The questionnaire design was based largely on the International Men and Gender Equality Survey⁸ and the cross-national Survey of Family and Changing Gender Roles,⁹ while taking into account the gender-stereotypical behaviors reviewed in the previous section and the particular influence of Neo-Confucianism, which plays an important role particularly in East Asia.

The questionnaire consists of three sections. Section one collects basic demographic information, such as age, sex, education level, and nationality of the respondent. Section two includes five questions measuring the level of K-pop consumption of the respondent, such as the time and money spent/ willing to spend on K-pop related activities and products. In section three, respondents’ gender attitudes are measured using a set of eighteen statements. Most of these statements, gleaned from the above-mentioned surveys, contain gender stereotypical or sexist proclamations, for example, “In terms of housekeeping and taking care of children, women are naturally better than men.” Respondents are instructed to state their levels of agreement with each statement by choosing from a five-point Likert response scale (Strongly agree/ Partially agree/ Neither agree nor disagree/ Partially disagree/ Strongly disagree). Statements on gender attitudes are divided into four themes: gender identity; family life; child-raising practices; and the workplace. The complete list of statements can be found in the appendix.

The survey was disseminated in English, Korean, and Chinese. It was distributed online through three social network sites (SNS), Twitter, Facebook, and Weibo,¹⁰ for three weeks from October 21 to November 8, 2014. The distribution of the survey received major help from two K-pop idols whose images largely conform to mainstream concepts of masculinity and femininity in K-pop: Ok Taec-yeon from the boy group 2PM and Fei from the girl group Miss A. The survey collected a total of 6,317 effective observations from 100 countries, of which 98% reported to be K-pop fans.

This study uses several dependent variables to measure gender attitudes. First, a Gender attitudes index was constructed; it is a composite multi-item measure combining the eighteen attitude-related statements in the questionnaire with equal weights.¹¹ The level of agreement to each statement has been coded between 1 and 5 according to the five-point Likert scale. The index, which is the simple sum of the 18 individual variables, is therefore bound between a minimum score of 18 and maximum score of 90 to which higher values correspond to more egalitarian gender attitudes. Besides the aggregate index, three specific statements have been selected to serve as dependent variables as they can be regarded as more illustrative than a combined index. These three statements are: (A) “In terms of housekeeping and taking care of children, women are naturally better than men”; (B) “It’s embarrassing if husbands earn less than their wives”; and (C) “It is natural for men to occupy leading positions at work.” The first statement captures attitudes towards division of labor within the household with a particular focus on unpaid work. The second statement, while also focusing on division of labor, emphasizes the role of earnings, which plays an important role in the intrahousehold bargaining processes (e.g., Bethmann & Rudolf, 2016) and, therefore, in marital outcomes. It refers to the traditional norm in which the man should be the main breadwinner. Finally, the third statement concentrates on issues related to gender segregation at the work place (see also Rudolf, 2014). The top of Table 1 indicates that the mean of the gender attitudes index for the entire sample of 6,317 respondents is 52.1 (SD = 13.4). Statements are coded such that higher numbers indicate stronger disagreement with the non-egalitarian statement. Thus, as for the aggregate index, for statements A to C higher values also correspond to more egalitarian gender attitudes. Simple sample means show that the disagreement is strongest with statement B (norm of man as main breadwinner, sample mean = 3.28) and lowest with statement A (traditional division of (unpaid) labor within the household; mean = 2.66) among the three statement.

As described below, the independent variables are the respondents’ demographic characteristics and four variables reflecting the level of K-pop consumption.¹² First, K-pop fan length is measured by the question “How long have you been a fan of K-pop/ Hallyu?” Respondents can choose from the following answer categories: “For 10 or more years / For 6-9 years ago/ For 3-5 years ago/ For 1-2 years ago/ For less than a year/ Since recently/Not a K-pop fan.” The responses were coded from 0 to 6, where a higher number reflects more time being a K-pop fan. According to Table 1, 41.1% of all respondents have been following K-pop for 3 to 5 years at the time of the survey, while another 22% report to being a K-pop fan for 6 to 9 years, and another 17.8% for 1 to 2 years.

Second, K-pop fan expense reflects fans’ willingness to pay and is measured by the question “How much per month are you willing to spend on your favorite K-pop stars related commodities/ activities? (in US$)”; response categories are: “Above 1000/ 500–1000/ 200–500/ 100–200/ 50–100/ 0–50.” The responses have been coded using the mid-point value of the respective expense category. Table 1 shows that 40.4% of all respondents are willing to spend between $US 0 and 50 per month on K-pop related items; 25% are willing to spend between $US 50 and 100 per month; while 32.5% of respondents are willing to expend more than $US 100 a month on K-pop.

Third, K-pop fan exp./ Family income is the share of monthly family income that respondents are willing to spend on K-pop commodities and activities. Table 1 shows that about half of all respondents spend between 2 and 10% of their family budget on K-pop products, a non-negligible share.

Fourth, K-pop fan activities are measured via the question “How often do you listen to/ watch/ follow your favorite K-pop stars’ music/ video/ TV programs/ SNS?” Respondents can choose from the following response categories: “Multiple times per day/ Once per day/ Multiple times per week/ Once per week/ 2–3 times per month/ Once per month/ Rarely/ Not a K-pop fan.” Responses of this question were coded from 0 to 7 with a higher number indicating more frequent fan activities. Table 1 shows that most respondents (76.2%) are engaged in K-pop-related fan activities multiple times per day.

Table 1 provides additional descriptive statistics on other sociodemographic characteristics of survey respondents that are typically expected to affect gender attitudes and will be used as control variables in the regression analysis in the next section. The sample’s average age is 23.1; 96% of the respondents are female and 93% have never married. Most of the respondents are from Asian countries, and Chinese account for the largest group (27% of the whole sample), followed by Thai (26%). This is driven in part by the K-pop idols that promoted the survey.¹³ Average years of schooling are 17.3, indicating that the typical survey participant has graduated from college or is expecting to graduate from college.

How much survey participants are willing to spend in terms of money on K-pop related items is an important indicator for measuring K-pop consumption and fandom. Table 2 examines the relationship between different levels of K-pop fan expense and K-pop fan exp./ Family income and the Gender attitudes index.

Table 2 suggests that both absolute and relative K-pop consumption are negatively related to egalitarian gender roles. In particular, we see that the gender attitudes index is lower for respondents who report higher expenses on K-pop related commodities and activities.

Individual gender attitudes are a complex construct of cultural, historical, environmental, and religious factors that drive social norms and gender roles in a particular country. What is taken for granted in one country can be viewed as a taboo in another, particularly when it comes to the rights and freedoms of women (Nussbaum, 2001). For example, in Asia, we find that East Asian countries’ gender roles are driven by Confucian beliefs of a clear family hierarchy of men above women. However, these Confucian gender norms have largely disappeared if countries experienced a communist system during the past century (e.g., China and Vietnam). On the other hand, majority-Muslim countries, such as Malaysia and Indonesia, are known to cultivate traditional gender roles imposed by their religion. Western and Latin American countries are known to have more egalitarian gender roles (GGGR, 2015).

Table 3 presents average gender attitudes by country or region of nationality. As expected, we find the most egalitarian gender attitudes in North America (the U.S. and Canada), followed by Latin America and Europe. Of the major Asian countries included in our survey, respondents from Singapore, Vietnam, and the Philippines have the highest scores on the aggregate index, while the lowest scores can be observed for Indonesians, Thai, Koreans, Malaysians, and Japanese. Interestingly, the Chinese scores are only slightly higher than that of Koreans and Japanese. One can also observe interesting heterogeneity when comparing individual statements across countries. For example, Europeans and Chinese do not differ significantly in how they view the gender division of housework and childcare as well as gender segregation at work. Yet, Europeans are more open than Chinese to the possibility that a wife can earn more than her husband.

To test the hypothesis that a relationship exists between K-pop consumption and gender attitudes, this paper uses multivariate OLS regressions. Eq. (1) below summarizes the model, in which GAi is the dependent variable and refers to gender attitudes of individual i.

Kpop refers to a vector containing a set of variables to account for individual i’s level of K-pop consumption, while X is a vector that contains variables related to individual i’s personal demographic and socioeconomic background.

Table 4 presents the results of estimating equation (1) using the gender attitudes index as the dependent variable. Models (1) to (4) estimate over the entire sample, but each uses a different combination of the four K-pop consumption variables. All models control for a number of sociodemographic characteristics and the nationality of the respondent.

Models (1) to (3) include the three measures of K-pop fandom individually. Several observations can be made. The coefficient of K-pop fan length is positive, but not statistically significant in model (1).¹⁴ In model (2), the willingness to pay (WTP) for K-pop items is expressed as a share of family income. Findings show that the more respondents spend on K-pop products (relative to family income), the lower they tend to score on the gender attitudes index. For example, compared to the reference group (with a K-pop WTP of 2–5% of family income (the most frequent category in the sample)), individuals with a WTP of 10–15% of their family income score on average 3.489 points lower on the index. Model (3) controls for the frequency of K-pop fan activities and suggests that more frequent activities are related to less egalitarian gender attitudes (coefficient = -0.277).

Model (4) includes all K-pop consumption variables at the same time. It confirms the finding that K-pop expenses are negatively related to egalitarian gender attitudes. The more individuals spend on K-pop, the lower their score on the aggregate gender attitudes index. Compared to models (1) and (3), however, the coefficients for K-pop fan length and fan activities have changed. Fan length is now positive and significant, while the coefficient of fan activities has turned statistically insignificant. Given that model (4) has the highest value of R-squared, it is our preferred specification.

Regression results further suggest that respondents have more egalitarian gender attitudes (i.e., higher index scores) if they are older, better educated, if they live in smaller households, if they live in families with higher incomes, and if they have never married.

Regarding cultural differences across nations that were discussed in the last section, we find striking differences across nationalities. In line with the summary statistics in Table 3, regression results in Table 4 confirm that some nationalities have, on average, more egalitarian gender attitudes than others.¹⁵ The most egalitarian nationalities in our sample are North Americans, followed by Latin Americans, Europeans, Singaporeans, Vietnamese, and Filipinos. On the other hand, the least egalitarian attitudes can be found among Indonesians, Thai, Malay, Koreans, and Japanese. In model (5), we take the respondents from these five less-egalitarian nations and run the regression over only this subsample of respondents. Respondents from more egalitarian nations are then used in model (6). The positive coefficient of fan length in model (4) was puzzling. It actually indicates that, taking the negative effect of K-pop spending on gender attitudes aside, the longer an individual follows K-pop, the higher will be her gender attitudes index score. So, there seems to be a more complex relationship between being a K-pop fan and gender attitudes. To examine this further, we divided the sample into participants from less egalitarian nations and more egalitarian nations and estimate the regression over each subsample in models (5) and (6). This yields a very interesting finding: For participants from less egalitarian nations, fan length is no longer statistically significant and thus seems to play no role. At the same time, the negative effect of fan expenses has become stronger as compared to model (4). In contrast, for participants from more egalitarian nations, the negative effect of fan expenses has become weaker while fan length shows a significant, positive effect which is larger than in model (4). These findings suggest that K-pop indeed might have heterogeneous effects depending on who consumes it. On the one side, it appears to have a strong negative correlation with egalitarian gender attitudes among followers from nations with already high gender gaps. On the other side, negative effects of K-pop expenses seem to be moderate and might even cancelled out by fan length for followers from more egalitarian nations.

Table 5 presents the results of estimating Eq. (1) using the three thematic statements about gender attitudes as dependent variables. All three statements are typical gender stereotypes, and refer to commonly observed gender discrimination in unpaid work, pay, and promotion. It should be noted that the dependent variables are measured as the level of disagreement with each statement. Thus, it is consistent with the coding of the gender attitudes index which means that a higher number indicates more egalitarian gender attitudes.

In column (1) the dependent variable is the level of disagreement with statement A—“In terms of housekeeping and taking care of children, women are naturally better than men.” K-pop fan length and the frequency of K-pop fan activities show statistically insignificant coefficients. The share of K-pop fan expense in family income, however, shows a negative and statistically significant correlation. This indicates that a respondent with higher K-pop expense is more likely to be in favor of a traditional division of labor in which women take care of the home and the family.

With regard to statement B—“It is embarrassing if husbands earn less than their wives”—results again confirm the relationship between K-pop fan expenses and gender attitudes. K-pop fans are more likely to support the traditional view that the man should be the family’s main breadwinner. The effect found here is stronger in magnitude compared to the one in the previous regression. K-pop fan length is positive and significant; however, its coefficient is rather low, indicating that the length of being a fan does have, at most, a limited effect on gender attitudes.

Finally, results of regressing statement C—“It is natural for men to occupy leading positions at work”—are shown in column (3) of Table 4. Findings indicate that individuals with higher K-pop fan expenses are more likely to support gender segregation at work. Hence, K-pop fans appear to be more supportive of a gendered division of labor also within the work sphere.

With regard to the effects of demographic characteristics, the results from Table 5 are largely consistent with the findings in Table 4. Interestingly, while females do not differ from males in their attitudes towards housekeeping and childcare (statement A), they are more likely to disagree with men being the main breadwinner (statement B) and with gender segregation at work (statement C).