10/08/2020 Business Administration
dr. Frederik Situmeang Academic year 2019/2020
Bachelor's Thesis and Thesis Seminar Management in the Digital Age
Semester 2, period 1-3
Sequel art online - the effect of sentiments in the
first edition on sequels’ ratings.
Statement of Originality
This document is written by Student Alexey Lischuk who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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.
Abstract:
Nowadays sequels create most of our entertainment consumption but not many people realize or understand what goes into creating them or why they work. This paper aims to shed some light on this through the use of linear regression by looking at the sentiments in the reviews of the first installations and sequel scores. The data was collected from metacritic.com. Results revealed that critics’ sentiments have no effect, while users do. Moreover their sentiments are mediated by the first installation rating.
Section 1: introduction
The second half of the 20th century saw a significant increase in the production and sales of means that transmit - via satellite - information and data. It is there that the history of the TV began to flourish - production and sales skyrocketed, becoming ever so attractive as a good for the consumer as a possibility for profit for the producers. Today, the modern entertainment industry has grown to become one of the worlds most significant industry sectors. Consisting of various franchises, its value today lays well above many billions of dollars. These franchises are usually built upon sequels of original installation, meaning that they need a foundation on which to build their empires upon . One of these franchises is The Star Wars, originally born from the trilogy of films created in 1977, 1980, and 1983; and having since largely contributed to the history and development of the entertainment industry. The success of the films later led to the franchise being extended by prequel-movies, introduced in the years 1999, 2002, and 2005. However being successful in the production, a specific focus needs to be attributed to the fact that the majority of fans of the original trilogy did in fact not enjoy the prequels, whereas those who were unexposed to the franchise previously seemed to find enjoyment in the movies. Without taking the quality of the prequel movies into account, there can be a certain reason for the disparity in opinions This can be attributed due to the reason that in the minds of the fans the
original trilogy created a legacy, and, as a result, led to high expectations by them which the prequel trilogy did not uphold or meet. This is an indication that a consumer/customer’s judgement and evaluation of the sequel is dependent and may as well depend on the behaviour previous to the viewing of the movies - thus meaning that informing oneself about the previous installation and/or evaluation of it can influence perception of a product or service. This has also been confirmed by the study conducted by Situmeang et al., (2014). Moreover, the study confirmed that the evaluations/reviews of the previous installations influence success of the sequels and that it attributes this behaviour to a phenomenon called the carry over effect. Furthermore, carry over effect extends throughout the whole series (Situmeang et al., 2014). It can be argued that because of this effect, reviews of previous editions can affect/predict success of a next iteration in a franchise. As such, Kposowa 2015 examines the dependency of the sequel success on the reviews, coming to a conclusion testifyin in favour of the first installation reviews affecting sequels’ success. Both studies ultimately note that the reviews include both critic and customer evaluations of a product.
Reviews have become an essential part in the digital age as a tool for the evaluation of a product or service by a consumer. Thus, it should come with no surprise that the popularity of a product is dependent on the reviews of the same product. Consumer behaviour is influenced by the consumers ability to access information through Information and Communication Technologies which then will direct or influence a consumer choice to purchase a specific product. Information gathered by a customer through product description and/or online review ultimately results in that the electronic word of mouth constitutes an element responsible for reducing uncertainty in the consumers decision making, as a study conducted by Sharma et al. (2011) can testify for.
With the spread and the increased use of the internet, online reviews created by customers have also experienced a rise in importance to other customers (Situmeang et al., 2014), and as a result there has been an increase in the area of research into online reviews. The increased interest into the relationship between online market behaviour and online customer reviews can be attributed to two main reasons: first one being that firms would be able to understand their customers closely by reading feedback and closer targeting practices, and second reason being the creation of a new field of study of online consumer behavior. The concepts from traditional consumer behavior would change when applied to the online behavior. One of these new/modern concepts is the electronic word-of-mouth. In their study, King et al (2014) conducted an extensive literature review on eWOM, where they proposed several suggestions for further research. One of such literature gaps was a question regarding the ways consumers analyze eWOM content in a textual form - in other words, an online review. Moving in that direction, Purnawirawan et al (2015) conducted a meta-analysis on a topic of online reviews, more specifically their valence. What they found was the fact that the perceived usefulness of reviews depends whether they are mostly negative or positive. In the case of them being completely negative - reviews are untrustworthy, however if they are balanced in negativity then there is no such effect. Using the existing literature review and the research that has been conducted in this field, the main implication of this study to this paper is to generate an idea of what constitutes a balanced review, be it negative or positive.
To contribute to the validity of this paper, it is necessary to understand as to what elements create a trustworthy online review. It is an important question as online reviews, because of the power of the internet, tend to have a fast and wide spread across the globe. Currently, due to the work of Packard and Berger (2017), it is known that one of such elements is the text itself. More specifically, whether it is explicit or implicit, with the former being
illustrated by “I recommend it”, while latter - “I like it”. Therefore, language and exact words of a message have a role in the credibility of product reviews. Still, it is possible to delve even deeper into the reviews and research the review sentiments and the extent of their influence. Lee, et al. (2017), found that the negative reviews seem to be more trustworthy. However, the same study later concluded that the extreme negative emotions in a review tend to decrease trustworthiness of a review.
Combining everything said previously, this paper will investigate the effect of sentiments on a sequels' ratings', or to be more precise the polarity scores in the reviews. Thus, this paper will answer the following research question: To what extent does sequel’s rating depend on a polarity score of the first installation?
This paper will analyse the academic theory necessary to generate a good understanding of the topic regarding the research question, as well as the hypotheses. Section 3 and section 4 will describe the methodology and the results of an analysis, respectively. Finally, this study will discuss the results, as well as study the limitations and weaknesses of this study - this section will also include suggestions for further research and the implications this study could have on the academic and business world. The final section of this study will be the conclusion
Section 2: theoretical framework
First of all, this paper starts by understanding what a sequel is. A sequel is a continuation of a franchise in the entertainment industry. Sequels, as well as the original works they branched off, are, in nature, experience goods, that is those which are evaluated post-purchase. Therefore, in order to secure time and money people depend on the reviews of the experience goods. At the very least, expert reviews convey information regarding quality of the good (Hilger et al., 2011). Thus, several independent studies have been conducted to explore relationships between reviews
and movies’ success. It was found that the success of the movies depends on the reviews of it (Liu, 2006, Hilger et al., 2011, Duan et al.,(2008), Kim et al., 2013).
During these studies two success factors were found: volume and valence of the reviews, however initially only volume was largely credited. Valence of a review is a degree of positive or negative assessments in a review (Bae & Kim 2013). Liu (2006) was one of the first to research the volume effect. It was found that the amount of talk surrounding the movie contributed to more potential customers becoming exposed to it and consequently watching said movie. Later however, Cui et al., (2012) found that while volume is more important than valence, especially if it is in the early sales stage, valence still plays a role in selling an experience good. The findings were reinforced by the study of Bae & Kim 2013. Their research suggests that volume of reviews is significant only in the first week after product release, while any time after that valence is more prevalent. Judging from these studies, it is fair to say that volume and valence of reviews affect success of the experience goods
If experience goods are affected by the reviews, then it is fair to say that sequels to movies, books, video games etc. should be affected by the two types of independent reviews: reviews for the sequel itself and reviews for the previous installation. While the effect of the first type can be clearly seen, the second type seems to be less obvious. One of the researchers who inquired about this effect was Kposowa. In their study Kposowa (2015) found that one of the reasons for the change in return on investment are reviews of previous installations from critics and general public. It is possible to attribute these results due to the information and a general attitude generated by the original installation, also known as carry-over effect.
The carry-over effect is partially responsible for the sales of a sequel according to Situmeang et al. (2014). Their study concluded the existence of a positive relationship between positive past evaluations and sequels’ sales, thus confirming carry-over effect. Moreover, it was
found that the carry-over effect exists across the whole series, meaning that all past evaluations have a role in the success of a sequel. Due to this effect it is possible to theorize whether the sentiments in the reviews can carry over to sequels as well. Thus it is natural to discuss the relationship between sentiments and reviews
Current literature confirms the existence of a relationship between sentiments in the reviews and the product/service being reviewed. For instance, a study by Li, Xiaolin 2019 found, on the example of tablets, that sentiments as well as ratings in reviews significantly affect sales of the product. Sentiment analysis often includes objectivity and subjectivity as a lense of view. Thus, Jeong, Hyo-Jin 2015 found that the message usefulness is affected the highest if the review is objective and negative, while positive messages (both objective and subjective) influence attitudes towards buying a product. From this, an inference can be made that positive messages, and by extension sentiments as a whole, affect sales of a product. Moreover, similar results were found in the study of Zablocki, Agnieszka 2016 but regarding the emotions. More specifically that positive emotions positively affect product attitude, while negative and mixed decrease the attitude. Overall, it can be claimed that sentiments can be analyzed through the polarity score, that is whether the word/sentiment is inherently positive/negative/neutral/ or both
The extent that subjectivity has on the review has been researched before. A research by Liu et al.(2018) has shown us that subjectivity affects purchase intention of people depending on two factors: whether the customer is a male or female and whether the context is hedonic or utilitarian. Their study concluded that a combination of a male consumer with hedonic context and a female consumer with utilitarian context are affected by the degree of expressed subjectivity in the reviews. However this paper will not deal with distinguishing the genders of reviewers, yet the context might be of use. Reason being that this paper researches video games
and they are widely considered to be of hedonic nature, even if some subcategories of video games can be classified otherwise.
At the same time, it is necessary to realize that the source of the review plays an important role in its perceived usefulness and credibility. In an online environment two most popular and known sources of reviews are either reviews by the expert or by the regular consumers. A study conducted by Wang et al.(2020) researched how people react to disparity in the opinions of experts and average consumers in their respective reviews. Their results were twofold. Firstly, in the case of the majority of user reviews being mostly negative, while experts are positive people are more prone to be affected by the latter. On the other hand, negative expert reviews do not seem to persuade consumers to listen to their opinion when the majority of user-created reviews are positive. In that case, consumers will side with the other consumers. Overall, study of Wang et al.(2020) confirms “positivity bias”, meaning that positive evaluations seem to be more trustworthy.
Considering everything written above, this paper presents two hypotheses with two sub-hypotheses. It is done, because this paper is interested in knowing how critics and users evaluate the product. As such, the following hypotheses were created: Hypothesis 1(A): critics’ polarity score of the first installation will have an effect on the sequels’ rating given by the critics. Hypothesis 1(B): users’ polarity score of the first installation will have an effect on the sequels’ rating given by the users.
Hypothesis 2(A): critics’ effect of initial installation polarity score on sequels’ rating given by the critics will be mediated by the rating of the first installation given by the critics. Hypothesis 2(B): users’ effect of initial installation polarity score on sequels’ rating given by the users will be mediated by the rating of the first installation given by the users.
This paper employed cross-sectional design in order to answer its research question. Convenience sampling was used, as all the data was collected from online video game database metacritic.com. Sampling consisted of video games which were sold for the playstation video game consoles. This includes games sold for all the consoles between playstation 1 and playstation 4. Further segmentation of sample was conducted by leaving out the video game series without the sequels, and those without critic or user reviews. Overall, the targeted sample consisted of 113 video game series, with at least 1 sequel which were published on the playstation 1, 2, 3, and 4. Examples of this are: “Mass Effect” and “Mass Effect 2”.
Because this research aims to explore relationships between sentiments and ratings of the video games, naturally independent variables would be sentiment scores while ratings would be dependent variables. However, due to hypotheses which distinguish between critics and reviews, it was decided to have two independent variables each for its own hypothesis. Thus, hypotheses regarding critics had critics’ average sentiment score of the first installation in the video game series as its independent variable, while users - users’ average sentiment score. In similar vein, dependent variables were chosen to be critics’ average rating of the sequels and users’ average rating of the sequels. For clarification, further in this paper “average sentiment score of the first installation in the video game series” will be called “initial sentiment score” for the sake of brevity. At the same time, the second set of hypotheses have a rating score of the first installation as its mediator.
In order to calculate the sentiments in the reviews by the critics and users alike, the MPQA subjectivity lexicon was used. The purpose of this dictionary was to assign the numeric values to the words appearing in the reviews based on their prior polarity. According to MPQA polarity shows whether the word in itself, that is without any context, is causing something positive or something negative. Thus the words with positive prior polarity were given a value of 1, negative -1, while both or neutral received the value of 0. With this information, each review was analyzed and was assigned with summarization of the polarity scores. However, there would have been a problem to continue without modifying this summary. For instance, the review consisting of five words and four of them being positive would score less than the review with five hundred words consisting of two hundred positive and hundred negative. Thus it was
decided to divide the sum of polarity scores in each review by the number of words in a review. This would bring everything to the same scale. As a result, sentiment scores could be reliably analyzed.
After all of this, hypotheses could be tested. In order to do so it was decided to use linear regression. Main reason being that hypotheses are interested in the linear relationship between independent and dependent variables. For hypotheses 1(A) and 1(B), independent variables would be the initial sentiment score of critics for 1(A) and the same but for users for 1(B). Dependent variables, on the other hand, would be sequels’ rating given by critics 1(A) and users 1(B). The mediator for the regression analysis was chosen to be the score of the first installation in the video game series given by the critics or users depending on the hypothesis being tested. In order to test the direct/indirect effect it was chosen to employ PROCESS macro of Hayes 2018 in SPSS. Moreover, in order to do that, it was decided to normalize the data in order to bring all variables on the same scale. During all of the analyses, the sentiment score of the sequel given by the critics 1(A), 2(A) or users 1(B), 2(B) was chosen to be a control variable.
Table 1: means, standard deviations, and correlations table Variables M SD 1. 2. 3. 4. 5. 6. 7. 8. 1.Critic, 1st, sentiment score 0.09 0.45 - 2.Critic, 1st, rating 72.74 8.66 0.458** - 3.Critic, sequel, sentiment score 0.08 0.51 0.272** 0.587** - 4.Critic, sequel, rating 73.65 11.67 0.251** 0.585** 0.469 - 5.User, 1st, sentiment score -0.22 0.64 0.09 0.05 0.05 0.11 - 6. User, 1st, rating 7.43 1.53 0.07 0.01 -0.05 -0.09 0.41** - 7.User, sequel, sentiment score -0.26 0.64 0.07 0.06 0.13 0.15 0.46** 0.123 - 8. User, sequel, rating 7.43 1.59 -0.1 -0.08 0.06 -0.001 0.00 0.21 0.36* * - ____________________________________________________
** significant at p-value <0.01, * significant at p-value <0.05 n=113
Table 1 serves as an initial glimpse into the data because it presents means, standard deviations, and correlations of all variables. First of all, in order to compare the ratings between users and critics it is important to clarify that essentially they do not differ that much from each other. Reason for this is that metacritic.com allows critics to assign rating scores with any number between 0 and 100, while users can only assign in the range from 0 to 10. Therefore, if user ratings scores were to be multiplied by 10 similar ratings could have been achieved. As such, the means of the rating scores of the first installation in the video game series and their subsequent sequels by the critics and users appear to be similar, but not their standard deviations.
Thus, the ratings for the first installation have a mean of 72.74 (SD=8.66) by the critics and 7.43 (SD=1.53) by the users, whereas for the sequels it is 73.65 (SD=11.67) from the critics and 7.389 (SD=1.59) from the users. Interesting notion is that while means seem to be similar, standard deviations differ from each other, except for the variables related to ratings given by the users. At the same time, table 1 also shows the means of the sentiment scores. As can be clearly seen, critics tend to write their reviews with positive or neutral subjective words with means of 0.078 (SD=0.45) for the first installations and 0.09 (SD=0.51) for the sequels. On the other hand, users tend to achieve negative polarity scores with the means of -0.22 (SD=0.66) for the first installations and -0.26 (SD=0.64) for the sequels.
Moving next, table 1 presents the correlations. It is possible to see that most of the variables correlate with each other. Among those all variables related to critics have significant correlation with each other. It can be seen that critics’ average sentiment score has moderate positive correlation with the critics’ first installation rating score and weak positive correlation with the sequel score rating given by the critics(r=0.457705, p<0.00, and r=0.251173, p=0.007286, respectively). On the side of user variables, user sentiment score has no correlation with sequel score and has moderate positive correlation with the score of the initial installation(r=0.412035, p=0.000006, and r=0.000314, p=0.997368, respectively). After conducting descriptive analysis, all variables were standardized in order to have the same scale for comparison sake.
Because this paper proposes a linear relationship between the variables it is necessary to ascertain that linear regression assumptions are met. Critic related variables had shown to have met all the assumptions except having several outliers. These outliers were cut from the sample bringing it from 113 to 107 data points. Therefore critic related variables are safe to use in linear regression. The same could not have been said about user related variables. Reason being that most of the assumptions, except multicollinearity, for the linear regression have not been met. Nevertheless, the linear regression was still conducted. However it is imperative to notify that all results related to the user variables should be taken with extreme caution. Moverover, the user's independent variable was found to have 4 outliers which were cut from the sample, thus bringing total count to 109.
Table 2 shows the results of the conducted linear regression. Hypotheses 1 proposed the positive linear relationships between average sum of sentiment in the first installation of the video game series and subsequent sequels’ ratings. Looking at the table it can be seen that hypothesis 1 (A) is rejected because, first of all, even if the model explains 17% of the overall variance it is not significant ( p<0.00). Even more so, the main variable of interest indicates that it is nonsignificant and has an effect of β=0.124 ( p=0.123) on the overall dependent variable. Unlike hypothesis 1(A) 1(B) has found significant results. Overall, the model explains 0.17 of variance while being statistically significant, (p=0.066). Moreover, sentiment score β=-0.145 (p=0.066). Therefore, it can be said that one hypothesis was rejected while the other was not.
Table 2 also presents results of hypotheses 2. To test the hypothesis, which is that users’/critics’ effect of initial installation polarity score on sequels’ rating given by the users/critics will be mediated by the rating of the first installation given by the user/critics, PROCESS macro (model 4) of Hayes(2018) was used. Results clearly indicate that there is no direct effect of critics’ sentiment score of the first installation on the sequels’ rating given by the critics(β=-.1129, p=0.1169, 95%CI=[-0 -.2544; 0.0287 ]). However, they also show that there is a significant indirect effect of the independent variable through the first installations’ rating score (β=-0.2367, 95%CI=[0.1134; 0.3711]). In dissimilar vein, results of the user related variables show that there is an existence of direct effect (β=-0.2402, p=0.0814, 95%CI=[-0.4016; 0.0788]), but there is an indirect one (β=0.0953, 95%CI=[0.0319; 0.2104]).
Section 5: Discussion, limitations, and suggestions for further research.:
The goal of this research was twofold: First, to discover the extent of linear relationship between sentiments in the first installation of a movie, video game, book, etc. franchise and rating score of a sequel. And second, to see the extent of mediation in that relationship when the first installation rating is involved. Regarding the first goal, results clearly show that there is an effect of sentiments in a review by the users of a first installation on a sequels’ ratings, but there is no effect when it concerns critic related reviews. For critics that means that the emotions they put and convey through their reviews of the first installation do not influence their judgments of
sequels. On the other hand, user generated reviews and sentiments in them decrease the ratings of the subsequent sequels.
Secondly, the conducted analysis seems to bring more favorable results to this research. This means that Hypotheses 2(A) and 2(B) were found to be true. They state that the rating score of the first installation in the franchise mediates the relationship between the sentiments in the first installation and the sequels’ rating. Results indicate that sentiments in both user generated and critic generated reviews have an effect on the first installations’ rating. This in turn affects sequels’ ratings. The implication of this is as follows: the first installation in a franchise’s rating is influenced by the sentiments in the reviews, which in turn, affect sequels’ rating. Therefore, from the very first review fellow users and critics alike affect most if not all subsequent ratings in a franchise.
When comparing the results of this research to others it is important to note the difference in measuring success of, in this case, video games. As it turns out, most of the research regarding video games use games sales as an indicator. However this research employed ratings as a substitute. Therefore, from here on both ratings and sales are seen as the indicators of success for a video game series, all in order to compare the findings of this research with others.
With this distinction in mind, the first comparison is with the study conducted by Li, Xiaolin 2019. This research is similar in the way that it also finds the confirmation of ratings mediation on the sentiments in the reviews and their effect on the success of a product. The difference lies in the products being reviewed. This paper focused on video games and their sequel, which are experience goods, whereas Li, Xiaolin 2019 on consumer electronics in the form of tablets. Moreover, this research made a distinction to divide reviews into ones from critics and consumers alike. As a result of this, it can be seen that the effects of these two groups do not coincide. Therefore, it is advised, when analyzing a review to group them according to the expertise level of a reviewer.
Furthermore, results indicated that the first installations in a video game series have an impact on all the succeeding games after it. This seems to coincide with results of the research which inspired this paper, that is, a study of Situmeang et al. (2014). The main difference is that this paper aimed to discover the reason for this, firstly, by proposing sentiment effect and,
secondly, mediating the role of the first installation rating. Results showed that with critics, whereas sentiments in the first installation by themselves do not carry any effect, they do have an effect once they are mediated by the first game rating. On the other hand, user sentiment scores in reviews affect sequel ratings with and without mediation. Although, it needs to be said that in case of mediation it gains a positive effect . As such, both expert and non-expert reviews can make a difference in the video game franchise through the sentiments in the reviews of the very first installation. To play the devil’s advocate, these findings might be a result of the model variables having a high correlation with each other.
Going further, the findings of this study have a contradiction with one findings of Kposowa (2015) which stated the fact that success of a movie sequel depends on the critics’ and casual filmgoers’ reviews of the previous installation. Before discussing the differences, it is important that this paper as well as one of Kposowa (2015) agree on the fact that user reviews decrease the score of the subsequent sequel or in more general terms, that parent score is higher than the sequel one. The difference lies in the fact that sentiments in, and by extension critic reviews have no effect on the score of a sequel, whereas Kposowa (2015) argues the opposite. The difference can be attributed to the fact that one research looks into the video games while the other into movies.
Limitations and suggestions for further research
As with every study, this one has several limitations. The most obvious and important one is, undoubtedly the fact that user-related variables have been found to violate most of the assumptions of normality. As a consequence, it is advisable to approach results of this research with extreme caution. On the other hand, the same results seem to be in line with the results of different studies, therefore there might be some truth in them.
Because of the decision to focus on ratings, this research was constrained by the lack of academic research on the ratings of a sequel product as a dependent variable. As it was mentioned above, most of the studies focus on sales rather than ratings. Therefore, it makes a comparison harder and less reliable. With all that said, it is also an advantage because there is not
much focus on ratings themselves. Reason for this is that ratings can serve as a heuristic in the consumer's choice process during the reading a review. As a result, future research should focus on the effects sentiments have on sales.
Next to that stands the notion that Another limitation of this paper, as well as a point of interest for future studies is that this research was interested in the effect of sentiments in reviews on the whole series of a franchise. This serves as a good starting point in realizing that this effect does exist, therefore it makes sense to research the effect in more depth. For instance, it might be worth looking into the relationship between immediate sequels. For example, if the sentiment effect from the first installation affects second installations in the same way it affects thirds and so on. Main implication of this is that it would help predicting success of sequels based not only on the immediate “parent” video game, movie or book, but also on the series as a whole.
Lastly, study of Tsao, Hsiu-Yuan ; Chen, Ming-Yi ; Lin, Hao-Chiang Koong ; Ma, Yu-Chun(2019) showed that brand strength plays an important role through moderation. Thus because this paper did not include any moderators it might suffer in that regard. It is important as Tsao, Hsiu-Yuan ; Chen, Ming-Yi ; Lin, Hao-Chiang Koong ; Ma, Yu-Chun(2019) found that stronger brands suffer from negative reviews more than smaller brands, and that at the same time that positive reviews do not affect strong brands as much as smaller ones. Therefore, future research might want to incorporate that into their moderators or control variables.
The results of this study propose several implications for business as well as the academic world. This research might become useful for businesses because online word of mouth is an important aspect of marketing. As such, businesses should note that the results showed the fact first installation is important for the whole franchise. Though the direction of the second installation is significant, as when it diverts from the fans’ liking of the first installation, their expectations might not be met or exceeded. Thus, when making plans for creating a franchise, companies should put extra effort into creating successful first installations as well as interpreting the liking of it. As it is said “you have only one first impression”, therefore make it count, because then due to carry-over effect sequels retain some positive impression of the first installation. Moreover, the results of this study give insight into the reviews sentiments and their
importance. More specifically, according to the findings, user created reviews can seem as a good indicator of the future product’s rating.
As for academic implications, it can be said that the results of this paper add more insight to the existing literature on the topic of sentiments in the reviews. It adds that critic and consumer generated reviews have different effects on the product being reviewed and it’s future continuations.
Section 6: Conclusion:
In conclusion, there existed a gap in the literature on the topic of sentiments effect on sequels rating. Initially, this paper proposed for the existence of a relationship between sentiments in the very first installation and sequels rating. Results indicated that depending on who wrote the reviews, that is critics or users, the sequel’s rating will change accordingly. Moreover, this paper was interested in seeing whether mediation takes place through the first installation score. As it turned out it did. In the end of it all, the main message that this paper presented is that the companies and/or independent users who want to create a franchise should pay extra attention to how the first installation fares and act accordingly.
Bibliography :
Chen, M., 2016. Can two-sided messages increase the helpfulness of online reviews? Chen, Ming-Yi Online Information Review,, Vol.40(3), pp.316-332
Cui, Geng ; Lui, Hon-Kwong ; Guo, Xiaoning. The Effect of Online Consumer Reviews on New Product Sales International Journal of Electronic Commerce, 01 October 2012, Vol.17(1), pp.39-58
Duan, Wenjing ; Gu, Bin ; Whinston, Andrew B. Do online reviews matter? — An empirical investigation of panel data Decision Support Systems, 2008, Vol.45(4), pp.1007-1016
Feng, Z., & Xiaoquan Z., (2010). Impact of Online Consumer Reviews on Sales: The Moderating Role of Product and Consumer Characteristics. Journal of Marketing, 74(2), pp.133-148
Hilger, J., Rafet, H., 2011, Expert Opinion and the Demand for Experience Goods: An Experimental Approach in the Retail Wine Market, S ofia Review of Economics and Statistics, 2011, Vol.93(4), pp.1289-1296
Hou Yong ; Wang Tie-Nan ; Li Xiang-Yang, 2013. How successful movies affect performance of sequels: Signal theory and brand extension theory in motion picture industry. International Conference on Management Science and Engineering 20th Annual Conference Proceedings, pp.798-806
Jeong, Hyo-Jin ; Koo, Dong-Mo. Combined effects of valence and attributes of e-WOM on consumer judgment for message and product Internet Research, 02 February 2015, Vol.25(1),
Jungho, Bae ; Byung - Do, Kim. Is the electronic word of mouth effect always positive on the movie? Academy of Marketing Studies Journal, Jan, 2013, Vol.17(1), p.61(18)pp.2-29
Ken, H T, Sun, J. 2018
The Differential Effects of Online Peer Review and Expert Review on Service Evaluations: The Roles of Confidence and Information Convergence Journal of Service Research, , Vol.21(4), pp.474-489
Ketron, S.(2017). Investigating the effect of quality of grammar and mechanics (QGAM) in online reviews: The mediating role of reviewer crediblity . Journal of Business Research, Vol.81, pp.51-59
Kim, Sang Ho ; Park, Namkee ; Park, Seung Hyun. Exploring the Effects of Online Word of Mouth and Expert Reviews on Theatrical Movies' Box Office Success Journal of Media Economics, 01 April 2013, Vol.26(2), pp.98-114
King, R. A., Racherla, P., & Bush, V. D. (2014). What we know and don't know about online word-of-mouth: A review and synthesis of the literature. Journal of Interactive Marketing, 28(3), 167-183:
Kposowa K., 2015, The Financial Success of Franchise Film Sequels: An Exploration of the Relationship of Budget, Personnel Factors, and Reviews with Sequel Return on Investment
“Wii Will Rock You!” The Use and Effect of Figurative Language in Consumer Reviews of Hedonic and Utilitarian Consumption Kronrod, Ann ; Danziger, Shai Journal of Consumer Research, 01 December 2013, Vol.40(4), pp.726-739
Li, Xiaolin ; Wu, Chaojiang ; Mai, Feng. The effect of online reviews on product sales: A joint sentiment-topic analysis Information & Management, March 2019, Vol.56(2), pp.172-184
Liu, S. Q., Ozanne M., Matilla, A. S., (2018) Does expressing subjectivity in online reviews enhance persuasion?. Journal of Consumer Marketing, 35(4), pp.403-413
Liu. Word of Mouth for Movies: Its Dynamics and Impact on Box Office Revenue , Yong Journal of Marketing, 1 July 2006, Vol.70(3), pp.74-89
Packard, G., Berger, J.,(2017). How Language Shapes Word of Mouth's Impact. Journal of Marketing Research,, Vol.54(4), pp.572-588
Purnawirawan, N., Eisend, M., De Pelsmacker, P., & Dens, N. (2015). A meta-analytic investigation of the role of valence in online reviews. Journal of Interactive Marketing, 31, 17-27.
Situmeang, F., Leenders, M., & Wijnberg, N. (2014). The good, the bad and the variable: How evaluations of past editions influence the success of sequels. European Journal of Marketing, 48(7/8), 1466-1486.
Sharma, Ravi S. ; Morales - Arroyo, Miguel ; Pandey, Tushar, 2011.The emergence of electronic word-of-mouth as a marketing channel for the digital marketplace Journal of Information, Information Technology, and Organizations,Vol.6-7, p.41(21)
Tsao, Hsiu-Yuan ; Chen, Ming-Yi ; Lin, Hao-Chiang Koong ; Ma, Yu-Chun, 2019. The asymmetric effect of review valence on numerical rating Online Information Review, Vol.43(2), pp.283-300
Wang, J., Molina, M., Maria D., Sundar S., 2020When expert recommendation contradicts peer opinion: Relative social influence of valence, group identity and artificial intelligence.
Zablocki, Agnieszka ; Makri, Katerina ; Houston, Michael J. Emotions Within Online Reviews and their Influence on Product Attitudes in Austria, USA and Thailand.(Report)(Author abstract) . Journal of Interactive Marketing, 2019, Vol.46, p.20
Appendix: Critic:
● Normality: ○ linearity:
○ Normality of residuals:
○ Homoscedasticity:
○ Multicollinearity
Users:
● Normality:
○ Linearity:
○ Homoscedasticity of residuals:
○ multicollinearity
Descriptives:
Regression: user:
