Factors affecting consumer loyalty in role-playing and life-simulation video games
Consideration of the main factors influencing consumer loyalty in role-playing games and video games that simulate life. General characteristics and features of the conceptual model of online games loyalty. Acquaintance with modern gaming technologies.
| Рубрика | Менеджмент и трудовые отношения |
| Вид | курсовая работа |
| Язык | английский |
| Дата добавления | 26.08.2020 |
| Размер файла | 1,7 M |
Отправить свою хорошую работу в базу знаний просто. Используйте форму, расположенную ниже
Студенты, аспиранты, молодые ученые, использующие базу знаний в своей учебе и работе, будут вам очень благодарны.
All items were measured via seven-point Likert scales, where 1 refers to strongly disagree and 7 is for strongly agree. Likert scale was developed to measure attitudes and personality traits and is frequently used by researches in social and business sciences. The most common Likert scale used by scholars is a 5- or 7-point scale applied to measure the degree to which respondents agree or disagree with a statement, where 1 - strongly disagree, 2 - disagree, 3 - somehow disagree, 4 - cannot agree nor disagree, 5 - somehow agree, 6 - agree, and 7 - strongly agree . It is also important to distinguish Likert scale data from Likert-type data. Likert-type questions are designed to measure one variable per question, whereas Likert scales include several Likert-type questions measuring 1 variable (Boone & Boone, 2012). These questions are then combined to provide estimates of the variable. In order to show high correlation between items within one measure, the reliability and validity of collected data were tested by the presented paper in the analysis part of the study.
3.2 Survey instruments and data collection
The data was collected by two online survey questionnaires. The two surveys were adapted from previous researches and adjusted to fit the specific settings of two examined games. The survey instruments for RPG and the sources of adaptation are illustrated in Appendix 1. Survey instruments for LSG were stated in the same way with the exception of using “Sims” instead of “WoW” in each question. Pilot tests with 25 RPG and LSG players were conducted to ensure that questionnaires were well interpreted by respondents. According to the comments of pilot tests' respondents, some questions were modified.
The subjects of the research are players of World of Warcraft (RPG) and The Sims (LSG). The users were reached via game groups in the Internet, streaming services' forums, and comment sections under the streamers' videos about the examined games. Although many researchers studying consumer behavior in online games use university students as survey subjects, such sample may not be very representative. According to Gamesparks, a game data service, the average age of video game players is 35 years old (Gamesparks, n.d.). Consequently, having the sample of more diverse age groups gathered directly from interest-based communities might be more informative and help to avoid the generalization problem (Wu et al., 2008). Altogether, 464 users completed two survey questionnaires, where 224 responses were for RPG and 240 - for LSG.
After the results were collected, an overview of respondents' profiles was obtained. The characteristics of RPG and LSG players were summarized and represented in Table 1 and Table 2, respectively. On the whole, about 95% of RPG respondents are male and 5% are female, whereas for LSG female predominate, composing about 82% of respondents, while male - 18%. Such dramatically different demographic results for two types of games of this survey can arise the problem of associating different results for loyalty factors with demographic differences of two games. However, this issue can be mitigated by the general demographic distribution of players in game industry. The studies of various researches suggest that the majority of RPG players were young male (Kendall, 1999; Taylor, 2009; Williams et al., 2008). Furthermore, the game analytics consulting company Quantic Foundry has reported that from 270,000 observations 18.5% RPG players are female (Yee, 2017).
Considering the LSG demographics, one of the biggest and famous companies producing simulation games, Electronic Arts, have reported that 51% of their customers are females (Electronic Arts, 2019). In addition, according to the Quantic Foundry's research the average percent of female gamers for family/farm simulations accounted for 69% (Campbell, 2017). The slight difference between simulation games statistics and obtained results for LSG can be explained by less male representatives in the game communities. Taking into account the survey results obtaining for this research and overall demographic statistics for RPG and LSG market, the problem of relating loyalty factors differences with gender of players is irrelevant. Moreover, the similarities between the industry statistics and survey results may contribute to more representative population sample.
To proceed with survey respondents' characteristics, more than 59% of role-playing gamers have an experience with the game for more than 6 years, whereas 39% of LSG users play it for >6 and 31% for 1-3 years. Such results advocate for overall high experience with the games and make the results for loyalty factors more accurate. Since the major aspect of how the game is experienced is intrinsic motivations, experienced users give more valid assessment of their game enjoyment (Mekler et al. 2014).
Table 1. RPG respondents overview
|
Measure |
Items |
Percent |
|
|
Gender |
Male |
95,1% |
|
|
Female |
4,9% |
||
|
Age |
<20 |
2,2% |
|
|
20-25 |
30,4% |
||
|
>25 |
67,4% |
||
|
Education |
high school |
8,0% |
|
|
college/university degree |
90,6% |
||
|
graduate institute |
1,3% |
||
|
Years of playing the game |
<1 |
2,2% |
|
|
1-3 |
8,0% |
||
|
4-6 |
30,4% |
||
|
>6 |
59,4% |
||
|
Hours a day playing the game |
<1 |
0,4% |
|
|
1-3 |
12,5% |
||
|
4-6 |
57,6% |
||
|
>6 |
29,5% |
Table 2. LSG respondents overview
|
Measure |
Items |
Percent |
|
|
Gender |
Male |
18,4% |
|
|
Female |
81,6% |
||
|
Age |
<20 |
29,3% |
|
|
20-25 |
54,4% |
||
|
>25 |
16,3% |
||
|
Education |
high school |
18,8% |
|
|
college/university degree |
70,7% |
||
|
graduate institute |
10,5% |
||
|
Years of playing the game |
<1 |
14,6% |
|
|
1-3 |
31,0% |
||
|
4-6 |
14,6% |
||
|
>6 |
39,7% |
||
|
Hours a day playing the game |
<1 |
22,2% |
|
|
1-3 |
55,6% |
||
|
4-6 |
10,5% |
||
|
>6 |
11,7% |
3.3 Analysis
As one can tell from the methodology discussion and the overall description of the study, the analysis process includes several important steps.
The primary stage of the process is descriptive analysis which presents the exploratory study of the gathered datasets. The goal of this part is to provide some insight on collected data and determine its internal patterns that can impact the interpretation of the results of the main analysis and limitations of the research. To achieve that goal the key estimates of the variables (number of observations, mean, standard deviation) are presented. Worth mentioning, Likert scale data that was obtained by the two questionnaires cannot be normally distributed, according to its definition (Norman, 2010). Even though Likert data has been of the most adopted methods of collecting and measuring character traits of respondents, there has been a lot of debate over whether parametric or nonparametric statistics should be used to analyze it. Likert data is statistically considered to be ordinal data and some researches argue whether it can be treated as interval data when converted to numbers (Carifio & Perla, 2008). Since the exact difference between “disagree” and “somehow disagree” can be distinct for respondents, the main argument against using Likert data is misinterpretation of means and standard deviation when applied to scale responses. With that being said, experts have argued that parametric tests should not be used for analyzing Likert scale data. Conversely, some researches hold the opinion that parametric statistics can be utilized with Likert scale data (Boone & Boone, 2012.; Norman, 2010). For instance, the study of Norman (2010) showed that both Pearson and Spearman correlations yielded same results even with non-normally distributed data and a small sample size (93 observations). Thus, many researchers claim that ordinal data should be analyzed with the median or mode as the central tendency measurements because the arithmetical manipulations are not appropriate for verbal statements (Clegg, 1984). However, regarding the Likert scales in particular, both previous studies and some statisticians agree that one should employ the mean as the measure of central tendency and treat the data as continuous interval because items that compile the scale convert rough ordinal measures into metric data (Chang et al., n.d.; Teng & Chen, 2014; Tseng et al., 2015). After thorough investigation of the topic in the available literature, the data was summarized. Table 3 represents the general view on Likert data and what descriptive statistics should be used for Likert Scale Data. The results of descriptive statistics based on the information presented above is discussed in the Results section of the study.
Table 3. Suggested method of analysis of Likert Scale Data*
|
Likert-Type Data |
Likert Scale Data |
||
|
Data type |
Ordinal categorical |
Continuous interval |
|
|
Central Tendency |
Median or mode |
Mean |
|
|
Variability |
Frequencies |
Standard deviation |
The second step of the analysis process is to conduct reliability and validity tests of the data. This stage is required since for any statistical analysis it is essential to critically evaluate the measures with weak psychometric characteristics, otherwise the results of the analyses will be meaningless (Kline, 2011).Reliability and validity tests were conducted in SPSS software via Factor Analysis to establish internal consistency between items within each theoretical construct. The detailed description of the process and the results are presented in the Results section of the study.
The third and the main step of the analysis process is implementation of Structural Equation Modeling (SEM) using Partial Least Squares method (PLS). SEM is a statistical modelling technique that is strongly relied on theoretical constructs, which are displayed as latent variables (Kline, 2011). These latent variables are hypothetical factors hence they cannot be measured quantitatively. The relationships between those theoretical constructs are represented by path coefficients between the latent variables. The structural equation modelling implies a prior input of a model to be analysed. More specifically, it requires the information about which variables are expected to affect and be affected, and the direction of these effects. The model is often visualised graphically as a path diagram. From the technical point of view, structural equation modelling consolidates traditional multivariate techniques that measure relationships between two or more variables: factor analysis, linear regression analysis, and discriminant analysis (Hox & Bechger, 1999). As one can understand from the theoretical background part of the study, the theoretical model applied for the present research is Conceptual Online Game Loyalty Model developed by Zhao and Fang (2009).
While latent variables represent hypothetical factors (theoretical foundation), observed variables are the data depiction. These two different types of variables could be linked via the structural model. In the presented paper, observed variables include constructs' measurements obtained from the two questionnaires: GE1, GE2, GE3, SN1, SN2, SN3, etc. Latent variables are the theoretical constructs: Game Enjoyment (GE), Social Norms (SN), etc. The most used estimation method for SEM is Maximum Likelihood (ML) method (Hox & Bechger, 1999). Another common method is Partial Least Squares (PLS) that uses least-square procedure. Both methods rely on resampling to a large extent - bootstrapping. However, Sharma and Kim (2013) found that PLS bootstrap outperformed ML on smaller sample size (<500) by producing smaller bias. Generally, PLS is believed to have larger or equal statistical power on smaller sample size (<250 observations) and almost no bias estimating composite model population data (Rigdon et al., 2017).That is due to the fact that PLS approach is based upon separate ordinal least squares regressions, thus the data is not imposed with any distributional assumptions (Rodrнguez-Entrena et al., 2018). Moreover, the current research is exploratory in nature, and PLS is known for being a better fit for an exploratory approach (Chin, 1998). Subsequently, in the presented paper, Partial Least Squares method is used.
Finally, in order to obtain the significance level of indicators and path coefficients Bootstrapping procedure to 500 resamples need to be carried out. T-statistic values on 95% confidence level must be obtained. As suggested by Zhao & Fang (2009), non-parametric bootstrapping should be implemented in order to assess the results for statistical significance.
The steps of the analysis process discussed above require special software that possess a variety of statistical algorithms suitable for the research tasks. Thus, this study uses such statistics programs as SPSS, Smart PLS 3.0 and Microsoft Excel, that carry the necessary resources.
3.4 Limitations
The study has several essential limitations to be mentioned. First, one of the initial constraints is associated with generalization of the findings. Restricting the study to only two games of their respective genres may cause some problems while interpreting the results in regard to every game in the genre. The reason behind this approach has already been discussed in “The statement of the research question” part of the study and is cause by the time and scope restrictions of the research. However, the limitation it has on the interpretation of the results of the research is worth mentioning. Even though the games that are included into a certain type possess the same characteristics, for some of them it might not be true. Thus, it may not be exactly correct to assume that a certain game fully represents its type. For instance, World of Warcraft, despite being a perfect example of a role-playing game, is not exact copy of another RPG, Fallout. However, these two games are similar in their mechanics and technological factors. Consequently, the results may be biased, because the responses are influenced by impressions of a single game, for which the survey is designed for. For this reason, test findings should be viewed with precaution in regard to all games of the genre as more research is required to fully assure the applicability of the results to every RPG and LSG game.
Second, taking into consideration that the current paper is a survey-based study, another limitation associated with the users' responses exists. The tendency to report socially acceptable answers and to conceal socially undesirable behavior causes the social desirability bias (Beck & Ajzen, 1991). This limitation may affect the findings of the present research. However, the survey questionnaires were designed and adapted from the previous research in such way to eliminate this bias. Thus, this study suggests that its effect on the results should not be regarded as significant.
Third, it is feasible that cultural and geographical differences of the respondents may influence the results. The survey questionnaires were spread in communities and platforms that can be accessed from various geographical locations. However, among all theoretical constructs, the influence of those differences might be only significant to social norms factor and its impact on game enjoyment, intentions to play and game loyalty. Some cultures tend to rely more on social norms and socially acceptable behavior than other cultures (Gelfand, 2019). Thus, the extent to which social norms affect aforementioned factors could be possibly caused by the cultural differences of the respondents to some degree.
Finally, it is possible to associate the differences in the results of loyalty-affecting factors for two game genres with demographic differences of the respondents. Since the majority of LSG's respondents are female, whereas of RPG - male, this may also influence the conclusions of the research. However, as discussed in the previous subsection of this chapter, this problem is mitigated by the general statistic of entire LSG and RPG populations that claims similar demographical distribution for this game genres.
Overall, the established research question, hypotheses and theoretical foundation made it possible to come up with the respective methodology that involves data collection, its processing and analysis. However, inevitable limitations occur along the research process. The research design that was presented in this chapter allows to fill in the gap in current scientific understanding of consumer behavior in the gaming industry by examining consumer loyalty from the perspective of two game genres.
4. Results
4.1 Descriptive analysis
To give an overview of the data that was collected the results of a descriptive analysis are presented. A brief description and summary of the dataset, its internal patterns and characteristics that can influence the results of the analysis, are essential for interpreting the results and limitations of the research.
The overall dataset used for the research includes records of 20 variables (10 for each type of game), that were measured via Likert scales, each of which consisted of 3 Likert-type questions for each variable.
Consequently, taking into account that the current study employs Likert scale data, mean as the central tendency value and standard deviation as variability value should be measured. Table 4 and 5 shows the means and standard deviations of each construct for the results of the two questionnaires. The findings show that on average users responded positively on playing RPG and LSG. However, the majority of respondents of both games claimed that social norms are not important for their attitude towards the games. Moreover, for life-simulation players, the quality of online game community is not relevant. It is also worth noting that Likert Scale data cannot be normally distributed, which is followed from its definition and discussion above (Norman, 2010). As suggested by Zhao & Fang (2009), non-parametric bootstrapping should applied order to confirm the results for statistical significance, which imply that there is no necessity to test the data for normality.
Table 4. RPG: descriptive analysis
|
Mean |
SD |
||
|
GE |
6,98 |
0,17 |
|
|
SN |
1,99 |
1,17 |
|
|
IP |
5,96 |
0,87 |
|
|
GG |
6,28 |
0,94 |
|
|
GS |
6,53 |
0,70 |
|
|
GL |
6,53 |
0,78 |
|
|
PEU |
3,06 |
0,73 |
|
|
QS |
6,14 |
0,87 |
|
|
GC |
6,57 |
0,94 |
|
|
L |
6,07 |
0,77 |
Table 5. LSG: descriptive analysis
|
Mean |
SD |
||
|
GE |
6,38 |
0,91 |
|
|
SN |
1,50 |
0,78 |
|
|
IP |
6,18 |
1,20 |
|
|
GG |
6,04 |
1,03 |
|
|
GS |
5,36 |
1,58 |
|
|
GL |
5,68 |
1,25 |
|
|
PEU |
6,47 |
0,99 |
|
|
QS |
4,08 |
1,15 |
|
|
GC |
2,70 |
1,53 |
|
|
L |
5,83 |
1,07 |
4.2 Reliability and validity of the data
Based on the previous studies and recommendations of the theoretical model developers, Structural Equation Modelling (SEM) as the method for this research was employed. However, for any statistical analyses it is crucial to identify and critically evaluate measures with weak psychometric characteristics, otherwise the results of the analyses will be meaningless (Kline, 2011). Thereupon, reliability and validity coefficients must be reported for RPG and LSG measurements. Overall, the data obtained from the questionnaires is deemed acceptable.
In order to check data reliability and validity, a Factor Analysis in SPSS was conducted. Reliability was tested using composite reliability method (ICR), which is similar to Cronbach's Alpha since both of the methods are used to measure reliability. This statistic estimates internal consistency reliability, which is the degree to which responses of the surveys are consistent among the questions (items) within a measure (Kline, 2011). After obtaining Rotated Component Matrix (Appendix 2 and 3), factor loadings were used in order to further investigate and generate Average Variance Extracted and Composite Reliability. Then, using factor loading squared and Error Variance, Average Variance Extracted and Composite Reliability were computed. As reported by Brown (2006) and Brunner & SЬв (2005), a composite reliability of 0.70 and above is acceptable. Average variance extracted is bearable when 0.50 or above. Table 6 and Table 7 illustrate the results for Internal Consistency Reliability analysis using Composite Reliability. It shows that composite reliability ranges from 0,83 to 0.95 (RPG) and from 0.79 to 0.94 (LSG), while average variance extracted ranges from 0.65 to 0.87 (RPG) and from 0.54 to 0.85 (LSG). Thus, all ten items demonstrate high loadings (>0.70) on their corresponding constructs and load more robust on their respective constructs than on other constructs.
Table 6. RPG: composite reliability
Table 7. LSG: composite reliability
To establish construct validity, convergent and discriminant validity were calculated. Average variance extracted higher than 0.50 is not sufficient for proving construct validity. Thus, discriminant validity calculated using Fornell and Larcker Criterion was used. First, average loading of each variable was found, following by average variance extracted and average variance extracted between two variables. Then, using component correlation matrix (Appendix 4 and 5) obtained by the factor analysis squared correlation was calculated. Tables 8 and 9 show that average variance extracted of each variable is higher than the squared correlations between variables and all other variables of two examined datasets. Consequently, construct validity was established. To sum up, the results in Tables 6, 7, 8 and 9 provide sufficient evidence to argue that reliability and convergent and discriminant validity of the measurement instruments were established.
4.3 Validity of the sample size
SEM is a technique that requires large number of observations in order to avoid technical problems and achieve a reliable outcome with high statistical power. Furthermore, the complexity of the model is directly proportional to the sample size it requires (Kline, 2011). With the sample size being such a crucial and sensitive factor, researchers have come up with a variety of fit indices. According to Jackson (2003), the minimum sample size is the number of cases (N) to the number of model parameters (q). The ideal N:q ratio would be 20:1 (Kline, 2011). Consequently, if there are q=1o theoretical constructs in our research model that demand statistical estimates, the ultimate sample size of observations is 200. Thus, the number of observations collected for the research (224 and 240) is acceptable.
Another support for the sample data size in the current study is the average number of observations that is applied to SEM-based studies. Such data was collected by Breckler (1990), who reviewed 72 articles, and Shah & Goldstein (2006), who analysed 93 management science articles. Their observations approved 200 cases to be the median sample size of SEM applications, while the studies with samples less than 200 are often rejected by journal submissions' reviewers.
4.4 Test of the model and hypotheses
Since the main method of the analysis is structural equation modeling with PLS approach, the hypotheses were tested by examining the path coefficients between variables and their significance levels in the structural model. This method of analysis is analogous to standardized beta weights in a regression analysis. The path coefficients are the standardized regression coefficients from separate linear regressions computed for each relationship between the latent constructs, as shown in Figure 3 and Figure 4. This study implemented Smart-PLS 3.00 (2019) software to perform the calculations, the original results can be found in Appendices 6-9. The significance level of indicators and path coefficients were evaluated via bootstrapping procedure to 500 resamples to obtain t-statistic values on 95% confidence level.
Figure 3 and Figure 4 show path coefficients and significance levels for each hypothesis in the case of two games. The variance for the three dependent constructs, game enjoyment, intention to play and game loyalty, is explained in bold. The variances are the percentage of the variance explained in each construct. For example, in Figure 3, the theoretical model explains 67% of the variance in game loyalty. More specifically, for RPG, game technology factors, such as game story, game graphics, game length, perceived ease-of-use and game services, as well as game community together explain 31% of the variance in online game enjoyment. Game enjoyment, social norms and game community account for 42% of the variance on players' intention to play. The total variance in game loyalty is explained by 67% by the conceptual model, which indicated that the overall model is acceptable in explaining the variance (Falk & Miller, 1992) . For LSG, game technology factors and game community explain 27% of the variance in game enjoyment, while game enjoyment, social norms and game community explain 33% of the variance in intention to play. The total variance in game loyalty is explained by 62% by the conceptual model. To sum up, the study reveals that both RPG and LSG players' loyalty can be explained by the conceptual online game loyalty model (R2=67% for RPG, R2=62% for LSG).
Figure 3. RPG: Results of Structural Modeling Analysis
Figure 4. LSG: Results of Structural Modeling Analysis
The results of the structural modeling analysis presented in the figures above, allow to support or reject the formulated hypothesis. For RPG, 11 out of 13 hypotheses are supported, as shown in Table 10, which summarizes the results of the analysis. Consistent with the prediction, social norms, intention to play and game community are significant predictors of online game loyalty for RPG players, thus hypothesis 1, 2 and 3, which are directly related to loyalty, are supported.
Considering the indirect factors, their effects are caused through the main mediator, intention to play, which is significantly influenced by social norms and game enjoyment. Social norms have some effect on intention to play, although it is not very powerful, thus hypothesis 4 is supported. Game enjoyment is important predictor of intention to play RPG games, thus hypothesis 6 is supported. However, intention to play is not influenced by game community, so hypothesis 5 cannot be confirmed. Considering game enjoyment, hypothesis 7, which posits that social norms predict game enjoyment, is not supported since the path between these two constructs is insignificant, according to the analysis. Game technology factors are all significantly related to game enjoyment, hence hypotheses 8-12 are supported. The effect of game community on game enjoyment is also substantial, which confirms hypothesis 13.
For the life simulation game, 9 out of 13 hypotheses are supported, as shown in Table 11. Hypotheses 1 and 2, that posit social norms and intention to play to be positive predictors of game loyalty, are supported. However, in line with the prediction, the importance of game community for game loyalty is insignificant, thus hypothesis 3 is not supported.
Speaking about the factors that affect game loyalty indirectly, the main mediator, intention to play is significantly affected by both social norms and game enjoyment, supporting hypotheses 4 and 6, respectively. However, intention to play is not affected by game community, thus hypothesis 5 is failed to be supported. Similar to the RPG results, social norms do not relate to game enjoyment, which results in fail to support hypothesis 7. All game technology factors are significantly related to game enjoyment, supporting hypotheses 8, 9, 10, 11 and 12. Moreover, the relation of game community to game enjoyment is insignificant, thus hypothesis 13 is not supported.
4.5 Additional results
As stated by the conceptual model that is used for the analysis, both direct and indirect effects predict game loyalty. Hence an additional information about the impact of the indirect factors can be provided. In order to evaluate indirect effects on game loyalty, a mediator analysis of the model should be carried out. To perform a mediator analysis, the significance of direct and indirect paths should be taken into account. According to Zhao et al., (2010), if p1 and p2 is the indirect effect, p3 - direct, and the indirect path p1*p2 is insignificant, while p3 is significant, no mediation occurs, and the effect is direct-only. Moreover, the total effect is measured by the following formula: (p1*p2) + p3 (Zhao et al., 2010). Tables 12 and 13 show total effects that includes both direct and indirect effects. Following from this, for the RPG players game enjoyment has a statistically significant indirect effect on loyalty through intention to play. Moreover, game enjoyment shows full mediation via intention to play, meaning that it has indirect-only effect on loyalty. Similarly, for the life-simulation gamers, game enjoyment presents indirect-only effect on loyalty via intention to play. The statistical significance of the indirect effects was also measured via bootstrapping to 500 resamples, after which the t-statistic was obtained, as shown in Table 14 and Table 15. Thus, the indirect effects provide additional information on what can possibly predict online game loyalty. The importance of these effects will be further discussed in the following section of the study.
Game technology factors are not essential predictors of game loyalty, as shown in Tables 12 and 13. However, their effect on game enjoyment is significant, whereas game enjoyment has a strong indirect effect on game loyalty. Thus, this study suggests that the importance of game technology factors should not be underestimated. The indirect effect of technological factors of the two examined games may be insignificant inasmuch as the sample size is not sufficient to examine the indirect effects through two mediators.
4.6 Goodness of fit of the model
Goodness of fit for a model is based upon a comparison of the covariance matrices of the saturated versus estimated model. However, the process of PLS-SEM does not imply covariance-based structural equation modeling, so the covariance matrices is not what the model is built on. Thus, some researchers advise to be extremely cautious while reporting model fit for PLS-SEM (Hair et al., 2017). Even though others have requested to use new model fit indices for partial least squares approach to SEM, the proposed methods are not fully understood yet and are often not very useful (SmartPLS, 2019). However, the approximate model fit for PLS can be estimated via SRMR and NFI indices. The model fit is considered satisfactory with the certain threshold: SRMR<.08 and NFI>.90 (L. Hu & Bentler, 1998). Table 16 and Table 17 provide the mentioned indices and support a satisfactory fit of the models for two datasets.
5. Discussion
5.1 Summary of RPG results
The results of the current study show that for a role-playing game intention to play, online game community and social norms are important predictors of customer loyalty. Delving deeper, intention to play is the strongest predictor of game loyalty, whereas intention to play itself is predicted by social norms and game enjoyment, where game enjoyment is the most powerful influential factor. Interestingly, game community possesses no effect on intention to play RPG, but significantly impact both game enjoyment and game loyalty. Similar result was shown by the study of Zhao and Fang (2009), which indicated the direct effect of quality of online game community on game loyalty, but no effect on neither game enjoyment nor intention to play. As discussed in the theoretical part of the study, game communities provide useful information for gamers, as well as the social community where users can interact and share their thoughts or game achievements. Thus, being a source of information about the game, communities may not influence the players' intention to play the game itself. Moreover, some gamers even prefer to participate in game communities without playing the game. Therefore, this study believes that game community is not a crucial factor for understanding gamers' intention to play RPG, however, it is an essential predictor of game loyalty for role-playing games.
The results also point out the explanatory variables of game enjoyment of RPG. All game technology factors, including game story, graphics, length, perceived ease-of-use and customer support have significant impact on game enjoyment. As discussed above, game community also has a strong effect on game enjoyment. On the contrary, social norms does not have any crucial effect on game enjoyment, corresponding to the results of Zhao and Fang (2009). In addition, according to the additional results of the analysis, game enjoyment has an indirect-only effect on RPG game loyalty. This outcome proves the initial idea of the current study that it is critical to investigate not only direct factors affecting loyalty, but the indirect causes as well.
To conclude with the results of the RPG analysis, eleven out of thirteen paths were found statistically significant, thus eleven out of thirteen hypotheses were confirmed. The remaining two paths, from social norms to game enjoyment, and from game community to intention to play, were not found statistically significant. Such results confirm the predictions of the analysis, which highlighted the importance of game community for game loyalty of RPG players. Consequently, the answer to the stated research question sounds as following: for the role-playing games, intention to play, game community and social norms affect game loyalty directly, whereas game enjoyment has an indirect effect on loyalty via intention to play. The contribution of these outcomes will be further addressed in the next part of the study.
5.2 Summary of LSG results
The analysis of the LSG structural equation modeling has confirmed the significance of intention to play and social norms for game loyalty, while online game community does not play any significant role in repetitive game consumption. Intention to play is the strongest predictor of game loyalty, while intention to play is predicted by social norms and game enjoyment. Remarkably, LSG's community does not have a significant effect on any of the dependent variables. Thus, this study suggests that there is no important correlation between game community and game loyalty for life-simulation gamers.
Another conceivable finding is that game enjoyment of players of life simulators is predicted by game technology factors only, whereas neither game community nor social norms have a significant effect on game enjoyment. Consistent with the findings of a supplemental analysis, game enjoyment has an indirect-only effect on game loyalty through intention to play. Thus, both RPG and LSG loyalty is influenced not only by direct effects, but by indirect factors as well, supporting the idea that game loyalty should be examined from all perspectives for any game genre.
To conclude with the results of the LSG analysis, nine out of thirteen paths were found statistically significant, thus nine out of thirteen hypotheses were supported. The remaining four paths, from game community to game loyalty, from game community to intention to play, from game community to game enjoyment, from social norms to game enjoyment, were found statistically insignificant. Thus, the final outcome confirms the expected results of the analysis, which assumed that quality of the game community is not a decisive factor for predicting life-simulation games customer loyalty. Hence the research question concerning LSG can be answered: intention to play and social norms are the factors that affect LSG loyalty directly, whereas game enjoyment impact LSG loyalty indirectly.
5.3 Contribution and key insights
As a result of structural modeling analysis and presentation of the findings, this research makes several contributions to the overall understanding of consumer behavior and customer loyalty in online video games. First, a major contribution of the current study is that it highlights the difference between factors affecting game loyalty for two game genres, role-playing games and life-simulators. Prior research in online gaming industry examined the impact of social norms, intention to play and game community on players' loyalty towards the games. Moreover, the differences and similarities of the investigated game genres were presented, assuming that factors that would influence game loyalty vary for RPGs and LSGs. This research indicates that quality of online game community is important influential factor for RPG loyalty, whereas for LSG this factor is insignificant. Thus, these results fill in the gap correlated with the lack of research on different game genres and contribute to the study of customer behavior on online gaming.
Second, the present study confirms the findings of the previous research conducted in the field of consumer behavior in video and online games. More explicitly, the results support positive direct relationships between intention to play online games and customer loyalty, and between social norms and game loyalty for both examined game genres. This outcome is consistent with the prior research on game loyalty, which provides additional evidence that aforementioned factors are crucial triggers for game loyalty (Choi & Kim, 2004; Teng & Chen, 2014; J. Wu et al., 2008; F. Zhao & Fang, 2009). In addition, the current study confirms the critical factor of intention to play online games, game enjoyment, as well as its significant indirect effect on game loyalty, also supported by the studies of Chang et al. (n.d.), J. Wu & Liu (2007) and Zhao & Fang (2009).
Another important contribution of this study is that it tests the Conceptual Online Game Loyalty Model proposed by Zhao & Fang (2009) in the context of two different game genres, while previous research has concentrated on online games as a whole. In addition, as mentioned in the theoretical background section of the study, conceptual model implemented for testing the hypotheses is an adaptation of TRA. Thus, consistent with the theory and the conceptual model, social norms and players' positive attitude (i.e. game enjoyment) towards playing have significant impact on intention to play both RPG and LSG.
Fourth, this study investigates the predictors of game enjoyment, which is an important factor explaining users' behavioral intentions to play online games and customer loyalty. Consistent with the findings of Zhao & Fang's research (2009), game technology factors have significant impact on game enjoyment of both examined game genres. However, quality of online game community is not critical predictor of game enjoyment of LSG, while for RPG it plays an important role for players' positive game experience.
5.4 Implications for future research
The present study is based on the assumption that factors affecting game loyalty vary across different game genres. Hence it is motivated by the need to understand the relationships between game-related factors and players behavior towards different types of games. This study and its results yield the following theoretical implications for future research.
First, this study presents the evidence of dissimilarities in the degree to which several factors affect online game loyalty of two game genres. This information provides an important implication for further research of consumer behavior in different game genres. On the one hand, as for the fact that this study examines only two games of respective genres, a more thorough research on other games of these genres is required. As stated by the limitations of the current study, examining only two games of each genre may have some restrictions on the implications of the results to the overall game community. Thus, future research can reexamine the correlations between the factors and game loyalty in regard to several games of LSG and RPG or any other genres. On the other hand, this study finds that game community does not possess any significant effect on game loyalty for LSG. This implies that game communities, being a source of information about the game, may or may not influence continuous intention to play. Due to such distinctions in the results of two examined game genres, more research on game communities and its drivers is needed. Moreover, future research should reexamine the relationship between these two constructs to determine if this result is valid.
Second, the present study shows the significance of game enjoyment for predicting behavioral intentions of both games. Such results, also supported by previous research, strikes the attention of the researchers on consumer behavior and is worth further investigation. This study finds out that game technology factors impact game enjoyment, as well as quality of online game community (only RPG). Consequently, more investigation is required to determine what are the other possible factors that can influence players' enjoyment of online games.
5.5 Practical implications
As noted in the introduction section of this study, this research is aimed to develop several insights for the gaming industry practitioners. Hence the present study generates the consequent implications for game developers, vendors and marketing teams in gaming industry worldwide.
First, online game loyalty for RPGs and LSGs is predicted by social norms. Thus, this study suggests that game vendors and marketing teams should build strong relationships with opinion-makers and influential people in the gaming community. For instance, a partnership with Twitch streamer who has a mass following and regular streams may influence players to stay longer in the game. Thus, opinion leaders that take part in building the social norms can impact users' participation in the game and increase average retention time.
Second, online game community significantly impact loyalty of RPG players, whereas for LSG players it does not have any substantial effect. Thus, this research suggests that RPG developers and technical teams should concentrate on keeping game communities alive and thriving, while investing in LSG communities is not a necessity.
Third, because game enjoyment is important indirect predictor of online game loyalty regardless of the game genre, managers and game developers should try to stimulate users' intrinsic motivations. This can be done through factors that directly affect game enjoyment: game technology factors, and, in the case of RPG, game community. Thus, developers and designers of any game genre should strive to develop a game with captivating game story and game graphics, appropriate game length, easy-to-use interface and game design, and helpful game service. Furthermore, RPG game community should also be maintained in order to achieve maximum game enjoyment of the players, which may result in increased customer loyalty.
Conclusion
The present study investigated factors crucial for loyalty of role-playing and life-simulation games. Considering the nature of online video games, the examined factors related to social norms, game technology factors, quality of game communities and intrinsic motivations are vital for improving consumer retention and loyalty towards the games. The conceptual model used for the analysis of the relationships between the factors and game loyalty is valid and reliable in explaining customer behavior in the context of online games.
One of the objectives of the study was to compare the results for the two examined games in regard to the factors affecting game loyalty. Dependent upon the findings of the analysis, one of the most impactful factors for RPG gamers' loyalty is online game community, which contributes not only to game enjoyment but to continuous intention to play above all. On the contrary, loyalty of LSG players is not predicted by game community, although the impact of social norms, intention to play and game enjoyment is as significant for LSG loyalty as for the RPG loyalty. Consequently, these results incite important managerial application of the study. Game developing companies should strike their attention on establishing and growing online game communities for role-playing games, whereas life-simulation games are not required that precise attention from managers to their online communities since it does not influence players' loyalty in any way. Noteworthy, this study also revealed the significant indirect effect of game enjoyment on game loyalty for both examined game genres. Hence it provokes another crucial practical implication related to improving game technology factors that are involved in influencing the feeling of enjoyment while playing the game.
The results of this research, however, should be accepted and interpreted with caution for a number of reasons. First, the initial difficulty is the limitation on generalization. Since only two games of their respective genres are analyzed, it may restraint the interpretation of the findings in regard to all games of RPG and LSG genres. Such approach was chosen due to the time and scope restrictions of the research. However, as discussed earlier, the games chosen for examination possess all the features that describe their genres. Thus, although this limitation exists and influences the results, this effect is not ravaging for the research findings. Second, taking into account that this research is a survey-based study, a social desirability bias may also affect the results of the analysis. The bias is caused by the tendency of the online survey respondents to report socially acceptable answers and to conceal undesirable behavior. Finally, the limitation of geographical location and demographic characteristics of respondents can be present too. The survey questionnaires were spread in communities and platforms that can be accessed from various geographical locations, resulting in possible influence of cultural specificities on the examined factors. Regarding the respondents' demographics, this problem was mitigated by the statistic on overall population and is no longer relevant.
The aforementioned results and limitations present the areas for future research. In order to fully confirm the differences in the loyalty-affecting factors for RPG and LSG and reassure validity of the relation of game community to game loyalty, further research involving more games of each genre is required. In addition, as shown by the results, game enjoyment is important predictor of game loyalty. Thus, more research is needed to determine what are other possible factors that can influence players' intrinsic motivations to play online games.
Reference list
1.Ajzen H, & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Prentice-Hall, Inc., Englewood Cliffs, NJ, 1980. https://www.scienceopen.com/document?vid=c20c4174-d8dc-428d-b352-280b05eacdf7
2.Anderson, R. E., & Srinivasan, S. S. (2003). E-satisfaction and e-loyalty: A contingency framework. Psychology and Marketing, 20(2), 123-138. https://doi.org/10.1002/mar.10063
3.Arts, E. (2019, January 16). Diversity & Inclusion--About EA. Electronic Arts Inc. https://www.ea.com/about/diversity-and-inclusion
4.Bajaj, A., & Nidumolu, S. R. (1998). A feedback model to understand information system usage. Information & Management, 33(4), 213-224. https://doi.org/10.1016/S0378-7206(98)00026-3
5.Beck, L., & Ajzen, I. (1991). Predicting dishonest actions using the theory of planned behavior. Journal of Research in Personality, 25(3), 285-301. https://doi.org/10.1016/0092-6566(91)90021-H
6.Boon, R. (2002). Forty Hour Milestone. Computer Trade weekly, 877, 2002. https://blog.ihobo.com/2015/09/forty-hours.html
7.Boone, H. N., & Boone, D. A. (2012). Analyzing Likert Data. Journal of Extension, 50(2). Retrieved:http://www.joe.org/joe/2012april/t t2.php
8.Breckler, S. J. (1990). Applications of Covariance Structure Modeling in Psychology: Cause for Concern? Psychol Bull, 107, 260-273.
9.Brown, G. H. (1953). Brand Loyalty--Fact of Fiction. The Trademark Reporter, 43, 251.
10.Brown, T. A. (2006). Confirmatory factor analysis for applied research. Guilford Press.
11.Brunner, M., & SЬв, H.-M. (2005). Analyzing the Reliability of Multidimensional Measures: An Example from Intelligence Research. Educational and Psychological Measurement, 65(2), 227-240. https://doi.org/10.1177/0013164404268669
12.Campbell, C. (2017, January 20). Which games are women and girls playing? Polygon. https://www.polygon.com/2017/1/20/14337282/games-for-women-and-girls
13.Carifio, J., & Perla, R. (2008). Resolving the 50-year debate around using and misusing Likert scales. Medical Education, 42(12), 1150-1152. https://doi.org/10.1111/j.1365-2923.2008.03172.x
14.Chang, K. T.-T., Koh, A. T.-T., Low, B. Y.-Y., Onghanseng, D. J. S., Tanoto, K., & Thuong, T. S. T. (n.d.). Why I Love This Online Game: The MMORPG Stickiness Factor. in Proceedings of the 29th International Conference on Information Systems, Paris, France.
15.Chen, Y., Chen, P. S., Hwang, J., Korba, L., Song, R., & Yee, G. (2005). An analysis of online gaming crime characteristics. Internet Research, 15(3), 246-261. https://doi.org/10.1108/10662240510602672
16.Chin W. (1998). "The Partial Least Squares Approach to Structural Equation Modeling," in Modern Methods for Business Research, G. A. Marcoulides (ed.), Lawrence Erlbaum Associates, Mahwah, NJ, 293-336, 1998.
17.Choi, D., & Kim, J. (2004). Why People Continue to Play Online Games: In Search of Critical Design Factors to Increase Customer Loyalty to Online Contents. CyberPsychology & Behavior, 7(1), 11-24. https://doi.org/10.1089/109493104322820066
18.Chuan-Chuan Lin, J., & Lu, H. (2000). Towards an understanding of the behavioural intention to use a web site. International Journal of Information Management, 20(3), 197-208. https://doi.org/10.1016/S0268-4012(00)00005-0
19.Clegg, F. (1984). Simple Statistics: A Course Book for the Social Sciences. Cambridge University Press.
20.Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User Acceptance of Computer Technology: A Comparison of Two Theoretical Models. Management Science, 35(8), 982-1003. JSTOR.
21.Drake, C., Gwynne, A., & Waite, N. (1998). Barclays Life customer satisfaction and loyalty tracking survey: A demonstration of customer loyalty research in practice. International Journal of Bank Marketing, 16(7), 287-292. https://doi.org/10.1108/02652329810245993
...Подобные документы
Value and probability weighting function. Tournament games as special settings for a competition between individuals. Model: competitive environment, application of prospect theory. Experiment: design, conducting. Analysis of experiment results.
курсовая работа [1,9 M], добавлен 20.03.2016Impact of globalization on the way organizations conduct their businesses overseas, in the light of increased outsourcing. The strategies adopted by General Electric. Offshore Outsourcing Business Models. Factors for affect the success of the outsourcing.
реферат [32,3 K], добавлен 13.10.2011The audience understand the necessity of activity planning and the benefits acquired through budgeting. The role of the economic planning department. The main characteristics of the existing system of planning. The master budget, the budgeting process.
презентация [1,3 M], добавлен 12.01.2012The concept, essence, characteristics, principles of organization, types and features of the formation of groups of skilled workers. The general description of ten restrictions which disturb to disclosing of potential of group staff and its productivity.
реферат [29,7 K], добавлен 26.07.2010Critical literature review. Apparel industry overview: Porter’s Five Forces framework, PESTLE, competitors analysis, key success factors of the industry. Bershka’s business model. Integration-responsiveness framework. Critical evaluation of chosen issue.
контрольная работа [29,1 K], добавлен 04.10.2014Considerable role of the employees of the service providing company. Human resource policies. Three strategies that can hire the right employees. Main steps in measure internal service quality. Example of the service profit chain into the enterprise.
презентация [338,7 K], добавлен 18.01.2015The main idea of Corporate Social Responsibility (CSR). History of CSR. Types of CSR. Profitability of CSR. Friedman’s Approach. Carroll’s Approach to CSR. Measuring of CRS. Determining factors for CSR. Increase of investment appeal of the companies.
реферат [98,0 K], добавлен 11.11.2014M.A. Rothschild is a German banker and the founder of the Rothschild banking dynasty, business leader, which is believed to have become the wealthiest family in human history. A brief sketch of his life and career. Main stages of empire building.
презентация [425,6 K], добавлен 10.06.2014The concept and features of bankruptcy. Methods prevent bankruptcy of Russian small businesses. General characteristics of crisis management. Calculating the probability of bankruptcy discriminant function in the example of "Kirov Plant "Mayak".
курсовая работа [74,5 K], добавлен 18.05.2015Major factors of success of managers. Effective achievement of the organizational purposes. Use of "emotional investigation". Providing support to employees. That is appeal charisma. Positive morale and recognition. Feedback of the head with workers.
презентация [1,8 M], добавлен 15.07.2012Evaluation of urban public transport system in Indonesia, the possibility of its effective development. Analysis of influence factors by using the Ishikawa Cause and Effect diagram and also the use of Pareto analysis. Using business process reengineering.
контрольная работа [398,2 K], добавлен 21.04.2014Analysis of the peculiarities of the mobile applications market. The specifics of the process of mobile application development. Systematization of the main project management methodologies. Decision of the problems of use of the classical methodologies.
контрольная работа [1,4 M], добавлен 14.02.2016Factors that ensure company’s global competitiveness. Definition of mergers and acquisitions and their types. Motives and drawbacks M and A deals. The suggestions on making the Disney’s company the world leader in entertainment market using M&A strategy.
дипломная работа [353,6 K], добавлен 27.01.2016Origins of and reasons for product placement: history of product placement in the cinema, sponsored shows. Factors that can influence the cost of a placement. Branded entertainment in all its forms: series and television programs, novels and plays.
курсовая работа [42,1 K], добавлен 16.10.2013History of development the world leader in the production of soft drinks company "Coca-Cola". Success factors of the company, its competitors on the world market, target audience. Description of the ongoing war company the Coca-Cola brand Pepsi.
контрольная работа [17,0 K], добавлен 27.05.2015Nonverbal methods of dialogue and wrong interpretation of gestures. Historical both a cultural value and universal components of language of a body. Importance of a mimicry in a context of an administrative communication facility and in an everyday life.
эссе [19,0 K], добавлен 27.04.2011Description of the structure of the airline and the structure of its subsystems. Analysis of the main activities of the airline, other goals. Building the “objective tree” of the airline. Description of the environmental features of the transport company.
курсовая работа [1,2 M], добавлен 03.03.2013Процесс стратегического планирования: области выработки и базисные стратегии. Шаги определения стратегии предприятия. Выбор стратегии. Разработка стратегии для ООО "Москва-online" на основе SWOT-анализа. Методы и стадии реализации стратегии предприятия.
курсовая работа [46,4 K], добавлен 24.01.2008Методологічні аспекти вивчення впливу поля діяльності організації на її діяльність. Загальна характеристика та критерії оцінки поля діяльності організації, фактори та умови, які на нього впливають. Аналіз поля діяльності компанії "Life Card Group".
курсовая работа [79,8 K], добавлен 06.07.2013Праксеологічні аспекти вивчення особливостей розробки, прийняття і реалізації оперативних управлінських рішень на прикладі компанії "Life Card Group". Кількісні (аналітичні та статистичні) та якісні методи обґрунтування оперативних управлінських рішень.
курсовая работа [63,4 K], добавлен 04.07.2013
