1
Whose reviews are adopted by whom?
The Influence of online movie reviews explained by source – receiver interaction.
Master’s Thesis by Samira Baars
Amsterdam Business School, University of Amsterdam MSc. In Business Administration – Marketing Track Supervisor: dhr. dr. F. Situmeang
Student number: 5660521
2
Statement of originality
This document is written by Samira Baars, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Table of Contents Acknowledgements……….……..5 Abstract……….6 1. Introduction……….7 2. Literature review………..……….14
2.1. WOM versus eWOM.………..…………..14
2.2. Online reviews within the movie industry………..16
2.3. The power of reviewer expertise.………..…………..……18
2.4. Reviewer expertise within the movie industry………...….19
2.5. Differentiating between expert and consumer reviews……….…….….22
2.6. Comparing mainstream and arthouse movies……….….………..24
2.7. Mainstream and arthouse movie audience.………..26
2.8. Movie audience and perceived review source similarity………..28
2.9. Movie audience linked to ELM.………30
3. Method……….……….35
3.1. Design and procedure……….….….…35
3.2. Participants………..……….…37
3.3. Manipulation of experimental stimuli……….……….……….37
3.4. Manipulation checks………..……40
3.5. Independent variables……….……….……40
3.6. Dependent variables………..………40
3.7. Process variables……….……41
4 4. Results………..44 4.1. Descriptive statistics………..……44 4.2. Analysis strategy………...…………42 4.3. Hypotheses testing.………..….….47 4.4. Explorative analyses….………..……61 4.5. Additional analyses.……….…..62 5. Discussion………...64
5.1. Integration and interpretation of findings………..64
5.2. Theoretical contribution………..68
5.3. Practical implications.………69
5.4. Strengths and limitations.……….…….71
5.5. Future research.………...…..72
5.6. Conclusion.……….….………73
6. References………..…….….…….76
Appendix 1 – Means, standard deviations and intercorrelations among variables……..….…..91
Appendix 2 – Interaction effects of movie preference and review type……….…..95
Appendix 3 – Interaction effects between movie preference and source identity……….…….98
5
Acknowledgments
First, I would like to thank my supervisor Frederik for his optimism and support, quick feedback and for always being approachable and willing to help.
Then, I want to thank my family and friends for believing in me. In particular, I want to thank Rafael for his extra patience, love and care during these months, my mom and sister for persistently being dedicated to my graduation, my brother for once again repairing my laptop, and Elif and Bas for unexpectedly offering their help during data collection next to continuously being supportive.
I also wish to express my gratitude to all the people that were willing to participate in this study. By contributing each in their own way, all aforementioned people have been essential in bringing this thesis to completion.
6
Abstract
Given the fact that the movie industry is characterized by uncertainty and high risk, and that movie selection is strongly influenced by eWOM, it is crucial to understand the impact of eWOM on potential movie audience. As an attempt to explain the inconsistencies in online movie review literature, as well as answering the call for a focus on qualitative review aspects, this study investigated if, in what way, and why online movie reviews written by either experts or consumers differently influence movie audience. In doing so, this study distinguished between arthouse and mainstream movie audience. Based on their proposed different characteristics in terms of need for cognition and movie watching motives, it was expected that arthouse audience would be stronger influenced by expert reviews, whereas mainstream audience would be stronger influenced by consumer reviews. This interaction effect was proposed to be a result of different levels of perceived similarity and elaboration. Hence, theory of Similarity-Attraction (Byrne, 1991) and ELM (Peggy & Cacioppo, 1986) were applied in this context. To test the hypotheses, a quantitative experimental study was
performed in which manipulated reviews were presented as part of a questionnaire. Data of 188 moviegoers revealed that in general, arthouse audience was stronger influenced by expert reviews, while mainstream audience was stronger influenced by positive consumer reviews and negative expert reviews. The findings furthermore indicated that perceptions of similarity and levels of elaboration contributed to this interaction effect. This study has shown that the influence of online movie reviews is dependent the interaction between the review source and the receiver. Strengths, limitations and practical implications of the study are discussed.
7
1. Introduction
Each year, hundreds of movies are being produced worldwide and subsequently, watched by a particular audience. Between this huge amount of offer, how do people select a movie? Just as with other experience goods, characterized by attributes that are not known until using the product, this selection is strongly influenced by word of mouth (WOM; Bei, Chen & Widdows, 2004; Liu, 2006). For example, the box office success of movies as ‘The Blair Witch Project’ or ‘Star Wars’ has been attributed to the WOM caused by these movies (Liu, 2006). Here, it is worth noting that for every ten major movies produced, six to seven are
unprofitable (Vogel, 2001). Given the fact that the movie industry is characterized by uncertainty and a high level of risk (d’Astous & Colbert, 2002), every factor that contributes to the success of their movie is highly valuable to movie-producers and marketers.
Compared to other marketer-generated sources of information, as advertisement, WOM is often perceived as more credible and trustworthy and more easily accessible (Liu, 2006). WOM is furthermore unique because of the positive feedback mechanism it can generate (Duan, Gu & Whinston, 2008), meaning that WOM results in sales, which in turn generate more WOM that leads to even more sales (Godes & Mayzlin, 2004). In other words, WOM is not only an antecedent, but also an outcome of product sales (Duan, Gu & Whinston). Moreover, since television advertising as well as consumers’ trust of advertising is decreasing, WOM entails a way to obtain product success, and ultimately a significant competitive advantage (Sweeney, Soutar & Mazzarol, 2008).
The emergence of the Internet has subsumed traditional WOM in the form of advice from friends and relatives, by making it possible to exchange WOM in an online
environment with an unlimited amount of other people. In previous research, several types of the resulting electronic version of WOM (eWOM) are described, such as blogs (e.g.
8 Thorson & Rodgers, 2006), social media (e.g. Trusov, Bucklin & Pauwels, 2009), discussion forms (e.g. Cheung et al., 2009) and product reviews (e.g. Hu, Liu & Zhang, 2008). EWOM exerts a strong influence on decision making processes of consumers (Filieri & McLeay, 2014). In comparison with traditional WOM, eWOM possesses several advantages as being instantly, ubiquitous and permanently available, easily accessible and transparent
(Lindgreen et al., 2013). Next to this, it is often perceived as objective compared to commercial sources (Godes & Mayzlin, 2004). Consulting eWOM prior to purchasing is widespread; worldwide, 90% of consumers read online product reviews, and 83% report that they rely on them (Channel Advisor, 2001). This also applies to the movie industry, where online movie reviews are considered as highly influential (d’Astous & Colbert, 2002) and in addition, more prevalent than other types of WOM (Duan, Gu & Whinston, 2008). In other words, these online reviews can be of critical value for the success of a movie, which underlines the importance for movie producers and marketers to acquire a thorough understanding of how these reviews can influence their future audience.
Over the past decade, scholars have become increasingly interested in the
antecedents and consequences of eWOM (King, Racherla & Bush, 2014). This has resulted in a rich, yet fragmented body of literature because of different contexts as eWOM type, platform type or product category. Broadly, this literature can be divided into the
antecedents of sending versus searching eWOM, and its consequences to the senders versus
receivers; the power of eWOM.
Since the power of eWOM describes the effects of eWOM on consumers and hence, it’s effect on valuable outcomes as sales, revenues and stock prices (Chevalier & Mayzlin, 2006), it is no surprise that the last literature stream is also the largest (King, Racherla & Bush, 2014). With regard to this stream of literature, various numerical features of eWOM
9 have been extensively studied, for example the effects of review volume or review valence (e.g. Chevalier & Mayzlin; Chen, Wu & Yoon, 2004; Duan, Gu & Whinston, 2008; Duan, Gu & Whinston, 2008; Li & Hitt, 2008; Pavlou & Dimoka, 2006). However, scholars have called for a focus on review quality rather than quantity (Mudambi & Schuff, 2010; Resnick et al., 2000). That is, eWOM messages that are comparable in a numerical way do not necessarily have an equal effect on receivers. Marketing communications literature furthermore underlines that message persuasiveness not only depends on message characteristics, but also on sender and receiver characteristics (Pentina, Bailey & Zhang, 2018). This would explain why the influence of messages that are comparable in content can still differ per context. Since research in this area has mainly focused on message characteristics, this disaggregate effect of eWOM has hardly been covered in literature, and is therefore described as still being an open question (King, Racherla & Bush).
On different online platforms, an extensive amount of eWOM in the form of reviews can be found for every released movie. Here, reviews created by experts or consumers are often presented separated from each other. Expertise as source characteristic has been included in multiple studies (e.g. Basuroy, Chatterjee & Ravid, 2003; Chakravarty &
Mazumdar, 2010; Li & Zhen, 2012; McAuley & Leskovec, 2013; Petty, Cacioppo & Goldman, 1981; Sussman & Siegal, 2003; Vermeulen & Seegers, 2009). According to early studies, experts are more persuasive than non-experts, as why traditional marketing usually let their products or services be reviewed by experts (McGinnies & Ward, 1980; Petty, Cacioppo & Goldman, 1981; Wiener & Mowen, 1986). However, the superior influence of experts over non-experts might not be as straightforward as previously assumed, which is underlined by ambiguous findings regarding the influence of experts in more recent literature (Berger et al., 2010; Bickart & Schindler, 2011; Racherla & Friske, 2012; Tsao, 2014).
10 This ambiguity can be illustrated within the movie industry. Movies as ‘Fifty Shades of Grey’ and ‘Twilight’ sketch how movies can be hyped by consumers irrespective of expert review dispraise. On the other hand, some movies that receive excellent expert reviews seem to be only appealing to a relatively small group of consumers. A weak correlation between expert ratings and movie popularity is indicated in literature as well (Holbrook, 1999), just as a stronger influence of consumer reviews compared to expert reviews (Tsao, 2014) and contradicting findings (Kennedy, 2007; Reinstein & Snyder, 2005).
Before concluding that reviewer expertise within the movie industry is irrelevant, this study proposes a more nuanced approach by bringing forward differences between movie audiences in terms of their movie preferences as an explanatory factor. Segmenting movie audience by means of their preferences has been shown to be effective previously (Hixson, 2005). The current study will distinguish between arthouse and mainstream movie audience, and will investigate if and why this audience is differently influenced by online movie reviews. In doing so, this study aims to answer the following research question:
What is the influence of movie reviews on moviegoers (mainstream versus arthouse
audience), when review source expertise (consumer versus expert) is taken into account?
In this paper, it is argued that these audiences can form meaningful segments because of their different characteristics and motives that are not only related to their movie
preferences, but also to the way these audiences process a review. Subsequently, it is expected that movie audience’ preferences can affect the way this audience is influenced and persuaded by this review. It is proposed that mainstream audience is more strongly influenced by reviews stemming from consumers, while arthouse audience is more strongly
11 influenced by reviews stemming from experts. The hypotheses in this study are based on two complementary theories, namely Similarity-Attraction Theory (Byrne, 1971) and the
Elaboration Likelihood Model of persuasion (Petty & Cacioppo, 1986). The explanatory role
of both of these theories will be tested. The hypotheses will be further clarified in the theoretical framework of this paper, and illustrated in the conceptual model (Figure 1). To answer the research question, a quantitative study was performed by means of a questionnaire that contained manipulated reviews written by both consumers and experts. Data of 188 participants were analysed to be able to draw our conclusions.
By testing our propositions, deduced from premises that are based on grounded theories, this study contributes in several ways to the growing literature of eWOM in general and movie marketing in specific. The fragmented eWOM literature is lacking an encompassing approach that includes both source and receiver factors while investigating review influence. Such qualitative approach would furthermore address the
abovementioned literature gap regarding the identification of the disaggregate effects of eWOM on its receivers.
This study addresses this gap by simultaneously taking into account source, receiver and message characteristics, rather than studying them separately. In doing so, this study advances the literature stream regarding the power of eWOM in a nuanced way. By further narrowing down to one form of eWOM (online reviews) within one industry (movie
industry), this study can provide a more specific and less simplified framework that more accurately reflects reality. This way, this study aims at solving a piece of the fragmented and incomplete puzzle of eWOM literature. Simultaneously, it contributes to the clarification of the previously described inconsistencies regarding the relationship between reviewer expertise and eWOM influence.
12 This study furthermore adds to the growing literature on movie marketing by
providing insights in the effectiveness of segmenting movie audience based on movie preferences. Targeting movie audiences by means of their demographics and
psychographics is not feasible and stable over time (Hixson, 2005). On the contrary,
targeting movie audiences based on their movie preferences has been shown to be effective for other marketing methods as movie trailers (Hixson, 2005). Hence, it would be a valuable addition to assess if movie marketing in the form of online reviews is also affected by movie preference.
This study is unique in the way that – to the best of our knowledge – it will be the first that examines the influence effect of online reviews in an experimental setting by differentiating between movie audience preferences. This allows for a more in depth investigation of the ways in which reviews are differently processed by receivers and in doing so, for elaborating on previously suggested relationships that could only be assumed because of real data limitations (e.g. Gemser, van Oostrum & Leenders, 2007; Jeon & Jiao, 2012; King, 2007). Moreover, this method allows us to draw causal conclusions by
controlling for the influence of external confounding variables as for example offline WOM or advertising campaigns, in contrast to methods making use of real data. The observed review influence in studies that make use of real data may be spurious, since this influence could be induced by underlying unobservable quality signals (Reinstein & Snyde, 2005). The experimental setting simultaneously enables an investigation of the effect of reviewer expertise by comparing ordinary consumers and professional experts as review source. The clear separation between consumer and expert reviews on websites that present movie reviews, as for example Metacritic or IMDb, is barely reflected within the relevant stream of literature. That is, most studies focus on either expert or consumer
13 reviews, or ignore the distinction all together (e.g. d’Astous & Touil, 1999; Chen, Liu & Zhang, 2012; Duan & Whinston, 2008; Gemser, van Oostrum & Leeders, 2007; Jeon & Jiao 2012; Kennedy, 2007; Reinstein & Snyder, 2005).
From a practical perspective, movie producers and marketers can benefit from the in-depth insights this study aims to provide. The previously described risks attached to producing a movie, the crucial role of eWOM in the selection process of choosing a movie and the inherent positive feedback mechanism of eWOM, all underline the urgency to understand the impact of eWOM on the potential movie audience. Most of the studio’s and movie distributors use the internet as a critical marketing venue by establishing websites and discussion forums for their movies to reach their audiences (Duan, Gu & Whinston, 2008). Movie distributors are described to shape or even ‘manipulate’ critical opinion and to sometimes purposely include movie review quotes in their advertising (e.g. Gemser, van Oostrum & Leenders, 2007). By knowing which reviews will affect their movie audience, movie marketers can better manage and structure review platforms, strategically contribute to these platforms, and improve the selection of reviews for marketing purposes by
matching the review to the target audience.
The remainder of this paper is structured as follows. The next section presents prior relevant literature that composes the theoretical framework of this study followed by the resulting hypotheses, which are outlined in the conceptual model. Then, the method that was applied in this study is described as well as the sources of data. Subsequently, findings of data analysis are described in the result section. Finally, the conclusions, strengths and limitations of this study are discussed, followed by implications for theory and practice and ending with suggestions for future research in this area.
14
2. Literature review
2.1. WOM versus eWOM
The emergence of the Internet has resulted not only in numerous extra options for
organizations to market their products and services; it has also provided consumers with a powerful tool. By means of online reviews, every person with internet access is nowadays able to share experiences, opinions and information openly and uncontrolled with a broad public that consists of an infinite number of Internet users. For all product categories, including electronic goods, movies, books, or restaurants, consumers increasingly employ online platforms to post their evaluations (Zhang, Law & Li, 2010). This way, traditional word of mouth through personal face-to-face conversations has been transferred to online
contexts, resulting in a different, yet even more influential type of communication: eWOM. In literature, eWOM is defined as ‘any positive or negative statement made by potential, actual or former customers about a product or company, which is made available to a multitude of people and institutions via the internet’ (Hennig‐Thurau et al., 2004, p. 39). Just as with traditional WOM, eWOM allows consumers to exchange product information, to socially interact and to make informed purchase-decisions (Blazevic et al., 2013). However, its online-only and written nature implies that eWOM differs from traditional WOM in multiple ways, which are described in detail in the overview of King, Racherla and Bush (2014). Compared to WOM, the reach and volume of eWOM is
enhanced. Furthermore, eWOM can take place across a broad range of communities and is often enhanced by community engagement. In addition, it is persistently available,
anonymous and thereby it implies a risk for deception. Then, eWOM imposes the possibility to express an opinion in a numerical way, for example by stars (King, Racherla & Bush). Since the concept of eWOM is in essence distinctive from traditional WOM, it has resulted in an
15 entire new stream of literature.
Broadly, studies to eWOM can be divided into four literature streams. First, the motivations for consumers to talk online – or the antecedents of sending eWOM. For example, it is found that people can post their opinions online out of altruism (e.g.
Dellarocas & Narayan, 2007), because it results in feelings of self-enhancement (e.g. Angelis et al., 2011) and because of neuroticism (e.g. Picazo-vela et al, 2010). The second stream concerns studies to antecedents of receiving eWOM. Here it is found that consumers search for eWOM for social reassurance (Bailey, 2005), and to reduce uncertainty, search and evaluation efforts (e.g. Dabholkar, 2006), and risk (e.g. Sweeney, Soutar & Mazzarol, 2008). The literature on consequences to the senders on the other hand, describes that generating eWOM seems to provide social capital and reputation to the communicator (e.g. Chen et al., 2010), next to increased learning and impression management (Muniz & Schau, 2005). The last stream relates to the consequences to receivers, or ‘the power of eWOM’. This included more informed decision making, since consumers can better decide which products best meet their needs and preferences after consulting eWOM (e.g. Dellarocas, 2003). Moreover, eWOM has been found to enhance consumers’ willingness to pay (e.g. Pavlou & Dimoka, 2006), trust and loyalty (e.g. Awad & Ragowsky, 2008) and consumer engagement (e.g. Nambisan & Baron, 2007). As a result, the effects of eWOM translate into previously described valuable firm-level outcomes as increased product sales, revenues and stock prices (Chevalier & Mayzlin, 2006). For this reason, it is not peculiar that marketers and scholars have become increasingly interested in the power of eWOM, specifically in the form of online reviews for multiple product categories. Not for nothing, eWOM is referred to as the new element of the marketing communication mix (King, Racherl & Bush.
16 2.2. Online reviews within the movie industry
The movie industry has become a popular setting to study online reviews, and has by far received the most attention in marketing literature on (e)WOM (Liu, 2016; Duan, Gu & Whinston, 2008). The movie industry constitutes an excellent research context, because of the frequent introduction of new products and the high volume and availability of the reviews that appear as a result (Chen, Liu, Zhang, 2012). In general, the emerging studies are based on the premise that (e)WOM influences moviegoers (Lui, 2016). That is, negative movie reviews are believed to harm movie performance, whereas positive movie reviews are considered to enhance box office revenues (Gemser, van Oostrum & Leenders, 2007). Liu (2016) brings forward two possible explanations for this widespread belief. The first is the fact that movies are a prime category of culture goods. Because of this, they generally receive great public attention and interest, which in turn will lead to active
communication about movies. Secondly, just like other entertainment products as concerts, restaurants and books, movies fall in the category of experience goods. Concerning these markets, the product quality is uncertain prior to consumption, making consumers rely more heavily on secondary sources (Reinstein & Snyder, 2005; Jeon & Jiao, 2012). Hence, because of the intangible and experiential nature of movie consumption, consumers consult online reviews to be able to make a more informed decision before viewing a movie. For the movie industry, (e)WOM is therefore a critical factor for the ‘staying power’ of a movie on cinema screens, which in turn results in its financial success (Elberse & Eliashberg, 2003). This indirect effect of (e)WOM on demand when holding product quality constant is referred to as the influence effect of movie reviews and should be distinguished from the prediction
effect. That is, the prediction effect refers to the correlation between reviews and demand
17 movie will become a success (Eliashberg & Shugan, 1997).
Studies on online reviews in the context of the movie industry have been mainly focusing on the review volume which measures the total amount of eWOM interactions, review valence, often measured as positive, negative or as star ratings, and the variance in reviews (Godes & Mayzlin, 2004). The explanatory power of review volume on box office revenues is often confirmed (e.g. Duan et al., 2008; Liu, 2006), while the relationship between valence and sales seems to be mixed (Liu, 2006), and the relationship between variance and sales unsupported (Godes & Mayzlin, 2004). This is later explained by an indirect influence of review valence on box office sales through eWOM volume, measurable through a long term approach (Duan et al., 2008). Studies furthermore reveal that online reviews not only affect box office sales through consumers reading them, but also through vendors who allocate more screens to movies that generate more positive eWOM (Elberse & Eliashberg, 2003).
As aforementioned, in most studies eWOM is analysed in a numerical way, assuming that the impact on consumers is equal. This observation is described as a major void of eWOM literature in the review of King, Racherla and Bush (2014). That is, prior to adopting the eWOM messages, consumers not only look at the quantity of reviews; they also
carefully assess the quality of the content and the source characteristics of the review. As a consequence, some online reviews are more influential than others. To consider these subtleties of online interactions (Resnick, 2000), scholars have recently called for defining the socio-psychological aspects by analysing factors as source characteristics and source-consumer similarity (King, Racherla & Bush, 2014). In the next section, this call will be addressed by describing the current state of research that focuses on reviewer expertise as source characteristic.
18 2.3. The power of reviewer expertise
Unlike traditional WOM, eWOM does not enable consumers to directly examine the trustworthiness of the review source because of its anonymous nature (King, Racherla & Bush, 2014; Huang, Hsiao & Chen, 2012). Hence, to assess if a review is worth relying on, consumers have to look for alternative cues to evaluate the source. A study investigating the factors resulting in consumers’ value perceptions and subsequent adoption of eWOM shows that expertise in the relevant domain leads to positive evaluations, and that this relationship is furthermore dependent on consumer characteristics (Weiss, Lurie & MacInnis, 2008). King (2007) mentions three factors that can contribute to the market power of experts’ (movie) reviews in general. First, these reviews are widely available, not only online but also in newspapers and magazines. Secondly, experts can be considered as relatively objective and therefore credible. Lastly, experts can function as opinion leaders for their readers. This applies specifically to the movie industry, where expert reviews have a strong advisory function since they can influence future purchase decision making (in contrast to reviews about passed events that are solely read out of intrinsic interest).
Indeed, the real world provides examples where the expertise of the reviewer is crucial for affecting consumer purchases. For instance, restaurant owners ache for
prestigious Michelin stars awarded in the rating system of restaurant experts, mainly since these stars are widely perceived as indicators of quality gastronomy. The influence of this rating system is described as ‘both impressive and fearsome’, since gaining or losing a star will directly translate into an enormous change in the sales and profitability of the
restaurant (Johnson et al., 2005, p. 172). The fact that the Michelin Guide literally guides restaurants’ target market in selecting a restaurant emphasizes the crucial influence of reviewer expertise on consumer decision making processes in this sector.
19 On the other hand however, there are examples of products that are extensively criticized by experts, yet extremely popular under mass consumers – and the same applies for the other way around. For a long time, car brand Skoda suffered from a low image among consumers despite positive expert reviews on its quality. Clearly, this concerns a product category in which brand relevance is high (Fischer, Völckner & Sattler, 2010) and where awareness and associations could surpass the influence of reviewer expertise. Yet, comparable examples can be found within the music, book or movie industry, in other words experience good markets where brand influence is less relevant.
This interesting ambiguity in real world observations seems to be reflected in eWOM literature, which provides inconsistent findings regarding the influence effect of expert reviews, or their superior influence over non-expert reviews. Berger et al. (2010) for example, found that negative book reviews resulted in increased sales. Some studies conclude that consumers ascribe more value to online reviews written by other consumers rather than professional reviewers because of higher perceived credibility and relevance (e.g. Bickart & Schindler, 2001), other studies do find positive correlations between source expertise and consumers’ behaviours (e.g. Gilly et al., 1998). These mixed findings indicate that the influence of reviewer expertise might be relevant to some contexts, outcomes or receivers, but irrelevant for others.
2.4. Reviewer expertise within the movie industry
Also within product categories, the influential effect of reviewer expertise seems to be ambiguous. Returning to the movie industry, Jeon & Jiao (2012) carried out a study on opening weekend revenues of movies released between January 2008 and August 2011. Movies that generated the highest revenues, like ‘A Christmas Carol’ and ‘The Twilight Saga:
20 New Moon’ were previously poorly rated in expert reviews (a score of 55 and 44 out of 100 respectively), counterintuitively suggesting that negative publicity stemming from expert reviews could result in positive influence effects. The authors explain their observation by the ‘any publicity is good publicity’ phenomenon, referring to the positive effect of product awareness.
In the study of King (2007) performed by means of Metacritic data, an independent effect of critical ratings on box office earnings is found. That is, every extra 20 points on a rating is associated with $26 million in box office earnings, but only for movies released on more than 1000 screens. According to d’Astous and Touil (1999), consumers carefully evaluate the motives of the expert reviews before they are influenced by the expressed judgement and recommendations. In the experimental study of Tsao (2014), results indicate that consumer reviews are more important for potential moviegoers’ movie selection than expert reviews – especially when the review was negative or when the moviegoer had lower expectations towards the movie. The previous findings are further nuanced in the study of Kennedy (2007). By means of retrieving Metacritic score data1 from 220 movies and data about their revenues, this study shows that positive expert reviews do positively affect sales but only when it concerns a widely-released movie. For ‘limited edition movies’, no
relationship is found. The author suggests that people will consult reviews for further information after they have noticed the movie through marketing or advertisements, in which limited edition distributors do not invest as much.
The results of this study however are contradicted by the study of Reinstein and
1 Metacritic score data is data retrieved from the website www.metacritic.com, that continuously monitors approximately 40 other sources of movie reviews and is therefore often consulted to retrieve big data regarding online movie reviews. These sources include daily and weekly newspapers and magazines, trade publications, and the internet. By means of special formula, the website assigns a score out of 100 to each review, allowing for comparisons between data regarding their valence (King, 2007).
21 Snyder (2005), who find the opposite result. Here it is found that positive expert reviews influence opening weekend box office revenue, but only for narrowly-released movies instead of widely-released movies. In contrast to the argumentation expressed by Kennedy (2007), the authors in this study argue why expert reviews are more important for arthouse movies than for mainstream movies: for commercial widely-released movies, press reports and advertising already provide sufficient quality signals. In addition, consumers of these movies have a different judgement of quality than critics.
The differentiation between arthouse and mainstream movies is also applied in the study of Gemser, van Oostrum and Leeders (2007), which focuses on the influence of
number and size of expert reviews. They find that for arthouse movies, expert reviews show an influence effect on box office sales. For mainstream movies on the other hand, a
predictor effect is seen and explained by the hypotheses that mainstream movie audience relies on alternative sources of information. The factor movie audience is also included in the study of Jeon and Jiao (2012), who report that while positive expert reviews of domestic movies discourage people to watch a movie, a positive influence is found for foreign movies. The authors explain their findings by the differences in domestic versus foreign movies’ target audience. They argue that because foreign movies in the study were perceived as art movies, audiences who would like to watch sophisticated movies are positively influenced by reviews written by experts, who are expected to have similar taste.
In short, there is no general consensus on the influence of expert reviews within the movie industry. Some studies assume that the ambiguous effects can be attributed to the movie audience that reads the review. In this study, this assumption will be investigated in more depth by assessing if, how and why arthouse and mainstream movie audience differently responds to reviews by differentiating between expert and consumer reviews.
22 2.5. Differentiating between expert and consumer reviews
Movie review experts can be defined as “persons, usually employed by newspapers, TV stations or other media who screen newly released movies and provide their subjective views and comments on the movie for the public’s information” (Cones, 1992, p. 120 as adopted by Eliashberg & Shugan, 1997). They are neither producers or consumers, and are known to function as opinion leaders by influencing consumers to watch a movie or not by virtue of specific knowledge attributed to them (Wijnbeg & Gemser, 2000).
Product reviews however, are not only submitted by experts, but also by ordinary consumers. These reviewers merely express their evaluations based on their experience, and in doing so represent mass taste and popular appeal (Clement, Proppe & Rott, 2007). Despite a clear distinction between the source of the review on most review websites, only few studies have translated this into comparing the influence of consumer and expert reviews (Na, Thura Thet & Khoo, 2010; Tsao, 2014).
Reviews written by consumers or experts are characterized by different aspects. In a study on sentiment expressions, their typical differences are extensively analysed and outlined (Na, Thura-Thet & Khoo, 2010). First of all, expert reviews tend to be more lengthy and comprehensive. While ordinary movie consumers may discuss only the aspects they like or dislike the most, movie experts try to provide a comprehensive analysis of a movie. Whereas consumers often discuss multiple movie aspects in a single sentence, expert reviews bring up the various aspects of the movie by discussing these aspects one at a time. Then, experts use more formal and complete sentences and unique terms. This in contrast to consumers, who generally express themselves in more informal and incomplete sentences and by using common words. In their reviews, experts mention the name of the movie director, the cast and story aspects much more frequently than consumers. Also, the
23 directors techniques, movie choreography, character and storyline development and acting skills are discussed in more detail by experts.
Furthermore, the article of Na, Thura-Thet and Khoo (2010) shows that consumers express 81 percent more often how they feel towards a movie. In doing so, they use informal words (e.g. ‘cool’, ‘funny’, ‘stupid’, ‘boring’). Acting skills are discussed in a much more simplified way in consumer reviews compared to expert reviews (e.g. ‘Robert Downey Jr is awesome’). In addition, it is found that scene aspects (e.g. setting, action and visual effects) are 41 percent more often discussed by consumers. Differences in word choice are also observed: experts tend to use milder terms and measured expressions (e.g. ‘weak plot’, ‘limp script’), and consumers more strong words (e.g. ‘awesome’, ‘fantastic’, ‘dumb’ and ‘horrible’) and informal expressions (e.g. ‘the script sucked big time’). Chakravarty, Liu & Mazumdar (2010) add that consumer reviews are typically affective and personally relevant, while expert reviews more often include technical and artistic movie aspects.
To summarize, expert movie reviews can be described as a) relatively longer, b) more comprehensive, c) more structured, d) more formal, e) more detailed (in terms of artistic and technical movie aspects, directors techniques, movie choreography, acting skills, and character and storyline development), and f) containing more complex vocabulary, g) more complete and longer sentences, h) more unique terms, i) milder and more measured expressions and j) less affective and personally relevant – when compared to consumer movie reviews.
Based on the different message characteristics of expert and consumer reviews as observed in an in-depth comparison (Na, Thura Thet & Khoo, 2010), it is expected that these reviews will be differently processed by the readers of these reviews and consequently, that they will result in a different influence effects.
24 2.6. Comparing mainstream and arthouse movies
Within the online review literature, it is recurrently suggested that the influence of movie reviews differs per movie category because of the consumers these movies attract (Gemser, van Oostrum and Leeders, 2007; Jeon & Jihao, 2012; Kenedy, 2007; Reinstein and Snyder, 2005). In different terminology, earlier studies have distinguished between mainstream and arthouse movies. These propositions however have not been tested before, and will be elaborated on in this study. Both mainstream and arthouse movies are characterized by typical features. To understand the difference, it is important to know what is meant by ‘arthouse’, and how this concept relates to ‘mainstream’.
The category arthouse emerged after the Second World war, when there was more prosperity and attention for creating and experiencing art and entertainment (Willinsky, 2001). Since then, small theatres in urban areas were screening these ‘offbeat’ movies as independent Hollywood, foreign language, and documentary movies. Typically, these
theatres created ‘a subtle snob appeal’, by exhibiting tasteful art in the lobby, serving coffee and asking premium prices. It attracted a well-mannered, sophisticated audience that was looking for an experience different than offered in mainstream theatres (Willinsky).
As with the concept of ‘art’, there is no general applied definition of what constitutes ‘arthouse movies’. Both concepts typically refer to their artistic quality, whereas
mainstream movies are characterized by their commercial qualities (Baumann, 2002). One could state that over the years, the term arthouse has referred to “feature movies made in the post-World War II period which display new ideas of form and content and which are aimed at high-culture audience” (Lev, 1993, p. 39). Arthouse movies differ from classical Hollywood movies in filming style and theme, which reflects their artistic, realistic and expressive image. Other key features of arthouse movies are “an emphasis on visual ad
25 character style, a suppression of action, and an interiorization of dramatic conflicts” (Neale, 1981, p. 103). Hence, arthouse movies form an alternative for movie audiences to
distinguish themselves from ‘ordinary’ movie audiences by watching content of higher artistic, esthetical and intellectual level (Willinsky, 2001). To avoid definition ambiguity, some scholars simply refer to arthouse movies as ‘movies shown in art theatres’, and to mainstream movies as ‘movies shown in mainstream theatres’ (Gemser, van Oostrum & Leenders, 2006; Zuckerman & Kim, 2003).
Arthouse movies can be distinguished from mainstream movies based on several characteristics, under which their content, plot, actor type, visual richness, budget,
production purpose, target group and share of screens (Gemser, van Oostrum & Leenders, 2006). The plot of a mainstream movie tends to be organized in a narrative way as a series of events causally following up each other and performed by well-defined characters (Bordwell, 1979). To appeal to a wide audience, mainstream movies stay simple in terms of images, plots, and characters, making them easy to enjoy and not challenging or difficult to understand (Baumann, 2001). This in contrast to the more complex content in the non-causal narrative structures of arthouse movies, which requires effort to be appreciated and which invites their audience to experience new ways of thinking (Canaday, 1980). To continue, mainstream movies are high-budget products. Their high production and
marketing budgets are mainly spent on cues that could attract the broad public. These cues include popular movie stars, catchy soundtracks and creating visual richness with elaborate special effects that arthouse movies often lack. Unlike arthouse movies, the role of the plot tends to be almost irrelevant (Bachella & Bechetti, 1999).
Consequently, mainstream and arthouse movies differ by the audience they attract. By means of low budgets, arthouse movies as a form of ‘art’ are mainly produced to give
26 expression to the idea of the director, while focusing on a niche market that wants to see more mature, intelligent movies (Wilinsky, 2001). On the contrary, mainstream Hollywood productions are ‘business’, aiming at pleasing the largest possible audience by means of their commercial qualities (Baumann, 2002). As a result, mainstream movies are shown on more screens and in more theatres than arthouse movies (Reinstein & Snyder, 2005). This commercial success could be considered the main driver of mainstream movie production. Since its origin, arthouse movie producers have balanced between differentiation and making profit (Wilinsky, 2001). That is, by decreasing the gap with mainstream movies, they might attract broader audience and hence, make more profit. But simultaneously this strategy would diminish the differential factors that characterize them. Generally spoken, arthouse movie producers pursue value in the form of cultural capital.
In short, arthouse movies can be labelled as ‘independent’ and ‘low budget’, and mainstream movies as ‘major’ and ‘high budget’ (Zuckerman & Kim, 2003).
2.7. Mainstream and arthouse movie audience
Because arthouse movies are targeting a specific niche market, researchers have been trying to sketch a profile of this type of movie audience in light of what differentiates them from mainstream movie audience. Though the article is dated, a serious attempt to do so is made by Austin (1983) by gathering in depth survey information from 226 art movie visitors. After analysing the data, the author describes arthouse audience by summarizing the outcomes, namely that: “1) they are highly educated, enthusiastic movie-goers; 2) they are more likely than other film audience samples to report movie-going as their favourite leisure activity; 3) they are frequent movie-goers; 4) they perceive their art film theatre as offering a unique alternative to commercial cinemas; 5) a majority planned their last attendance at the art
27 theatre at least one week in advance; 6) 30% attended alone the last time they went; 7) they are interested in learning about the films they see; and 8) unlike the majority of U.S. movie-goers, they press only a modest preference of American films over foreign movies” (Austin, 1983, p. 1.).
While these findings provide a detailed impression of arthouse audience, they do not provide a comprehensive understanding of what differentiates this audience from
mainstream moviegoers. In this paper, it is argued that the type of movie that consumers select to watch can be perceived as an indicator of several latent characteristics related to consumers’ decision making processes. Based on these different characteristics, arthouse and mainstream movie audiences could be expected to be differently influenced by expert and consumer reviews. One could think of the motives to watch a movie, or the general motivation and ability to learn and process new information. For example, when someone selects a mainstream movie instead of an arthouse movie, this could implicitly reveal that this person does not like to engage in complex thinking in general, or is just watching movies as light and enjoyable entertainment. Because of this, this person might be not very
motivated to thoroughly process detailed and complex information sources. As a
consequence, this moviegoer might be less prone to reviews written by experts given the typical beforementioned review aspects as being more complex. In addition, the moviegoer might feel that the expert review does not reflect his own entertainment values anyway, and might instead be persuaded by a review written by a fellow consumer. Following this logic, it is plausible to expect that movie preference not only relates to if consumers consult online reviews, but also to how these reviews are processed and subsequently affect their intentions to watch the particular movie.
28 theory of similarity attraction and the Elaboration Likelihood Model of persuasion (ELM). Deduced from these theories, this chapter will end with the proposition that mainstream audience will be more affected by consumer reviews than expert reviews, while arthouse audience will be more prone to the reviews written by experts compared to reviews written by consumers. The previous sections that highlight the differences between arthouse and mainstream movies and their audience, as well as the differences between consumer and expert reviews, form the starting point of these propositions.
2.8. Movie audience and perceived review source similarity
Not only review content influences the way in which reviews are processed, but also the review source and how this is perceived by the audience (Pentina, Bailey & Zhang, 2018). A synthesis of the literature makes it apparent that arthouse audience is more comparable to experts as review source, while mainstream audience is more comparable to consumers as review source. To recap, consumer reviews are suggested to represent mass taste and popular appeal of most mainstream audience (Clement, Proppe & Rott, 2007), while the nuanced writing style and detailed evaluation of expert reviews should reflect the
sophisticated taste of arthouse audience (Jeon & Jiao, 2012). To explain how both review sources can affect mainstream and arthouse audience differently, theory of
Similarity-Attraction is introduced here (Byrne, 1967). According to this theory, attraction between
individuals increases when the other person is perceived as more similar to themselves. For example, people are attracted to other people when they are similar in terms of
demographics, activities, experiences, attitudes or personality characteristics (Byrne, Griffitt & Stefaniak, 1967; Byrne, 1971; Tsui & O’reilly, 1989). The opposite also applies, in that perceived dissimilarity can decrease interpersonal attraction (Rosenbaum, 1986).
29 Studies further support the suggested similarity regarding movie reviewer and movie audience. For example, it is found that arthouse movie audience applies similar evaluative criteria as professional critics (Hirschman & Pieros, 1985), assuming that they are highly similar with respect to information processing and movie evaluation. In addition, the same study describes that mainstream movie audience handles different evaluative criteria - that might even conflict with those of movie experts. In line with this, it is found that experts give higher ratings to movies that are more complex, abstract and intellectually demanding in nature (Holbrook, 1999), in other words movies that reflect the characteristics of arthouse movies.
Therefore, it is reasonable to assume that similarity-attraction theory applies in this context. Additionally, prior studies have indicated that perceived similarity between
information seeker and source enhances information persuasiveness (e.g. Bansal & Voyer, 2000; Gilly et al., 1988). This applies especially for similarity in internal traits as opinions, taste and views on topics, since this indicates membership of the same social group, which further enhances persuasion (Lichtenthal & Tellefsen, 2001). Because of abovementioned, it is expected that:
Hypothesis 1: (a) Mainstream audience will report higher review source similarity for consumer movie reviews compared to expert movie reviews, while (b) Arthouse audience will report higher review source similarity for expert movie reviews compared to consumer movie reviews.
Perceived review source similarity then, is expected to be positively related to elaboration (Pentina & Taylor, 2010; Wheeler, Petty & Bizer, 2005). This relationship as well as the
30 concept of elaboration will be further explained in the next section.
Hypothesis 2: Perceived review source similarity is positively related to elaboration.
2.9. Movie audience linked to ELM
Movie entertainment is consumed for a variety of reasons. Oliver and Raney (2011) reason that people not only watch movies for ‘enjoyment’, but that entertainment can also be ‘gratifying’, for example moving or meaningful entertainment. To distinguish these motives, the authors refer to entertainment consumption in the pursuit of pleasure and amusement with ‘hedonic motives’, and as a need to search for life’s meaning, truths and purposes with ‘eudaimonic motives’. While hedonic motives refer to watching a movie for pleasure and positive valence, eudaimonic concerns may capture motives of seeking insights, personal development and expressiveness and contemplation of life.
In light of the typical beforementioned characteristics of arthouse and mainstream movies, arthouse movie audience may watch movies for eudaimonic motives, while
mainstream movie audience reflect hedonic motives. This idea finds support in the study of Oliver and Raney (2011), that shows hedonic motives matches fun and diversionary
entertainment, whereas eudaimonic motives were consistent with more meaningful, serious and thought-provoking entertainment. Following this logic, it is hypothesized that:
Hypothesis 3: a) Mainstream audience will report more hedonic than eudaimonic motives for watching a movie, while b) arthouse audience will report more eudaimonic than hedonic motives for watching a movie.
31 Because eudaimonic motives should reflect a motivation and tendency to contemplate, they are linked to the concept need for cognition (Oliver and Raney, 2011). A scale for the
concept need for cognition as a personality trait was developed by Cacioppo and Peggy (1982) and has become a popular term within literature on message persuasion. Need for cognition reflects the differences between people in terms of their “tendency to engage in and enjoy thinking” (Cacioppo & Petty, 1982, p. 116). People high in need for cognition are characterized as ‘thinkers’, with a ‘need to understand’.
Hence, in line with previous reasoning regarding the characteristics of mainstream and arthouse movies, it could be hypothesized that:
Hypothesis 4: Arthouse audience will report higher need for cognition compared to mainstream audience.
Need for cognition is associated with the amount of willingness to invest mental effort. While people high in need for cognition might highly elaborate when they process information, people low in need for cognition may process information in a heuristic way through low elaboration (Dole & Sinatra, 1998; Hall & Zwarun, 2012). Here, the Elaboration Likelihood Model (ELM) comes into play (Petty & Cacioppo, 1986). According to this widely adopted model, people choose either central or peripheral routes for interpreting and processing information before cultivating attitudes. The chosen route is dependent on two factors, namely the motivation and ability to elaborate on information. When people are more motivated and able to think about information that is relevant to the issue, they are more likely to elaborate. This way, information is conducted through the central route, implicating that these people will pay close attention to and think more about the message
32 content. However, when ability or motivation is lacking, people process the information through peripheral routes, which means they will elaborate less on the information and focus more on peripheral cues that can be automatically processed, like simple statements, number of reviews or ratings.
Hence, since need for cognition and eudaimonic movie watching motives can both be linked with willingness and motivation to elaborate on message content, the following is expected regarding the relationship between need for cognition as well as movie watching motives:
Hypothesis 5: Need for cognition is positively related to elaboration.
Hypothesis 6: a) Need for cognition is positively related to eudaimonic movie watching motives, and b) negatively with hedonic movie watching motives.
Hypothesis 7: a) Eudaimonic movie watching motives are positively related to elaboration, while b) Hedonic movie watching motives are negatively related to elaboration.
Arthouse audience as target group is thus hypothesized to be interested in intellectually challenging entertainment, to possess high levels of need for cognition and eudaimonic movie watching motives. That is, arthouse audience is characterized by high willingness and ability to process complex stimuli. For example, Austin (1984) reports that arthouse audience consciously select their movies, and show interest and enjoyment towards reading about these movies. As suggested by Gemser, van Oostrum and Leenders (2007), this conscious way of selecting and reading could translate into increased interest in
33 reviews. This would apply especially for more complex and cognitively demanding expert reviews, further underlining their motivations and ability to be intellectually challenged. Hence, one could assume that arthouse audience will be especially motivated to process expert reviews in more depth through the central route.
Mainstream audience on the other hand could be assumed to be less attracted to thoroughly process the complex, detailed and formal nature of expert reviews compared to the more accessible and simple reviews of fellow consumers. That is, mainstream audience is hypothesized to possess lower levels of need for cognition and mostly hedonic movie watching motives, resulting in less motivation and ability to elaborate compared to arthouse audience. Therefore, mainstream movie audience could be expected to be less elaborating through the central route on reviews stemming from experts, since these reviews inherently require more processing effort. Mainstream audience is expected to be able and motivated to process the more simple, informal and less detailed consumer reviews in more depth compared to expert reviews. In addition, studies show that review elaboration is enhanced by perceived source similarity (Pentina & Taylor, 2010; Wheeler, Petty & Bizer, 2005). Hence, given the characteristics of expert versus consumer reviews, and the hypothesized differences between arthouse and mainstream audience in terms of their need for cognition, movie watching motives and perceived similarity with the review source, it is expected that:
Hypothesis 8: a) Mainstream audience will elaborate more on consumer reviews compared to expert movie reviews, and b) Arthouse audience will elaborate more on both expert and consumer reviews compared to mainstream audience.
34 To summarize, it is hypothesized that mainstream audience will feel more similar to and elaborate more on consumer reviews, while arthouse audience will feel more similar to expert reviews and will generally elaborate more. ELM is the most tested framework of persuasion within marketing science (Kitchen et al., 2014). Both perceived similarity and elaboration are extensively associated with affecting message influence in literature (e.g. Bansal & Voyer, 2000; Gilly et al., 1988; Feick & Higie, 1992; Lichtenthal & Tellefsen, 2001; Petty & Caccioppo, 1984). The indicators for review influence in this study are the attitude towards the review (encompassing credibility, helpfulness, trustworthiness and
persuasiveness), movie expectations and behavioural intentions to watch the movie, reflecting a majority of literature in this area (e.g. Chakravarty, Liu & Mazumdar, 2010; Lee, Park & Han, 2008; Mudambi & Schuff, 2010; Pentina, Bailey & Zhang, 2018). This brings us to the final overarching hypothesis of this paper:
Hypothesis 9: a) Mainstream audience is more strongly influenced by consumer movie reviews compared to expert movie reviews – resulting in higher attitude towards both positive and negative consumer reviews, and in higher movie expectations and behavioral intentions in the same direction as the valence of the consumer review, and b) Arthouse audience is more strongly influenced by expert movie reviews compared to consumer movie reviews – resulting in higher attitude towards both positive and negative expert reviews, and in higher movie expectations and behavioral intentions in the same direction as the valence of the expert review.
35
Figure 1. Conceptual model.2
3. Method
3.1 Design and procedure
A 2 (implicit versus explicit source identity: between-subjects) x 2 (consumer versus expert review source: within-subjects) x 2 (positive versus negative review valence: within subjects) mixed factorial design was used to test the proposed model. To gather data that would closely match the population of interest (mainstream and arthouse movie audience),
36 approximately 600 moviegoers were approached at both arthouse theatres and commercial cinema’s. The approached moviegoers were invited to participate by receiving a small invitation card containing a link to the online Qualtrics survey. To stimulate participation, eight participants could win a cinema ticket in a raffle at the end of data collection, which was also mentioned on the card.
All parts of the experiment were conducted online. The survey started with an informed consent in which anonymity and confidentiality was ensured and voluntarily participation was emphasized. The informed consent furthermore included the procedure of the study as well as an unspecific description of the aim: it was globally mentioned that the study would focus on movie reviews and characteristics of the review reader.
The survey started with demographic questions and the assessment of involvement and prior experience with online reviews and movies, and continued with measuring movie watching motives. Then, participants were randomly assigned to either the ‘explicit
condition’ that presented reviews by including the source identity (consumer or expert), or the condition that presented reviews without mentioning the source identity (‘implicit condition’). Participants were asked to read four reviews, which were presented in a
random sequence to each participant. Each of the reviews was followed by items serving as manipulation check. Then, elaboration, perceived similarity and the dependent variables (attitude towards the review, movie expectations and behavioural intentions) were measured for each review. At the end, need for cognition was measured. In the last step, participants were asked to provide their e-mail addresses so they could be contacted in case they would win the raffle. There was one moment of measurement for each participant. The survey including the presented review types can be found in Appendix 2.
37 3.2. Participants
As a result of the recruitment process, 207 participants filled out the survey, indicating a response rate of 34,5%. Participants who did not complete the survey (n = 15) or who completed the survey within a very short (z ≥ -3) amount of time (n = 4) were deleted. This resulted in a dataset consisting of 188 participants.
The average age of participants (N = 188; 50.5% male, 49.5% female) was 32.8 years (SD = 16.12), ranging from 13 to 86 years. A percentage of 56.4% of the participants was higher educated (college or university degree). A minority of the participants (n = 76; 40.4%) reported to prefer arthouse movies over mainstream movies, and the remaining
participants (n = 122, 59.6%) reported to prefer mainstream movies over arthouse movies. In line with this, a minority of participants reported that the last cinema they visited showed mostly arthouse movies (n = 63; 33.5%), compared to commercial movies (n = 125; 66.5%). The minority of participants (n = 58, 30.9%) visited a cinema or movie theatre ten times a year or more, or three times or more during last month (n = 43, 22.9%). A percentage of 42.8% of the participants (n = 88) reported to read online reviews regularly before watching a movie, and 43.6% participants (n = 82) reported to be regularly influenced by online movie reviews with respect to selecting a movie.
3.3 Manipulation of experimental stimuli
The experimental manipulation consisted of online movie reviews written by either
consumers or experts. Two conditions were constructed to measure the influence of review source: one condition presenting reviews without mentioning the source identity (‘implicit condition’) and one presenting reviews that did include the source identity (‘explicit condition’). To enhance authenticity, the reviews were retrieved from the Internet Movie
38 Database (IMDb.com), which is an online database containing information about movies and video games, including reviews and ratings given by both consumers and critics.
As described before, expert reviews in general are “a) relatively longer, b) more comprehensive, c) more structured, d) more formal, e) more detailed (in terms of artistic and technical movie aspects, directors techniques, movie choreography, acting skills, and character and storyline development), and f) containing more complex vocabulary, g) more complete and longer sentences, h) more unique terms, i) milder and more measured expressions and j) less affective and personally relevant – compared to consumer movie reviews”. Expert and consumer reviews that in a high degree reflected the characteristics of this distinction were selected for this experiment, with the exception of review length. Since previous research has frequently shown that the length of the review influences it’s effect (e.g. Hamamura, et al., Mudambi & Scruff, 2010; 2009; Pan & Zhang, 2011), each selected review was of approximately the same length (between 287 and 298 words).
The outcomes of interest in this study are expected to be influenced in the same direction as the valence of the consumer review. That is, no clear conclusions could be deduced from outcomes of reviews without a valence. Comparing the influence of neutral or mixed consumer and expert reviews would therefore be troublesome. Hence, to be able to measure review influence as well as for reasons of comprehensiveness and to control for potential review valence effects (e.g. Chakravarty, Liu, & Mazumdar, 2010; Lee, Park & Han, 2008) it is chosen to include reviews of both positive and negative valence.
To prevent everything other than review source from influencing the outcomes, other factors were held as constant as possible. By including both positive and negative reviews per review source, it was ensured that review tone would not interfere with the
39 study. The selected movies were ‘Demain Tout Commence’ (2016)3 and ‘Papadopoulos and Sons’ (2013)4. Both movies fall within the genre of comedy drama and in-between the arthouse and mainstream movie category by containing both arthouse and mainstream characteristics (instead of reflecting particularly one category). Details that could reveal the movie (e.g. movie title or actor names) were either not mentioned or fictitiously created, so participants would not recognize the reviewed movie. This way, a direct influence of movie related factors (e.g. familiarity with the movie, movie category characteristics, movie genre, likeability and attractiveness of the actors) on consumer responses to the review was prevented.
As a result, four movie reviews were selected reviews: 1) a positive consumer review, 2) a negative consumer review, 3) a positive expert review and 4) a negative expert review. Both implicit and explicit condition contained the same reviews; solely the inclusion of the source identity was added in the explicit condition. The (nick)names of the reviewers were replaced by fictional ones to ensure that gender or country of origin could not be deduced. Furthermore, reviews were presented in a random sequence to each participant in order to prevent sequence effects (e.g. learning effect) from interfering in the study.
3 ‘Demain Tout Commence’ (2016) is a French independent movie from origin and covers both the genre of comedy and drama. However, it could be argued that the movie is characterized by mainstream movie features as well: (1) the plotline was a direct remake of the very successful Mexican mainstream movie ‘Instructions not included’ (2013), (2) the movie involves soundtracks developed by the famous singer Barbra Streisand, (3) the main character (Omar Sy) is widely known after his previous role in the successful arthouse movie ‘Intouchables’, (4) following ‘Intouchables’, ‘Demain Tout Commence’ gained success immediately after its release by appealing to the mass taste which resulted in a box office sale of more than 32 million dollars and (5) the producers spend a high budget of 17.5 million dollars to produce the movie.
4 ‘Papadopoulos and sons’ (2013) is an independent British film directed and self-distributed by its writer Marcous Markou. Just as ‘Demain Tout Commence’ (2016), the movie is characterized by both arthouse and mainstream movie features.
40 3.4 Manipulation checks
Review source. To verify that the presented reviews were perceived as such, the following item served as manipulation check: “I think this review was written by”, followed by answer options ‘a) a consumer’ and ‘b) an expert’.
Review valence. To check if review valence was perceived correctly, participants were asked if the opinion expressed in the review they read was predominately positive or negative.
Movie unfamiliarity. The movie described in the reviews was made irreducible by means of fictional names. To check if this was successful, familiarity with the movie was also checked.
3.5 Independent variables
Movie preference. To identify mainstream and arthouse movie audience, an indicator
item to define moviegoers’ attitudinal preference was designed for this study. Exploratively, moviegoers’ behavioural preference was measured as well. Behavioural movie preference was measured using the item: “The cinema that I visited most recently mostly shows:” and attitudinal movie preference was measured using the item: “In general, I prefer:”. For both questions, participants could choose either ‘(1) arthouse movies’ or ‘(2) commercial movies’ as their answer. To ensure participants understood the meaning of these movie categories, both answer options included a short description of their typical characteristics. Attitudinal and behavioural preference were strongly correlated (r = .79, p < .001; Cohen, 1988).
3.6 Dependent variables
