THE EFFECT OF MOVIE REVIEWS ON MOVIEGOERS’ INTENTION
TO SEE A MOVIE IN THE CINEMA
Research on the importance of online consumer reviews on the success of a
movie in the cinema depending on moviegoers’ involvement with movies and
their susceptibility to interpersonal influence
Master thesis, Msc Marketing, specialization Marketing Management
University of Groningen, Faculty of Economics and Business
January 14, 2019
2
ABSTRACT
3
TABLE OF CONTENTS
1 Introduction and Statement of the problem ………4
1.1 Word of mouth and Online Consumer Reviews ………...4
1.2 Online consumer reviews in movie industry ……….….5
1.3 Moviegoers’ intention to see a movie in the cinema ………..6
1.4 Movie reviews: sender and valence ………...7
1.4.1 OCR type of sender ………..8
1.4.2 OCR valence ………9
1.5 Research question ………...9
2 Literature Review ………10
2.1 OCR type of sender versus intention to see a movie in the cinema ………10
2.1.1 Amateur reviews ………10
2.1.2 Professional reviews ………..11
2.1.3 Amateurs versus Professionals ………..11
2.2 Review valence versus intention to see a movie in the cinema ………..13
2.3 Relative effect of OCR type of sender on OCR valence versus intention to see a movie in the cinema ………15
2.4 Impact of consumer characteristics ……….16
2.4.1 Moderating role of involvement with movies ………..16
2.4.2 Moderating role of susceptibility to interpersonal influence …………17
2.5 Conceptual framework ………19
3 Research design or methodology ………20
3.1 Research design ………..20
3.2 Population & Sample ………..20
3.3 Experimental procedure ………...20
3.3.1 Design of the stimulus ………..21
3.4 Operationalization: Variables and Table ………22
3.5 Results of the manipulation checks ………24
3.6 Plan of analysis ………..25 4 Results ………27 4.1 ANOVA ………...27 4.2 Regressions ………....28 4.3 Discussion of results ………..30 5 Conclusions ………...33 5.1 Summary of findings ………33 5.2 Managerial implications ………....35
5.3 Limitations and further research ………...36
References ………..37
4
1 – INTRODUCTION AND STATEMENT OF THE PROBLEM
The spectators usually go to the cinema in order to see a specific movie. Making the decision of which movie to see right there while watching the billboard is not an easy thing. One of the reason to know what movie we want to attend is because different media informed us or suggested us that a movie that fits our interest is available in our nearest cinema. Most likely we have seen an advertisement on TV, on the internet or at the bus stop. However, it is also possible that it has been recommended to us by someone we know. Sometimes we do not realize but we, as potential customers, are receiving all kind of recommendations every day. In the same way, when we leave the cinema after watching the movie, we can hear the audience commenting if they liked it or not. These same spectators will be those who tell their friends and relatives if they think it is worth paying a ticket to see this movie.
Commenting our opinion about what we have just seen is something natural to the human being (Godes and Mayzlin, 2004), especially in the entertainment industry where each person has subjective opinions. Indeed, the environment of cinema has always been closely linked to word of mouth (Elberse and Eliashberg, 2003).
The principal aim of this study is to learn more about the relationship between worth of mouth, through so-called online consumer reviews, and consumer behavior in the cinema environment, measured as people’s intention to go to the movies.
1.1 - Word of mouth and Online Consumer Reviews
To get started, the term word of mouth need to be defined since it will be the main topic of this research. Word of mouth (WOM) is a powerful type of communication among consumers about products and services (Liu, 2006). Its greatest value is to influence consumers’ use and adoption of these goods (Godes and Mayzlin, 2004) and it has been perceived as a standout amongst the most compelling resources of information transmission since the start of human culture (Duan, Gu and Whinston, 2008).
The term has evolved along with society. The arrival of the internet era has allowed this conventional form of oral communication to take a leap in a new media in which to spread, becoming person-to-person virtual communication or eWOM (Yeap, Ignatius and Ramayah, 2014). The term eWOM aludes to positive or negative statements about products, services or companies that a customer makes and can be viewed by other consumers through the internet (Hennig-Thurau, Gwinner, Walsh and Gremler, 2004). Every year the activity of customers who write their opinions on the internet increases through the online consumer reviews (Zhang, Craciun and Shin, 2010).
5
or service. Furthermore, in some occasions we can find a quantitative evaluation in addition to the qualitative one. In this way we can see, next to the text, the average score that other users have given to the product and, therefore, get an idea of whether it will be useful for us (e.g. one to five stars rating in movies or a scale from zero to ten in a restaurant). This online product quantitative rating is part of a total average that is shown next to the product or service to inform about its quality according to online reviewer community’s product evaluations (Sridhar and Srinivasan, 2012)
Besides, consumer reviews are not just a text or a number but they have a real influence that plays an importan role in many industries (Austin, 1984). People are increasingly influenced by online word of mouth on their consumer purchases (Shankar and Batra, 2009), so it is receiving attention of many researchers (Zhang, Craciun and Shin, 2010).
One of the reasons of playing this role is because there are areas where WOM is very useful. For example in new product development it plays a prominent role because consumers do not know the products properly and they need information to decide if purchasing or not (Mahajan, Muller and Kerin, 1984). Another reason why WOM is important is due to the experience goods markets, such as theaters or movies (Basuroy et al., 2003; Caves, 2000), because they are activities where consumers can not know if the experience is worth it before consumption (Boatwright, Basuroy and Kamajura, 2007). In these cases, reviews and opinions are considered information richness and truthfulness because they come from other clients who have been in the same situation previously and have experienced the outcomes (Dellarocas, 2003). However, it should be taken into account that not all criticisms have the same influence on consumers (Moon, Bergey and Iacobucci, 2010).
1.2 - Online Consumer Reviews in movie industry
This research focuses on the movie industry. In recent years, spending in the entertainment industry has been increasing. Cultural events such as shows, music, theaters, movies or books are frequent in our routine (Vogel, 2001). According to the website statista this market is expected to reach a worth of 2,2 trillion U.S. dollars in 2021 and the media market is going to grow by more than twenty five percent between 2014 and 2020 in the USA. Therefore, we are talking about a booming market.
Specifically in the film industry, in 2017 the global box office for all films released reached a record high of $40,6 billions, five percent more than in 2016, according to the statistics provided by the MPAA theatrical statics. Besides, the number of cinema screens increased eight percent globally in 2017, due in large part to the huge growth in the Asia Pacific region.
I have chosen the movie industry as my research context because industry experts consider WOM as a fundamental factor in the final success of a film due to its condition of experience good (Elberse and Eliashberg, 2003) and is believed that it strongly influence people’s decisions (Austin, 1989). It is even considered that the role of critics may be most prominent in the film industry than in any other (Eliashberg and Shugan, 1997).
6
the one before the premiere (first public screen), can have a strong impact on movie sales (Lehmann and Weinberg, 2000).
Consumers find difficulties in analyzing the preconsumption quality of these experience products (Holbrook and Hirschman, 1982) because of their intangible and experiential nature which makes it difficult to judge before trying them (Liu, 2006). They are based on sensory aspects, not conveyed by tangible attributes (Desay and Basuroy, 2005).
Every year, hundreds of movies are released by both Hollywood and independent filmmakers, and many consumers depend on the reviews of other consumers when deciding what movies are worth their time due to the huge quantity of them (Yeap, Ignatius and Ramayah, 2014), and the short life span that films have in theatres (Eliashberg and Shugan 1997). Consumers do not have neither the time nor the desire to see all movies so it makes consumer reviews a fundamental fact to the ultimate success of a movie because it helps moviegoers to make decisions (Liu, 2006). As a result, it directly affect the box office revenue (Elberse and Eliashberg, 2003).
Therefore, if there is an industry where critics have played a fundamental role, this is cinema. As Lopate (2006) states, it is demonstrable that in the last 50 years there has been no other art industry where the critics have put more energy, passion and analytical judgment and it is because movies tend to receive great public interest (Liu, 2006).
1.3 – Moviegoers’ intention to see a movie in the cinema
So far we have talked about the importance of word of mouth and its role in the movie industry. Next step is understanding if this term has a real influence on the people who go to the movies. The dependent variable of this research is the success of a movie within moviegoers measured in their intention to see it in the cinema. Extensive evidence indicate that other people opinions have an influence in this final decision of a consumer.
Social influence theory has an important part to do with this relationship. In a given social situation it is proved that individuals experience conformity pressures from the rest of the group (Sridhar and Srinivasan, 2012). As Cialdini and Goldstein (2004) state, we are affected by the acts of other people and not only by our own. Parks, Sanna and Berel (2001) go further and say that it is not necessary to have a direct observation of the other group members to be influenced by their opinion, simply communicating a rule in writing induces conformity. We are affected by the influence of people we do not know (Darley and Latane, 1966) and even by abstract reference groups (Cohen, 2003). If we look at the social influence process as a structure (Rashotte, 2009) we can see that consumers combine their own ideas with those of others when making a decision (Sridhar and Srinivasan, 2012).
Hence, social influence theory states that people are affected by opinion of others generating an intention of following their insights (Sridhar and Srinivasan, 2012) and, therefore, can generate a purchase intention in people affected by this social influence.
7
So, there are quite industries where this relationship has been proven to be weak. We have to make sure that this relationship is strong enough in the cinema environment.
Shugan and Swait (2000) developed a specific model for the movie industry and proved that moviegoers state intent-to-view with ultimate box office performance is linked. So, we can say that there is a direct relationship in this industry. This model agree with Ajzen (1985) theory of planned behavior which states actions to be preceded by intentions. Not all the intentions that we plan are fulfilled but they are a reliable indicator about our future actions according to these authors. So intention to see a movie is likely to end up in the action of going to see it.
Movies are an experience good because individuals do not know what the value of the movie is to them until they experience it (Shapiro and Varian, 1999). Individuals choose and use movies only for the experience and enjoyment (Holbrook and Hirschman, 1982), so consumption experience is an end itself (Reddy et al., 1998). When the spectators watches a movie, they enter into a purchase agreement with little knowledge of the particular movie; the form may be familiar, but not the content (Reddy et al., 1998).
Lack of knowledge about a particular movie is the reason that may lead the moviegoers to search for additional information before making a final decision (Chang and Ki, 2005). Because their previous knowledge is not enough, it is a natural thing for the audience to seek advice from other sources to find additional information in order to avoid risks such as wasting money or time (Rook and Hoch, 1985). Therefore, the feature that movies have of being an experience product is closely related to the purchase decision process of moviegoers which consists of seeking other opinions (Chang and Ki, 2005).
1.4 – Movie Reviews: Sender and Valence
There has been much research conducted to understand film reviews and how they affect consumers. Prior academic researches studied many variables such as the relationship between WOM and box office revenue (Liu, 2006; Neelamegham and Chintagunta, 1999), the genre of the movies (Desay and Basuroy, 2005), star power (Basuroy, Chatterjee and Ravid, 2003; Desay and Basuroy, 2005), valence (Liu, 2006), volume (Duan, Gu and Whinston, 2008) or dispersion of critics’ reviews (Godes and Mayzlin, 2004), as well as studies about platforms where the review has been written (Yeap, Ignatius and Ramayah, 2014) or the real role of critics as influencers or predictors (Basuroy, Chatterjee and Ravid, 2003).
8
1.4.1 - OCR type of sender
Thanks to the internet, we are facing a huge growth of information about movies in the form of electronic word of mouth in recent years (Chakravarty, Liu and Mazumdar, 2010). This has allowed many websites to be born and cover the new demand. They offer us information about the movies, discussion forums or critical reviews among other functions. Within the most famous, The Internet Movie Database (IMDB), Rotten Tomatoes (.com) and Filmaffinity (.com) stands out. This sites have become places of reference for moviegoers because unlike advertising, you can find negative reviews on them (Chen and Xie, 2005) and comes from a different source to movie studios, so they are supposed to be more credible and trustworthy (Murray, 1991). These websites use to provide both professional review critics and messages from amateur moviegoers.
There are differences between both type of review senders (Chakravarty, Liu and Mazumdar, 2008), because ordinary consumers and expert critics have dissimilar criteria when forming their opinions (Holbrook, 1999). Sometimes, amateur critics are at odds with professional ones because of their dissimilarities if we talk about preferences and experiences (Chakravarty, Liu and Mazumdar, 2008; Holbrook 1999; Wanderer 1970), so their opinions could affect readers differently.
In this research, we evaluate how both types of online consumer review sender affect the moviegoers’ intention to see a movie in the cinema. It is an interesting variable because moviegoers may be influenced in a different way or, at least, with a different strenght depending on the type of sender.
o Professional reviews: Much research has been done on the evaluation of professional critics (e.g. De Silva, 1998; Jedidi, Krider, & Weinberg, 1998; Litman, 1983; Litman & Ahn, 1998; Litman & Kohl, 1989; Prag & Casavant, 1994; Ravid, 1999; Sochay, 1994). Most of these authors state that there is a direct relationship between their opinions and movie success. Austin (1984) suggested that some of the functions of the critics are helping individuals in making a movie choice, understanding the content of the movie, developing an initial opinion of the film, and communicating movie information to others.
9
1.4.2 - Valence of OCR’s
Valence is one of the most important WOM attributes that have been studied previously (e.g. Anderson, 1998) but is not the only one. Other measures like volume, duration, intensity and dispersion are also important when evaluating consumer reviews (Eliashberg et al., 2000). Valence is defined as the semantic orientation of the review (Kennedy and Inkpen, 2006) and can be positive, negative or neutral, like a sentiment classification of opinions. It refers to the positive or negative nature of the statements in a message (Buttle, 1988). Positive valenced reviews indicate an acceptable performance or a satisfying experience while negative ones mean a non aceptable performance and are considered as a customer complaint (Weiner, 1985; Singh and Pandya, 1991).
It is a relevant variable for this research because of its interaction with the type of sender. A negative review may affect a moviegoer in a different way depending on whether it has been written by a professional reviewer or by an amateur community due to the differences that exist among these two groups of type of senders. Exactly the same happens with positive ones. In addition, it is interesting to check if consumers pay more attention to positive or negative criticism.
1.5 - Research questions
Combining the above, the main research question in this thesis is:
What is the effect of OCR type of sender and OCR valence on consumer intention to see a movie in the cinema?
So, the main purpose of this study is to measure the relationship between online consumer reviews and moviegoers final intention to see a movie.
To answer this, the following research questions will be studied in this thesis:
- How does review type of sender impact moviegoers’ intention to see a movie in the cinema?
- How does review valence impact moviegoers’ intention to see a movie in the cinema? - How does the interaction effect between review type of sender and review valence impact
moviegoers’ intention to see a movie in the cinema?
- Are there any consumer characteristic that might play a role in the relationship between movie reviews and moviegoers’ intention to see a movie in the cinema?
- What is the relative effect of these variables on moviegoers’ intention to see a movie in the cinema?
Structure of the thesis
10
2 - LITERATURE REVIEW
This chapter deals with how OCR type of sender and OCR valence affect the dependent variable, moviegoers intention to see a movie in the cinema. In addition, we talk about how the moderators, involvement with movies and susceptibility to interpersonal influence, impact these relationships. All relationships between variables are explained. To conclude, the conceptual model and the hypothesis are shown.
2.1 OCR type of sender versus intention to see a movie in the cinema
Reviewers, both professional and amateur, have several functions. They warn us about the release of a movie, create reputation, offer us an experience just by reading the review and influencing our preferences (Cameron, 1995). Besides helping the consumer to choose a movie, reviewers may help him to understand it, to form his own opinion and to communicate with criteria in social settings (Austin, 1984).
This influence of reviewers on consumers allows them to have a real impact on the final box office performance (Eliashberg and Shugan, 1997). As these scholars state, movie reviewers have the potential to shape the final revenue of a film in the theatre.
Talking in box office revenue terms, several studies show that critics play a role as both, influencers and predictors (Basuroy, Chatterjee and Ravid, 2003; Boatwright, Basuroy and Kamakura, 2007). In this research thesis we are interested in the part of influencers.
In the introduction part we have talked briefly about the different types of sender. Taking into account these differences, the reviews of each group may affect the success of a movie differently. In this part we are going to talk in depth about the relationship between both amateur and professional reviewers and the moviegoers’ intention to see a movie regardless of the valence. Review valence is not going to come into play until next point.
2.1.1 – Amateur reviews
Yet, Shankar and Batra (2009) state that eWOM by other moviegoers is progressively gaining importance in consumer decisions. There are websites specialized on cinema where we can find all kind of information from cast and plot to recommendations whether the movie merits viewing in cinemas (Chakravarty, Liu and Mazumdar, 2008). In websites like IMBD (imbd.com) or Filmaffinity (filmaffinity.com), amateur critics have a space to write their reviews. No requirement is necessary to write, except for being registered in the platform. This means that anyone can be an amateur critic.
These kind of reviews represent the mass tastes, because it is written by regular moviegoers and it is a review told from a personal and informal point of view (Holbrook, 1999). Its function is telling the rest of users their own experience with the movie and to recommend others to see it or not.
11
are receiving many visits and generating a significant amount of traffic on the Internet. In other words, amateur reviews are generating interest. These researchers continue saying that online amateur reviews likely play increasingly important roles in helping people make more informed decisions about the movies they want to watch. To conclude, their results confirm that Yahoo! Movies’ online amateur reviews have a positive and statistically significant influence on the amount of people going to the movies regardless of their valence (review valence does not play a role here). Specifically, a one-point increase in the grade rating of the aggregate score can induce 4–10% more people to attend the cinemas.
2.1.2 – Professional reviews
With the term professional we refer to critics who are specialized in the movie industry (West and Broniarczyk, 1998) and offer their particular opinion about a product. Unlike amateur moviegoers, they have more tools to evaluate in deep the artistic and technical aspects, because of their background and demand more details from the movies (Weiman, 1991). Professional reviewers are opinion leaders who are recognized with knowledge and experience in a particular area (Assael, 1984). Yet, there is debate among the scholars because some of them consider professional critics as a very important part in the success of a film while others consider that they hardly exert an influence. According to Chakravarty et al. (2010) this is because critical reviews only influence certain segments of moviegoers but not others.
There has always been an influence of professionals in consumer response to movies, cinema history is full of examples that show it. Zhang and Dellarocas (2006) state that from the earliest periods where films were made, even in the silent film era, film criticism has been recognized as an art and became a profession over time. The period from 1950s to 1970s was called ‘the golden age of movie criticism’ by Lopate (2006). There were some critics who were extensively read and were able to generate public agitation as was the case of Andrew Sarris or Pauline Kael. Zhang and Dellarocas (2006) continue with another example, the television program called ‘Siskel and Ebert’. The program turned its reviewers into a very common name between the society and has been successful for twenty years. It is difficult to imagine that these professional critics have been so popular for so long without having an influence on the decisión of amount of people going to the movies and, therefore, on the box office performance of the films.
Zhang and Dellarocas (2006) conclude with due to this reasons, movie industry studios consider that quality good reviews from professional critics help to attract moviegoers. So they use to quote sentences like “spectacular,” “thrilling,” and “two thumbs up” in their movie advertisements to get moviegoers’ attention.
2.1.3 – Amateur versus Professional reviews
We have seen that both professional and amateur reviewers can affect moviegoers consumer decisions as extenze literature has proved. In this section, the two types of reviewers are compared to each other deepening in their differences.
12
In the cinema industry, the opinion leaders are the movie critics. These professionals generally have more expertise or knowledge about the industry than the general public (Assael, 1984). They use to watch movies earlier than the rest of the spectators, so they can write about first hand information and, thereby, attract the attention of moviegoers (d’Astous and Touil, 1999). Depending on their positive or negative assessments, some people may be influenced (Eliashberg and Shugan, 1997) because of their capacity as influencers and endorsers (Ohanian, 1991). One fundamental aspect of professional critics is that their final reviews are more complete and well nuanced than the ones of amateur moviegoers (Chakravarty, Liu and Mazumdar, 2008). As said before, they have an important role in the launch phase of the films because they are usually invited to pre-release sessions so they have information before the rest of the users. At this time applicable WOM has not yet scattered among people in general, so professional movie reviews are seen as a fundamental source of data for moviegoers (Boatwright, Basuroy and Kamakura, 2007). This makes their reviews to have a greater impact (Boatwright, Basuroy and Kamakura, 2007) and allow them to shape, to some extent, the box office success (Litwak, 1986).
Conversely, in the opinion of Kirby (2000), amateur reviewers don’t have that power. They need to be a crowd to influence consumers with their opinions. He states that a consumer may not trust just one non expert reviewer opinion. But if 9 out of 10 amateur reviewers agree, probably they are right. Chen, Wu, and Yoon (2004) confirm that an increase in information sources could lead to more trust in the case of amateur reviewers. They show that as the number of consumer reviews increases, online reviews seem more trustworthy.
This makes sense with Austin (1984) opinion about the elist or snob position. He argues that professional critics appreciate movies differently than do public moviegoers. A critic will evaluate a film based on its artistic merits, or perhaps the way it represents a sociological or ethical issue. The public, on the other hand, will rate a movie based simply upon how entertaining it is.
Therefore, there are differences between both types of review senders. So, when a moviegoer is facing a professional or an amateur review we hypothesize that:
13
2.2 Review Valence versus intention to see a movie in the cinema
In this section online consumer review valence come into play. It is going to be anayzed without taking into account the type of sender. The relative effect of type of sender on review valence will come in next section (2.3).
Valence of consumer reviews is defined as any statement about a product or service (Hennig-Turau, Gwinner, Walsh and Gremler, 2004), that captures the nature of the message, whether it is positive or negative (Liu, 2006). Other authors previous research has demonstrated that the valence of the consumer review has an impact on product evaluation and customer choice (Chevalier and Mayzlin, 2006; Chiou and Cheng, 2003; Doh and Hwang, 2009; Lee et al., 2009; Sen and Lerman, 2007), so it seems clear that valence influences consumer product judgement. These studies found a direct relationship between online reviews valence and products or services sales (e.g. positive relationship if valence is positive and negative relationship if valence is negative) (Li and Hitt, 2008).
In this way, valence affects the consumers’ intention to see a film (D’Astous and Touil, 1999; Liu, 2006) because it influences consumer evaluation and ultimate purchase decision (Duan, Gu and Whinston, 2008). Moreover, it should influence the reader in the same direction of the valence because it provide a signal about the movie (Chakravarty, Liu and Mazumdar, 2008). According to Reinstein and Snyder (2005), when a movie is rated with a high punctuation by professional critics or amateur users, this is reflected later in an increase in revenue. This is because these high ratings have generated positive word of mouth between the moviegoers and this buzz contributes to more people going to theatres and having a positive experience (Sridhar and Srinivasan, 2012). Positive valenced reviews enhance the intention of watching a movie for moviegoers (Sochay, 1994; Prag and Casavant, 1994). Corroborating this theory, Austin (1984) argues that film attendance is bigger when the general evaluation of the public matches with the one of the professional critics, so we can conclude saying that regardless of the type of sender, a positive review valence leads to an increase in moviegoers’ intention to see a movie . In this research we are going to consider three types of valence: positive, negative and neutral. A positive consumer review gives us a recommendation for purchasing a product direct or indirectly. Negative consumer reviews appears when product denigration, rumor or private complaining are involved (Liu, 2006). So from the point of view of a consumer, positive review valence enhances expected satisfaction of a movie while negative valence reduces it. With a positive online consumer review a reviewer can benefit the success of a movie (in a small part) while with a negative one the reviewer can hurt that success (Levin, Levin and Heath, 1997). We talk about neutral valence when there is no clear direction in the content of the review. Therefore, as we can see in Image 1, the intention to see a movie in the cinema would follow an ascending line (red line) that goes from negative reviews to positive ones.
14 Image 1 – Moviegoers’ expected intention to see a movie when reading an OCR
It must be made clear that this graph is only an expected representation of moviegoers’ intention to see a movie. No data analysis has been performed yet.
Some previous research states that negative comments have more influence on consumers than positive ones (Ahluwalia and Shiv, 1997). Researches in various disciplines have called it the “negativity effect” (Ahluwalia, 2002). Basically it means that negative information has a stronger impact than positive information on evaluations of readers.
Arndt (1967) was the first researcher who explored WOM valence. He studied the influence of positive and negative WOM on purchase decision in a new food brands context. He found that the effect of negative WOM on reducing sales could be twice as strong as the positive one in increasing sales. In the context of entertainment, Chevalier and Mayzlin (2006) found that negative reviews had a greater impact on book sales than positive ones. Negative information is more diagnostic in categorizing the target into a certain category than positive information (Skowronski and Carlston, 1989). In other words, in our movie context, negative reviews about a movie help consumers categorize the product into a low-quality movie more so than positive reviews help them categorize the movie into a high-quality one (Bone, 1995; Herr et al., 1991). When consumers are neutral, negative reviews tend to be more salient than positive reviews (King et al., 2014). Tsao (2014) states that regarding past researches, negative views are less common than positive ones, and therefore information seekers perceive in them higher diagnosticity. As a result, negative reviews exert greatest influence on the judgement and decisions of consumers.
Therefore, as we can see in Image 1, the ascending line from negative reviews to positive ones would not be a straight line (constant) because its growth would not remain stable. The slope would be more inclined between negative reviews and neutral ones than between neutral reviews and positive ones.
15
2.3 Relative effect of OCR type of sender on OCR valence versus intention to see a movie in the cinema
As discussed in the online consumer review section, there are differences between both types of senders, regardless of the review valence. Taking into account these differences, it is possible that different senders exert a different effect on the relationship between movie review valence and moviegoers’ intention to see a movie in the cinema. In this relationship, we take into account both, the type of sender and the valence of the reviews.
As said before, online reviews are seen as a persuasive source of information in the consumer decision making process, influencing consumers’ attitudes and their purchase behavior (Bickart and Schindler, 2001; Chevelier and Mayzlin, 2006; Park and Kim, 2008; Senecal and Nantel, 2004). This persuasiveness of reviews has been partly explained by its source credibility (Bickart and Schindler, 2001; Brown et al., 2007).
Reviews are written by different types of sources (Mackiewicz, 2010), ranging from individuals with little knowledge about the product (i.e. normal consumer – amateur reviews) to individuals who obtained knowledge about a product as part of their profession (i.e. experts – professional reviews). Willemsen, Neijens and Bronner (2011) state that amateur consumers and experts have differential effects on perceived source expertise and trustworthiness. This makes them to have a different range of credibility. So in the case of films, moviegoers don’t see the same intensity in the valence of every consumer review. This valence intensity depend on the type of sender and his credibility.
We have hypothesized in the previous section that the valence of both (professional and amateur) online reviews influence moviegoers’ intention to see the movie. But due to the differences in trustworthiness that exists between both type of online consumer reviews senders, this influence may not have the same strength on moviegoers’ intention to see a movie. In other words, a negative valence review written by a professional may affect in a different strength to a moviegoer than a negative valence review written by an amateur because of the different trusworthiness of the sources.
16
2.4 – Impact of consumer characteristics
So far, the relationship between the main variables of the study and have been proposed. However, some consumer characteristics might play a role and affect these relationships. It is possible that not all moviegoers’ are affected equally by movie reviews, so this is where personal characteristics of the moviegoers come into play. Two consumer characteristics are tested in this research, involvement and susceptibility.
2.4.1 – Moderating role of involvement with movies
The first consumer characteristic that is going to be studied is involvement with movies. Involvement is defined as the perceived personal relevance of a product based on the individual consumer’s needs, values and interests (Griffith, Krampf and Palmer, 2001). Involvement with movies refers to a customer’s understanding or recognition of the cinema environment (Fiske and Taylor, 1991). Therefore, the higher level the consumer consideration of the product is called high involvement and the lower level, low involvement.
There are two types of involvement according to Celsi and Olson (1988), i.e. situational and enduring involvement. The former is a temporary increase of interest that fluctuates within the time frame of a purchase decision; the latter is a stable mood that represents the customer’s interest over a significant period of time.
In our research, situational involvement does not make much sense because we are not interested in a temporary concern about movies but in a stable attitude to them (Lin and Chen, 2006). So we are interested in enduring involvement which reflects the individual stable response to cinema environment. Houston and Rotschild (1978) state that enduring involvement is caused by consumer’s personal subjective appreciation system and previous customer’s experience with the product.
A common assumption in consumer research is that the higher level of involvement in the purchasing task, the greater the effort devoted to experience the product or service (Bloch, Sherrell and Ridgway, 1986; Clarke and Belk, 1979). So, the more involved the consumer is with the movies, the more likely he will go usually to the movies.
H4: There is a positive relationship between moviegoers’ level of involvement with movies and their intention to see a movie in the cinema
Because people highly involved with the movies seek more information about them than those who are low involved, they will be more affected by online consumer reviews. They are more interested in the content of the reviews and all of the information related to the movies environment.
Bloch, Sherrell and Ridgway (1986) report significant positive correlations between personal involvement with products (personal computers and clothes) and propensity to search for information about these products and consume them.
17
that these high involved moviegoers look for online consumer reviews more often than low involved.
Gu, Park and Konana (2012) argue in their research that high involved consumers tend to believe reviewers depending on the quality of the consumer reviews. In contrast, low involvement consumers tend to conform to the perspective of reviewers regardless of the quality of the consumer reviews. So it seems that highly involved moviegoers are more affected by professional reviews than by amateur ones, while low involved moviegoers do not make a difference between both types of review senders.
When the number of arguments in a consumer review increases, Petty and Cacioppo (1984) state that under low involvement, people agree with the message more strongly regardless of whether the arguments are cogent or specious. Nevertheless, under high involvement, the quantity of compelling arguments increases persuasiveness but the amount of specious arguments does not. Therefore, the difference of veracity between professional and amateur reviewers may be different in high than in low movies involvement customers.
Therefore, moviegoers’ involvement with movies moderates the relationship between OCR type of sender and moviegoers’ intention to see a movie in the cinema.
H5: Individuals highly involved with movies will be more influenced by consumer review type of sender than individuals who are less involved
2.4.2 – Moderating role of susceptibility to interpersonal influence of moviegoer
The second consumer characteristic that is going to be studied is susceptibility to interpersonal influence. In this section of the research we discuss how this moderating variable might affect the relationship between review valence and moviegoers intention to see a movie. So far we have seen that there are reasons to believe that movie reviews are influential in the intentions of the moviegoers. Now we want to know if these reviews influence them in a different way depending on their personal susceptibility.
As explained in chapter 1, people are often influenced in their purchase decisions by the opinions of others (Bearden and Etzel, 1982), especially in uncertain situations (Mitchell and McGoldrick, 1966). This consumer trait varies across individuals and is related to other individual traits and characteristics (McGuire, 1968). It depends on the receptivity of each individual. Janis (1954) suggests that some people consistently are amenable to social influence while others are consistently resistant.
‘Susceptibility to interpersonal influence’ is defined as the need to identify with or enhance one’s image in the opinion of significant others through the acquisition and use of recommended products; the willingness to conform to the expectation of others regarding purchase decisions and the tendency to learn about products by observing others or seeking information from third parties (Bearden, Netemeyer and Teel, 1989). These authors concluded saying that an individual who is more easily influenced by others will give word of mouth communications more weight than an individual who is less susceptible to influence.
18
order to make better decisions or mantain their social image. D’Astous et al. (2005) argued the consultation of film reviews to be positively linked to consumer susceptibility of being social influenced. Schroeder research (1996) agree with them and said that consumers highly susceptible to interpersonal influence have been shown to be more influenced by others when making purchase decisions. Liao and Cheung (2001) said that consumers who are highly susceptible to interpersonal influence are more likely to be affected by WOM and Park et al. (2011) conclude saying that there is a positive relationship between interpersonal susceptibility and eWOM effect.
So, we can say that when a moviegoer makes the decision to see a movie because of positive reviews (or, on the contrary, he refuses to go to the cinema because of the negative ones) he is being influenced. In other words, he is showing susceptibility to change his attitude about his final decision of going to see a movie. So, the more susceptible an individual is, the more his intention to go to the cinema will be influenced when reading valenced consumer reviews. An interesting fact contributed by McGuire (1968) is that conformity and persuasibility exist across occurrences. That is, people who conform to one source on one issue will likely conform to other sources on other issues. So moviegoers who are easily influenced by professional reviewers, will also feel the susceptibility with reviews from amateur ones. There is no difference in the influence of OCR type of sender.
It is expected susceptibility to moderate the relationship between online consumer review valence and moviegoers’ intention to see a movie. Taking this into account, the following hypothesis is raised:
19
2.5 Conceptual Framework
Image 2 – Conceptual framework
Hypothesis recapitulation
H1: Regardless of the valence, a review written by a professional critic has a stronger effect on moviegoers intention to see a movie in the cinema than a OCR written by an amateur critic. H2a: Regardless of the type of sender, the more positive the valence of the review, the greater the moviegoers’ intention to see a movie
H2b: Regardless of the type of sender, negative consumer reviews have a greater negative influence on the intention to see a movie than the positive influence of positive consumer reviews on the intention to see a movie
H3: Taking into account review valence, the impact of professional critic reviews on the moviegoers’ intention to see a movie in the cinema is stronger than the impact of amateur ones. H4: There is a positive relationship between moviegoers’ level of involvement with movies and their intention to see a movie in the cinema
H5: Individuals highly involved with movies will be more influenced by consumer review type of sender than individuals who are less involved
20
3 - RESEARCH DESIGN OR METHODOLOGY
This chapter explains the study’s research design, the choice of subjects, the manner in which the subject is conducted, the operationalization of variables and the methods of data analysis. 3.1 - Research design
The proposed hypotheses were tested by conducting a 2 (review type of sender: professional and amateur) x 3 (review valence: positive, neutral and negative) between-subject factorial design (Bryman and Bell, 2007). Type of sender and review valence were manipulated. With three movies as the subject matter of the experiments. A total of six scenarios were established. Each subject run through only one condition (one scenario of the factorial matrix) for the experiment. Both consumer characteristics, involvement with movies and susceptibility to interpersonal influence, were measured using a questionnaire to test wheter the results were moderated.
Table 1 - Scenarios
3.2 – Population & Sample
The participants in the experiment belong to three different populations. The first group was reached by social media, and was not limited to a specific demographic. This group was about 35% of the total respondents. The second group was composed of students of the University of Groningen and it was almost all made up of students form 22 to 25 years old. This group was approximately 30% of the total respondants. Lastly, the third group was composed of acquaintances of the researcher and was not limited to a specific demographic. This group made up the last 25% of the total number of respondents.
The survey was distributed and resulted in 203 responses. After clearing the answers of the participants who did not finish the survey correctly (30) and of those who did not pass the attention check (8), a number of 165 final respondents was reached.
3.3 - Experimental procedure
21
Next, subjects are asked to complete some questions with the purpose of measuring how the respondents are affected by the reviews and how they feel about the movie that the reviews talk about. The dependent variable is measured using a 7-point Likert scale ranged from 1, extremely unlike, to 7, extremely like. The specific questions can be found in the operationalization table (Table 2). Following that were the moderators scales, 10 questions to test for involvement with movies and 8 questions to test for susceptibility to interpersonal influence.
Lastly, they were asked to answer some general and demographic questions. 3.3.1 – Design of the stimulus
Eighteen reviews were created based on real reviews from websites specialized in cinema. Three reviews are used per cell/scenario to create the stimuli instead of only one in order to make a more consistent impact on readers. Each review include a picture of a clapperboard, the reviewer’s name and, most important, his/her condition of professional or amateur reviewer on the left side. The score that they give to the movie is in the top right side, and, finally, the content of the review. The length of the review is under control because it can affect the influence they exert on viewers (Chevalier and Mayzlin, 2004). The length of each review is sent at four lines approximately.
OCR type of sender: Professional and Amateur are choosen as the criteria of review sender. Professional reviews are seen as cinema-relevant, well nuanced, understandable and persuasive, with sufficient reasons based on cinema facts about the movie. Amateur reviews are seen as emotional, subjective and no information except expression of subjective feelings (Assael, 1984). In this experiment there is no difference in the content of the review between amateur and professionals, just the condition in the top of the review. This will help to measure to what extent just mentioning the condition of the reviewer influence moviegoers regardless of the content of the review.
OCR valence: Positive, neutral and negative are choosen as the criteria of review valence. Positive ones encourage the reader to go to the cinema, negative ones discourage the reader and neutral ones don’t drive in any direction.
Pretest:
Before the main experiment, a pretest was conducted to check whether these reviews are perceived as intended. Ten subjects, who didn’t participate in the final experiment, were asked to classify each review according to its valence. Only the reviews that all of them perceived with the same valence were used for the main experiment.
22
3.4 - Operationalization of variables
In this section we explain the process of defining variables into measurables factors. The objective is removing ambiguity and research by defining all relevant variables such that they can be objectively measured.
Operationalization table
Variables Source Items Scale FA / RA
Moviegoers intention to see a movie in the cinema (DV)
Park, Lee & Han (2007)
1. How likely is it that you will go to the cinema to see this movie? 2. How likely is it that you will recommend going to the cinema to see this movie to your friends? 7-point Likert Scale ranged from 1, extremely unlike, to 7, extremely like. KMO=0.500 B = p<0.001 𝐸𝑉1=1.83 (91.32%) α = 0.904 Involvement with movies Zaichkows ky (1994)
1. Movies are important to me 2. Movies are boring to me 3. Movies are irrelevant to me 4. Movies are exciting to me 5. Movies mean nothing to me 6. Movies are appealing to me 7. Movies are fascinating to me 8. Movies are worthless to me 9. Movies are involving to me 10. Movies are needed to me
7-point Likert Scale ranged from 1, strongly disagree, to 7, strongly agree. KMO=0.914 B = p<0.001 𝐸𝑉1=5.39 (53.94%) α = 0.872 Susceptibility to interpersonal influence Bearden, Netemeyer & Teel (1989)
1. I rarely go to the cinema to see a movie until I make sure it’s worth it due to the opinions of reviewers
2. I rarely go to the cinema to see a movie until I make sure it’s worth it due to the opinions of friends
3. It is important that others like the movies that I go to see to the cinema
4. When choosing movies, I generally select those ones that I know others will approve of 5. If I go with other people to the
cinema, I often let them to choose the movie
6. I like to know what movies make good impressions on others
7. To make sure I choose the right movie to see, I often observe what movies others are choosing
23 Attention
check
Hidden in question 6 of the covariate susceptibility to interpersonal influence
1. To see if you are paying attention, please select number 1 in this question
Manipulation checks
1. I think the people who wrote the reviews were …..…. critics
2. I think the reviews that I have read were ………. Q1: 7-point Likert Scale ranged from 1, amateur, to 7, professional Q2: 1 negative; 7 positive Other control variables
Gender (male, female), Age (open question) , Frequency of going to the movies over three months (0,1-2,3+), Country of origin (open question) and English level (elementary, intermediate, advanced, proficency)
Table 2 – Operationalization table
Dependent variable - Moviegoers intention to see a movie in the cinema
Moviegoers intention to see a movie was measured on a 7-point numeric scale. The scale items were based on Park, Lee and Han (2007). There were two questions, shown in the operationalization table, and the scale ranged from 1, representing extremely unlike, to 7, extremely like. Depending on the stimulus to which the respondent has been exposed, the answers may vary.
Kaiser-Meyer-Olkin Measure and Bartlett's test of sphericity indicated that a factor analysis may be useful with the data. Factorial and Reliability analysis showed that these two ítems can be combined into one new variable (ouput in Appendix 1).
Covariate 1 – Involvement with movies
It was a state of participants in the experiment. Respondents did not enter the experiment with exactly the same level of involvement with movies, so the initial level of involvement is something that we need to keep under control. We can use this data to see how our treatment affect different groups of population. Depending on the high or low involvement with movies of the respondents, they may perform different in the experiment.
Involvement with movies was measured as the revised personal involvement inventory. The ten questions were based in an article developed by Zaichkowsky (1994) regarding levels of product involvement. Zaichkowsky’s scale is considered a valid measurement for product involvement (Goldsmith & Emmert 1991). A seven-point Likert scale was used to measure the answers.
24
Covariate 2 – Susceptibility to interpersonal influence
It was also a state of participants in the experiment. Respondents did not enter the experiment with exactly the same level of susceptibility to interpersonal influence, so the initial level of susceptibility is something that we need to keep under control. We can use this data to see how our treatment affect different groups of population. Depending on the high or low susceptibility to interpersonal influence of the respondents, they may perform different in the experiment. Susceptibility to interpersonal influence was measured as the extent to which an individual's moviegoer choices were influenced by other people. The eight scale items were based on a study published previously (Bearden, Netemeyer & Teel, 1989). A seven-point Likert scale was used to measure the answers.
Kaiser-Meyer-Olkin Measure and Bartlett's test of sphericity indicated that a factor analysis may be useful with the data. Factorial and Reliability analysis showed that these eight ítems can be combined into two new variables (ouput in Appendix 1).
Other control variables
The experiment could be affected by other characteristics of the subjects such as their gender, their age,the frequency with which they go to the cinema, their country of origin or their english level. These demographic questions were asked in the last part of the survey.
Manipulation checks
After the participants had read the movie reviews by both professional and amateur reviewers, they were asked to answer two questions to serve as a manipulation check. The objective was to confirm that respondents were accurately affected by the manipulated scenario that would have been randomly assigned to them. The questions inquired whether the movie reviews read by respondents were written by a professional or an amateur reviewer and whether the valence of the review was positive, neutral or negative. The manipulation questions can be found in the operationalization table (Table 2).
3.5 – Results of the Manipulation Checks
The manipulation check questions inquired whether the movie reviews read by respondents were written by a professional or an amateur reviewer and whether the valence of the review was perceived as positive, neutral or negative. So, it was important to have a statistically significant difference between the options.
To test the type of sender a T-test analysis was run. Results proved that there was a statistically significant difference between amateur and professional movie reviews (𝑀𝑝𝑟𝑜 =
1.88 , 𝑀𝑎𝑚𝑎𝑡 = 1.12) (𝑡(1,163) = −14.81, 𝑝 < 0.001) so the manipulation of this variable was
succesful and found to be effective in the main study.
25
3.6 - Plan of analysis
The data was analyzed using SPSS, a software designed by IBM to perform statistical analysis (Meyers, Gamst and Guarino, 2013).
Attention Check and cleaning data
A question was hidden inside the questionaire to check if the respondents were paying attention. This question simply asked them to select a concrete option among the possible answers. For the subjects to qualify, they had to select the option asked correctly. 8 respondents failed the attention check. Incomplete responses (30) were also cleared making the final sample a total of 165 answers.
Reversed negatively worded items
In order to work properly with SPSS, 4 items in covariate 1 set of questions, involvement with movies, were reversed to a positive sentence.
Factorial Analysis and Reliability Analysis (outputs can be found in Appendix 1)
After that, both a factorial and a reliability analysis were performed for the direct variable (moviegoers’ intention to see a movie) and the covariate variables (involvement with movies and susceptibility to interpersonal influence). The purpose is to reduce a large number of variables into fewer numbers of factors and to check if the scales produce consistent results. Kaiser-Meyer-Olkin output should be over 0,500 to indicate that a factor analysis may be useful with the data; and Bartlett's test of sphericity indicate that a factor analysis may be useful with the data when the result is a small value (less than 0.05) of the significance level.
Outputs per variable can be found in the operationalization table (Table 2). At this point, is worth saying that:
Covariate 1, involvement with movies
Although FA & RA recommendation was to use two factors, only Factor 1 will be used (which explains 53,94% of the variance). The reason is that the only difference was that Factor 2 contained the reversed negatively worded items and Factor 1 the positive ones. Factor 1 fits the conceptual definition of the covariate enough attending the Parsimonious model efficiency. This model states that no more paramenters than necessary should be used. Only the right amount of predictors needed to explain the model well.
Covariate 2, susceptibility to interpersonal influence
Although FA & RA recommendation was to use two factors, only Factor 1 will be used (which explains 51,60% of the variance). The reason is that the only difference was the manner items were redacted. Items of factor 2 were more direct questions. Factor 1 fits the conceptual definition of the covariate enough attending the Parsimonious model efficiency.
Analysis of Variance
26
Multicollinearity and Regression
Finally, a multicollinearity and a regression analysis were performed per step. Multicollinearity is a type of disturbance, and if present in the data the statistical inferences made about the data may not be reliable. Regression analysis is used to explain the relationship between both the independent and moderator variables and the dependent variable.
In this model, the independent variable named as review valence, EV2, had 3 levels (positive, neutral and negative), so it was transformed in 2 dummy variables: EV2a is negative valence or not and EV2b is positive valence or not. They were created always versus the base of neutral, which is the reference level.
Step 1: Simple model
m1 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏
Multicollinearity → No (Tolerance and VIF tables in appendix 3) Step 2: Base (simple + interactions)
m2 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎+ 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏 Multicollinearity → No (Tolerance and VIF tables in appendix 3)
Step 3: Base + Covariate 1
m3 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎 + 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏+ 𝛽4𝐶𝑂𝑉1+
𝛽5𝐸𝑉1𝐶𝑂𝑉1
Multicollinearity → Yes (Tolerance and VIF tables in appendix 3) Step 4: Base + Covariate 2
m4 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎 + 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏+ 𝛽6𝐶𝑂𝑉2 + 𝛽7𝑎𝐸𝑉2𝑎𝐶𝑂𝑉2+ 𝛽7𝑏𝐸𝑉2𝑏𝐶𝑂𝑉2
Multicollinearity → Yes (Tolerance and VIF tables in appendix 3) Step 5: Full model
m5 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎 + 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏+ 𝛽4𝐶𝑂𝑉1+ 𝛽5𝐸𝑉1𝐶𝑂𝑉1 + 𝛽6𝐶𝑂𝑉2 + 𝛽7𝑎𝐸𝑉2𝑎𝐶𝑂𝑉2+ 𝛽7𝑏𝐸𝑉2𝑏𝐶𝑂𝑉2+ 𝜀0
Multicollinearity → Yes (Tolerance and VIF tables in appendix 3)
Table 3 - Legenda to understand the formulas
27
4 – RESULTS
The most relevant results of this research are adressed in this section. ANOVA and Regression results are shown and, then, interpreted in the discussion of results part.
First of all, the means of the six different review conditions studied in this research are presented. With this basic information we got a first impression of the results.
Table 4 - Means
4.1 – Analyisis of Variance
An Analysis of Variance was employed (see results in table 5). The outcomes proved that both direct relationships of independent variables with moviegoers’ intention to see a movie in the cinema were significant (type of sender to DV → [ 𝐹(1,159)= 4.95, 𝑝 < 0.05] and valence to DV → [𝐹(2,159) = 104.68, 𝑝 < 0.001]).
In addition, both the positive and the negative professional reviews were significantly stronger (positive more positive and negative more negative) than those of the amateur reviews
[𝑀𝐴𝑚𝑎𝑡−𝑁𝑒𝑔 = 3.2 , 𝑀𝑃𝑟𝑜−𝑁𝑒𝑔 = 2.78 , 𝑀𝐴𝑚𝑎𝑡−𝑁𝑒𝑢𝑡𝑟 = 4.41 , 𝑀𝑃𝑟𝑜−𝑁𝑒𝑢𝑡𝑟 =
5.35 , 𝑀𝐴𝑚𝑎𝑡−𝑃𝑜𝑠= 6.65 , 𝑀𝑃𝑟𝑜−𝑃𝑜𝑠 = 7.72].
In Image 3 is observed how the interaction take place in the negative valence reviews [𝐹(2,159)= 3.74, 𝑝 < 0.05]
Table 5 – ANOVA
28 Image 3 – ANOVA graphic representation
(Can be found in bigger size in Appendix 2)
4.2 – Regression analysis
By performing a regression analysis on this survey data, we can determine whether or not these variables have impacted overall moviegoers intention to see a movie in the cinema, and if so, to what extent.
A summary table with the most important information of the regression is provided. More in deep results per model can be found in the Appendix 4.
29
Five regression models were performed with the following results: Step 1: Simple
m1 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏 Regression → 𝑅2𝑎𝑑𝑗 = 0.546 (𝐹
(1,165) = 66.66 , 𝑝 < 0.001)
The model is significant [p < 0.01] and all EVs have a significant effect on moviegoers’ intention to see a movie in the cinema. Type of sender when reviewer is professional [β=0.137 , 𝑝 < 0.05], positive valence review [β=0.482 , 𝑝 < 0.01] and negative valence review [β= -0.370 , 𝑝 < 0.01].
Step 2: Base (simple + interactions)
m2 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎+ 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏 Regression → 𝑅2𝑎𝑑𝑗 = 0.561 (𝐹
(1,165) = 42.86 , 𝑝 < 0.001)
The model is significant [p < 0.01] and almost all EVs have a significant effect on moviegoers’ intention to see a movie in the cinema. Interations come into play. Type of sender when reviewer is professional [β=0.210 , 𝑝 < 0.05], positive valence review [β=0.485 , 𝑝 < 0.01] and negative valence review [β= -0.238 , 𝑝 < 0.01]. Professional negative reviews are significant [β= -0.199 , 𝑝 < 0.05] while professional positive reviews are not significant on moviegoers’ intention. [p > 0.05].
Step 3: Base + Covariate 1
m3 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎 + 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏+ 𝛽4𝐶𝑂𝑉1+
𝛽5𝐸𝑉1𝐶𝑂𝑉1
Regression → 𝑅2𝑎𝑑𝑗 = 0.580 (𝐹
(1,165) = 33.39 , 𝑝 < 0.001)
The model is significant [p < 0.01] and almost all EVs have a significant effect on moviegoers’ intention to see a movie in the cinema. Involvement with movies is the novelty here. Type of sender when reviewer is professional [β=0.207 , 𝑝 < 0.05], positive valence review [β=0.490 , 𝑝 < 0.01] and negative valence review [β= -0.237 , 𝑝 < 0.01]. Professional negative reviews are significant [β= -0.200 , 𝑝 < 0.05] while professional positive reviews are again not significant on moviegoers’ intention. [p > 0.05]. Involvement with movies is significant on the direct variable [β=0.205 , 𝑝 < 0.01] and moderates the relationship with review type of sender significantly [β= -0.150 , 𝑝 < 0.05] but in a negative way.
Step 4: Base + Covariate 2
m4 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎 + 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏+ 𝛽6𝐶𝑂𝑉2 + 𝛽7𝑎𝐸𝑉2𝑎𝐶𝑂𝑉2+ 𝛽7𝑏𝐸𝑉2𝑏𝐶𝑂𝑉2
Regression → 𝑅2𝑎𝑑𝑗 = 0.618 (𝐹
(1,165) = 34.15 , 𝑝 < 0.001)
30
Step 5: Full model
m5 → 𝑌 = 𝛽0+ 𝛽1𝐸𝑉1+ 𝛽2𝑎𝐸𝑉2𝑎+ 𝛽2𝑏𝐸𝑉2𝑏+ 𝛽3𝑎𝐸𝑉1𝐸𝑉2𝑎 + 𝛽3𝑏𝐸𝑉1𝐸𝑉2𝑏+ 𝛽4𝐶𝑂𝑉1+
𝛽5𝐸𝑉1𝐶𝑂𝑉1 + 𝛽6𝐶𝑂𝑉2 + 𝛽7𝑎𝐸𝑉2𝑎𝐶𝑂𝑉2+ 𝛽7𝑏𝐸𝑉2𝑏𝐶𝑂𝑉2+ 𝜀0 Regression → 𝑅2𝑎𝑑𝑗 = 0.624 (𝐹
(1,165) = 28.21 , 𝑝 < 0.001)
The model is significant [p < 0.01] and almost all EVs have a significant effect on moviegoers’ intention to see a movie in the cinema. Both moderators are taking into account here. Type of sender when reviewer is professional [β=0.206 , 𝑝 < 0.05], positive valence review [β=0.530 , 𝑝 < 0.01] and negative valence review [β= -0.239 , 𝑝 < 0.01]. Professional negative reviews are significant [β= -0.196 , 𝑝 < 0.05] while professional positive reviews are again not significant on moviegoers’ intention [p > 0.05]. Involvement with movies is significant on the direct variable [β= 0.135 , 𝑝 < 0.05] and moderates the relationship with review type of sender significantly [β= -0.110 , 𝑝 < 0.1] but in a negative way. On the other hand, susceptibility is significant on the direct variable [β=0.170 , 𝑝 < 0.1] and moderates the relationship with negative reviews significantly [β= -0.123 , 𝑝 < 0.1] while the relationship with positive reviews is not significant [p > 0.05].
4.3 – Discussion of results
The descriptive statistics associated with moviegoers’ intention to see a movie in the cinema across the two types of review sender and the three valence groups are reported in Table 4. It can be seen that the professional reviewer group was associated with a numerically stronger mean in both positive and negative valence (more positive in positive valence and more negative in negative valence) than amateur reviewer group. Professional are more influential on moviegoers’ intentions.
𝑀𝐴𝑚𝑎𝑡𝑒𝑢𝑟&𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 = 3.20 𝑣𝑠 𝑀𝑃𝑟𝑜𝑓𝑒𝑠𝑠𝑖𝑜𝑛𝑎𝑙&𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 = 2.78
𝑀𝐴𝑚𝑎𝑡𝑒𝑢𝑟&𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒 = 6.65 𝑣𝑠 𝑀𝑃𝑟𝑜𝑓𝑒𝑠𝑠𝑖𝑜𝑛𝑎𝑙&𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒= 7.72
In order to test the hypotheses raised in this research, both a between-groups ANOVA (H1, H2a, H2b and H3) and a Regression Analysis were performed (H4, H5 and H6). Both of them yielded statistically significant effects. Results can be found in Tables 5 and 6..
Analyzing the results we can draw several interesting conclusions. First, the relationship between type of review sender and moviegoers’ intention to see a movie in the cinema
[ 𝐹(1,159) = 4.95, 𝑝 < 0.05] was significant. This fits with the regression output [β=0.206 ,
