, 2007). For this model, the adjusted R2 is .492, meaning that around 50% of the variation in the dependent variable can be explained by the independent variables which are in the model. Overall, this model has a good fit with the data.

The parameter estimate, or regression coefficient, represent both the type of relationship (positive or negative) and the strength of the relationship between the dependent and the independent variable (Hair et al. 2005). Table 5 shows the parameter estimates for the variables in this model. When the p-value of a variable is below 0.05 (α), this gives a high level of assurance that this variable can be assessed as a predictor of appreciation for the 3D effect in a 3D movie. In this model, this is the case for eight variables excluding the constant, which is also significant (.000). These variables are; (1) the consumers’ perception on the general relative advantage of 3D cinema (PGRA), (2) their perception on the quality of the movies they have rated (PQM), (3) their perception on the justification of the price surcharge (PJS), (4) the Family/Adventure genre (GNR1), (5/6) both the animated and unanimated Sci-fi/Fantasy genre (GNR4A/ GNR4UNA ), (7) the IMDb rating of the movie (IMDb) and (8) the perception of

ANOVAa,b,c R-square

F-value P-value R2 Adjusted R2

92.258 .000 .497 .492 Parameter estimate Std. Error Stand. Beta t- value p-value VIF Constant -3642.273 832.918 -4.373 .000

Total number of 3D movies seen(N3D) 13.580 8.299 035 1.636 .102 1.191

Perception on General Relative Advantage of 3D cinema(PGRA) 504.254 32.130 .377 15.694 .000 1.501 Perception on the Quality of the Movie(PQM) 45.386 3.159 .315 14.367 .000 1.248 Perception on the Height of the Surcharge(PHS) 16.647 40.104 .009 .415 .678 1.173 Perception on the Justification of the Surcharge(PJS) 170.560 28.325 .137 6.022 .000 1.341

Genre: Family/Adventure(GNR1) 952.885 332.420 .153 2.867 .004 7.415

Genre: Action/Adventure(GNR2) -35.257 286.933 -.005 -.123 .902 4.178

Genre: Horror/Thriller(GNR3) 431.550 317.036 .041 1.361 .174 2.351

Genre: Animated Sci-fi/Fantasy (GNR4A ) 2683.041 447.721 .276 5.993 .000 5.495 Genre: Unanimated Sci-fi/Fantasy (GNR4UNA) 719.323 307.624 .110 2.338 .020 5.800

Budget of the movie(BGT) .266 1.254 .007 .212 .832 2.864

Runtime of the movie(RNT) 9.286 6.290 .072 1.476 .140 6.182

IMDb score of the movie(IMDb) -212.564 85.601 -.070 -2.483 .013 2.090

Perception on having to Wear Glasses(PWG) 114.059 33.727 .082 3.382 .001 1.523

Marijn de Boer June 12

Marijn de Boer

“Modeling the Added Value of the 3D Effect in Feature Films”

Author

Marijn de Boer

Schuitendiep 1001

9712KD Groningen

T 0617250301

E m.de.boer.31@student.rug.nl

S 1933655

University of Groningen

Faculty of Economics and Business

Master Thesis Business Administration

Specialization: Marketing Management & Marketing Research

Supervisor: Dr. J.E. Wieringa

Second Supervisor: Alec Minnema

Company: Wolff Groningen

Management Summary

Preface

Table of contents

1 Introduction

1.1 Background of the problem

1.2 Wolff Cinema Group

1.3 Research objective

1.4 Theoretical relevance

1.5 Thesis structure

2 Theoretical framework

2.1 Consumer characteristics

2.2 Movie characteristics

2.3 Conceptual Model

3 Research design

3.1 Plan of analysis

3.2 Data and variables

4 Results

4.1 General overview

4.2 Model 1: Ordinary Least Squares

∑