The research selects top 15 boy groups and top 15 girl groups as the research targets according to the ranking of boy and girl groups brand reputation monthly reports in 20181. The rankings can represent the group’s relative popularity in South Korea at that month.
In addition, the research collects these 30 idol groups’ discographies until December 31, 2018. The debut discographies of the idol groups were excluded because the research
intended to analyze the transformation of music genres, therefore it left a number of 268 discographies for 30 idol groups in the research. The variables of the research are as follows:
Dependent variable
I used the logarithms (with the base 10) of the audiences’ preference score of Music Bank’s K-chart to represent the consumers’ preference (called “Preference”), and I collectd the data from the official website of Music Bank2. The following is the reason why I use the K-chart score of the music program, Music Bank, to represent the performance of the discographies.
First of all, the music programs in Korea have their own music charts, and the ranking of the chart can represent the idol group’s relative grade of the discography in the period of comeback. Among the several music programs, Music Bank on KBS (Korean Broadcasting System), Inkigayo on SBS (Seoul Broadcasting System), and M Countdown on Mnet Media are the three main music programs that are approved by most K-Pop fans.
Secondary, Music Bank starts use K-chart since September, 2007, and the evaluation standard of Music Bank has little change in recent years. Also, the evaluation standard of
Music Bank includes the sale number of the album, the amount of music digital download, the broadcast times from the television channel, and the audiences’ preference, and this evaluation standard contains every dimension of the performance of the idol group’s comeback.
Finally, the audiences’ preference score is determined by the questionnaires sent randomly and the audiences’ vote. The audiences use text message to vote for their favorite idol groups, and for the purpose of supporting their idols whole-heartedly, the hardcore fans are willing to use text message vote to give their idol groups the huge supports.
Independent variables
I use the extent of the music styles transformation to represent the extent of brand image change, and the independent variable is called “Transformation” in the research. In order to present the transformation of music genres, I find these 30 groups’ discographies, and browse the data of the discographies on Wikipedia to determine the music genres.
However, I only take the leading tracks’ music genres because the idol groups promote their leading track on the music programs at every comeback.
In summary, 268 discographies are labeled on 58 music genres. According to these genres’
origin, these 58 music genres are categorized into 11 music styles: Hip-Hop, Contemporary R&B, Rock, Mixed Pop, Electro, House, Ballad, Blues, Reggae, Disco, and Pure (See Appendix B). Among them, the “pure” category aims to separate the innocent girl group’s bubblegum pop and teen pop genres from the other pop genres, such as electro pop, and synth-pop. Then, according to the history of music genres’ evolution, I draw a graph to show those 11 music styles’ relationships (see Figure 1).
This music genre framework helps to determine the image change extent between any two music categories. One unit of the distance (i.e., the image change extent) is defined as the stem connected two nearby music categories. For example, if the discography contains the Mixed Pop music genre, and the previous discography contains the Electro music genre, then the distance of this transformation is 2 (Mixed Pop→Hip-Hop→
Electro).
Figure 1. The relationship between each music styles.
In cases that a discography contains more than one music genre, I use the simple averages to represent the extent of image change. For example, if a previous discography contains the Mixed Pop, Electro, and House music styles and the current discography contains Electro and Ballad music styles, then I will calculate the extent of the image change by counting all the distance between any two music genres identified in these two discographies and calculate the simple average of the distances (Electro→Mixed Pop:2;
Electro → House:1; Ballad → Mixed Pop:1; Ballad → House:2; Ballad → Electro:3;
transformation= 2+1+1+2+3
5 =1.8).
Control variables
This research chooses five control variables: gender, the size of the management company, the discography type, whether the idol groups’ members have attended the talent show, and the number of the year goes through since the idol groups debuted.
The gender of K-Pop idol groups may result in differences in the scores of the audience preference of Music Bank K-chart because most of the idol group’s fandoms consist of the fans with the gender opposite to the idol groups. Furthermore, most of the K-Pop fans are female, so the boy idol group’s preference score is normally higher than the girl idol group’s score. The variable “Gender” receives the value of 1 if the idol group is the girl
idol group.
In the research, the size of management company is classified in three levels, listed company, subsidiary, and non-listed company. If the management company of the idol group has already gone public, such as JYP Entertainment and SM Entertainment, then the management company will be classified as the listed company; if the management company of the idol group has not gone public, but the company is a subsidiary of the listed company, such as STARSHIP Entertainment (the subsidiary of Kakao M) and Woollim Entertainment (the subsidiary of SM Entertainment), then the company will be classified as the subsidiary; neither big company nor middle company is classified as the non-listed company. The budget invested on the idol groups usually depends on the size of management company, and the more resources the company invests, the more likely the idol groups get good grades. Thus, I set two dummy variables, “Listed Company” and
“Non-listed Company”, to represent the size of the management company according to the above definition of the company sizes, and receive the value of 1 if the management company is classified as the listed company and non-listed company respectively, and the statistics is collected from Wikipedia and the website of Korea Exchange (KRX)3, Korean security exchange operator.
In K-Pop industry, the discographies of the idol groups can be categorized into three main types, single, extended play (as known as EP), and the full album. Each type of discography contains different number of songs, and releasing the full album implies that the idol group makes an effort on this comeback and the fans are willing to bring good grade to support the idol groups’ effort. Therefore, whether the discography is the full album or not might affect the performance of the comeback, and I set the variable “Album Type” receives the value of 1 if the discography type according to the data on Wikipedia is the full album.
In recent year, the talent shows, such as SIXTEEN and PRODUCE 101 series, become popular. Some idol groups are composed through the talent shows, the other idol groups debut after the talent show. Also, some management companies even assign the debuted idol groups’ members to participate the talent show and come back to the original group after the talent show ends. Even though the Korean idol talent shows broadcast on television for only two or three months, in the process of broadcasting, the show and the contestants can gradually accumulate many supporters, and after the talent show finishes, the supporters of the program probably become the fans of the talent show’s contestants.
Hence, the variable “Talent Show” receives the value of 1 if the members of the idol groups have attended the talent shows.
The research adds the number of the year goes through since the idol groups debuted as the control variable. If the idol groups have great reputation and the audiences’ preference, then the idol groups cannot be eliminated easily and can last for the long time compared to the other idol groups. Conversely, if the idol groups last for the long time, then the idol groups can gradually accumulate their fans and thus earn the higher preferences of the audiences. Hence, the variable “Year” indicates the number of the year between the debut day and the new discography launches.
Analysis
This research analyzes the data through R programming language and uses ordinary linear regression model to prove the hypothesis. In order to clarify the non-linear relationship, this research adds the independent variable’s general form (Transformation) and quadratic term (Transformation2) into the linear regression model. Therefore, this research shows Model 1 as the baseline model, Model 2 adds the general form of the independent variable, and Model 3 includes the general form and quadratic term of the independent variable to prove the hypothesis.
Result
Table 3 provides the means, standard deviations, and the correlation for all variables, and
Table 4 provides the variance inflation factor (VIF) table to show whether the multicollinearity problem exists. According to the variance inflation factor test, the VIF of the variables are all below the threshold of multicollinearity, thus the variables in the regression model do not exist the problem of multicollinearity.
Table 5 presents the result of the linear regression analyses. Model 2 shows that the extent of the transformation of music genres is statistically significant and has the positive effect on audience preference. Model 3, compare with the Model 2, adds the quadratic term of the extent of the transformation of music genres to test the hypothesis. The result of Model 3 indicates that the general form and quadratic term of the extent of the transformation of
Table 3. Description Data and Correlation Table
Variable Mean S.D. 1. 2. 3. 4. 5. 6. 7. 8.
1. Preference 2.09 1.10
2. Transformation 1.72 0.81 0.31***
3. Transformation2 3.62 2.26 0.26*** 0.95***
4. Listed Company 0.51 0.50 0.11 0.29*** 0.24***
5. Non-listed Company 0.31 0.46 -0.12 -0.15* -0.15* -0.67***
6. Gender 0.47 0.50 0.10 0.03 0.06 -0.03 -0.03
7. Talent Show 0.17 0.37 -0.13* -0.17** -0.16** -0.06 0.01 0.00
8. Album Type 0.24 0.43 0.16** 0.07 0.05 0.13* 0.14* -0.11 -0.06
9. Year 2.51 2.14 0.30*** 0.30*** 0.27*** 0.10*** -0.12* -0.03 -0.14* 0.14*
* p < .05 ; ** p < .01 ; *** p < .001
Table 4. Variance Inflation Factor Table
music genres are statistically significant, and the Model 3 also proves the hypothesis that the relationship between the extent of incongruity and the consumers’ reaction for corporate rebranding is presented an inverted U relationship.
Moreover, this research also uses likelihood ratio test to test the model comparison between two regression models. According to the likelihood ratio test, Model 2 is statistically significantly better than Model 1, and Model 3 is also statistically significantly better than Model 2. Thus, the conclusion of likelihood ratio test indicates that Model 3, including the general form and quadratic term of the extent of the transformation of music genres, is the best regression model to explain the statistical data.
Variables VIF 1/VIF
Transformation 1.22 0.82
Listed Company 1.98 0.50
Non-listed Company 1.88 0.53
Gender 1.02 0.98
Talent Show 1.04 0.96
Album Type 1.05 0.95
Year 1.14 0.88
Table 5. Ordinary Linear Regression Analysis
Likelihood ratio test for Model 2 compared to Model 1 and Model 3 compared to Model 2.
The value in the parentheses means the t-value of each variables.
N=268
Besides, the regression results also show some insights from the effects of the control variables. One is that the variable “Album Type” is statistically significant in Model 3, showing that relative to the single or EP, the full album may earn better market reaction.
Another finding is that the variable “Gender” is statistically significant in Model 3, and the result can be explained that the dependent variable of the research is determined by the audiences’ preference score on the K-chart of Music Bank, and this score is determined by the audiences’ text message votes. Normally, most of the fans tend to support the idol groups of the opposite gender, therefore, this result implies that the male fans are more likely to vote through the text message than the female fans, and the girl idol groups are statistically significantly better than the boy idol groups on this score.
The other finding is that the variable “Year” is statistically significant in Model 3, and result indicates that as the idol groups last for longer time, the opportunities of catching the audiences’ eyes increase.