1. Literature review

Scholars have long been interested in what factors predict public attitudes toward homosexuality, with personal religious beliefs receiving particular attention. Because most religions consider heterosexual intercourse as normal, homosexuality is often considered as abnormal, unnatural, blasphemous, and ‘impure’ (Yip, Reference Yip2005; Adamczyk and Pitt, Reference Adamczyk and Pitt2009). Active religious involvement and the fear of divine punishment further lead faithful followers to adopt anti-homosexual attitudes and pursue intolerant institute policies toward same-sex behaviour (Regnerus and Smith, Reference Regnerus and Smith1998; Burdette et al., Reference Burdette, Ellison and Hill2005; Wilcox and Carin, Reference Wilcox and Carin2010). Van den Akker et al. (Reference van den Akker, van der Ploeg and Scheepers2013) examine anti-homosexuality attitudes by studying four waves of the European Social Survey. Their findings show that among all religious people, Muslims disapprove of homosexuality the most, followed by Protestants, Eastern Orthodox, other non-Christians, and other Christians. In contrast with other religious and non-religious people, Jews and Catholics disapprove of homosexuality the least. Similarly, in Latin America, Evangelicals (mainline Protestant) and the Catholic Church have joined forces against the LGBT movement (Lodola and Corral, Reference Lodola and Corral2010).

Recent studies suggest that, based on their understandings of Christian values, social change, and interpretations of the Bible, Christians often view LGBT individuals as violating moral sexual values (Haddock et al.,Reference Haddock, Zanna and Esses1993; Riek et al., Reference Riek, Mania and Gaertner2006) and doctrines (Herek, Reference Herek1987; Rodriguez, Reference Rodriguez2010; Subhi and Geelan, Reference Subhi and Geelan2012; Marinaccio, Reference Marinaccio2016). They tend to regard LGBT individuals as symbolic threats and thus perceive themselves to be in zero-sum competition with homosexuality (Wilkins et al., Reference Wilkins, Wellman, Toosi, Miller, Lisnek and Martin2022). Such zero-sum beliefs often drive Christians to counter perceived LGBT threats in order to maintain their social position (Wilkins et al., Reference Wilkins, Wellman, Toosi, Miller, Lisnek and Martin2022, p.3). For example, prior to the enactment of the controversial 2014 Anti-Homosexuality Act in Uganda, scholars observed that American Christian missionaries not only instructed their Ugandan partners on why and how to frame opposition to homosexuality as a political and social issue but also provided them with financial assistance and ideological support (Sharlet, Reference Sharlet2010; Kron, Reference Kron2012; Boyd, Reference Boyd2013). Similarly, in Taiwan, scholars often attribute anti-LGBT movements to Christian churches. Most anti-homosexuality organizations, such as the Taiwan League for Preserving Family, the Family Guardian Coalition, and the Faith and Hope League, were established by Catholic and Protestant conservatives (Ho, Reference Ho, Chiavacci, Grano and Obinger2020). Scholars have further identified strong connections between foreign anti-LGBT religious groups – such as MassResistance, the International House of Prayer, Bread of Life Christian Church, and Agape Christian Church – and local religious conservatives (Cole, Reference Cole2013; Reference Cole2017; Ho, Reference Ho2019; Reference Ho, Chiavacci, Grano and Obinger2020; Yang and Kang, Reference Yang, Kang, Devries, Bessant and Watts2021, p. 65).

Other scholars are interested in the role that culture has in shaping public values, attitudes, thoughts, and subsequent behaviours toward anti-homosexuality. For instance, Dunbar et al. (Reference Dunbar, Brown and Amoroso1973) find that anti-homosexual attitudes are positively correlated with general sexual conservatism. Brooks (Reference Brooks2000) also suggests that the attitudes toward liberalization of civil liberty will gradually be extended to gays and lesbians. Similarly, Anderson and Fetner (Reference Andersen and Fetter2008) investigate the WVS data from 1981 to 2000 and attribute the increasing tolerant attitude among Canadians toward homosexuality to the rapid changes in culture, economy, and politics and the expansion of lesbian and gay subcultures in media. Studies of Asians’ conservative attitudes toward homosexuality have also referred to traditional Asian values. Cheng et al. (Reference Cheng, Wu and Adamczyk2016) suggest that because of its emphasis on family values and kinship networks, Confucianism in East Asia – similar to Christianity and Islam in other regions – plays a distinctive role in shaping public attitudes toward same-sex marriage.

The studies above provide important theoretical insights into how religious and cultural factors shape individuals’ anti-homosexual attitudes. Nonetheless, it is clear that these studies cannot fully explain the sudden increase in anti-LGBT sentiment in Taiwan, where Christians account for only 7.62% of the population (Fang et al., Reference Fang, Kaplan, Chiang and Cheng2020, p. 8). Previous studies consistently show that over the course of nearly a quarter-century, societal attitudes and public opinion in Taiwan have shifted markedly in favour of same-sex marriage, making it one of the most accepting countries of same-sex marriage in Asia. Table 1 further draws on successive waves of the TSCS to examine changes in attitudes toward same-sex marriage across different religious groups. Among Buddhists and adherents of folk religions – who are deeply influenced by Confucianism and traditional Chinese culture and are generally more conservative – only about 12% supported same-sex marriage in 1991. By 2015, however, this figure had climbed to over 50%, far exceeding the 33% of Christians who expressed support. Yet, following the anti-same-sex marriage referendum, the proportion supporting the justification of homosexuality dropped back to around 10%. If Confucianism were indeed the decisive and consistent factor shaping attitudes toward LGBT rights, such a dramatic rise in acceptance between 1991 and 2015 would not have occurred. In other words, we may reasonably infer that anti-LGBT forces effectively reactivated latent Confucian values during the 2018 referendum campaign.

Table 1.

Changes in Taiwanese people’s attitudes toward same-sex marriage (by religions)

Source: TSCS 1991, 2015, and WVS 2019.

¹ Responses of 1–3 to the question of whether homosexuality is justifiable are treated as agreement with the justification of homosexuality.

Even if we acknowledge the importance of Confucian traditions in shaping individual anti-homosexual attitudes, it remains unclear how these influences operate. Confucianism undoubtedly encompasses a broad range of topics, including humanity, morality, governance, filial duties, and manners, but it makes no explicit reference to sexuality. Which components, then, did religious conservatives draw upon to mobilize public support for their anti-LGBT appeals? Family is indeed one of the most important tenets of Confucianism, and in the WVS, 90% of Taiwanese respondents indicated that family is very important in their lives. However, our preliminary analysis shows that such a normative view has no significant association with public attitudes toward homosexuality.Footnote ¹ These findings call for a reconsideration of the literature on the LGBT movement in Taiwan and a reinvestigation of the strategies employed by anti-LGBT activists during the referendum campaign.

This paper sets out to explore two central questions: First, how have traditional values and beliefs rooted in Confucian ideology influenced people’s attitudes toward homosexuality? Second, how have Christian organizations managed to overcome the disadvantage of being a religious minority and successfully mobilize Taiwanese citizens to vote in favour of their preferred referendum outcomes? The following section outlines the processes and strategies through which anti-LGBT activists gained recognition and electoral support among the Taiwanese public. We argue that rigid norms of gender roles and the emphasis on marriage for reproduction constituted the core Confucian concepts associated with anti-homosexual attitudes. Moreover, by suggesting same-sex marriage as a threat to social order, anti-same-sex marriage groups allied with the conservative Kuomintang (KMT) and effectively divided DPP voters. Section three presents the post-referendum WVS data and the variables used in the analysis, along with the methodological approach. Section four reports the results of the empirical analysis of the relationship between Taiwanese attitudes toward homosexuality and the explanatory variables. Finally, section five discusses the findings and outlines possible avenues for advancing LGBT rights.

2. Background

Taiwan’s same-sex marriage movement can be broadly divided into three stages: the stage of isolated incidents without organized mobilization (before 2006), the parliamentary stage (2006–2012), and the mainstream stage (after 2013) (Chien, Reference Chien2012; Lee and Lin, Reference Lee and Lin2022). According to Ho (Reference Ho2019, p. 148), prior to 2006, same-sex marriage was not a salient political issue, as religious conservatives did not regard LGBT activists as a force requiring attention, nor did the LGBT community consider the early 21st century an appropriate time to put marriage equality on the political agenda. In 2006, the first same-sex marriage bill, proposed by DPP legislator Bi-khim Hsiao, formally triggered the conflict between religious groups and the LGBT community. Although the proposed bill was eventually voted down by the conservative KMT, it struck a deep nerve among religious conservatives. In March 2007, Catholic and Protestant leaders established the first anti-LGBT coalition, the Taiwan League for Preserving Family, to link churches together in opposing same-sex marriage and preserving traditional family values. To garner pro-LGBT support from political parties, legislators, and the broader public, the LGBT community also founded the Taiwan Alliance to Promote Civil Partnership Rights (TAPCPR). In 2012, TAPCPR proposed the ‘Diverse Family Formation Act’, which contained three major amendments to the Civil Code: marriage equality, contractual civil partnerships, and multiple-person families (Ho, Reference Ho2019). Although only the same-sex marriage bill garnered sufficient endorsements and successfully passed the first reading in the Legislative Yuan on October 25, 2013, it nevertheless marked a significant achievement for Taiwan’s LGBT movement.

This breakthrough attempt to legalize same-sex marriage unsettled conservatives. On November 30, 2013, under the banner of ‘protecting the next generation’s well-being’, the Family Guardian Coalition (FGC) organized a massive rally opposing same-sex marriage and the Diverse Family Formation Act, successfully attracting around 300,000 participants. The conservative rally, however, did not prevent the LGBT movement from advancing. The TAPCPR actively sought potential partners and collaborated with women’s rights organizations, human rights organizations, and sympathetic political parties and legislators (Pettinicchio, Reference Pettinicchio2012, p. 537; Krumbein, Reference Krumbein2020). The Sunflower Student Movement, in which individuals under 30 made up 77.2% of all participants, further provided the LGBT community with new momentum (Hsu, Reference Hsu2019). In addition to mobilizing the younger generation, LGBT activists also allied with two newly established political parties that emerged from post-Sunflower activism: the Social Democratic Party (SDP) and the New Power Party (NPP). More importantly, in October 2015, DPP presidential candidate Ing-wen Tsai officially announced, ‘I am Ing-Wen Tsai, and I support marriage equality’ (Jeffreys and Wang, Reference Jeffreys, Wang, Winter, Forest and Sénac2018, p. 218). The assertion made her the first Asian political leader to publicly declare support for same-sex marriage. After her inauguration in October 2016, President Tsai exercised her constitutional power to nominate seven Grand Justices of the Constitutional Court. Six of these justices officially supported the legalization of same-sex marriage in the process of reviewing the qualifications for the nomination in the Legislative Yuan. Shortly after these Grand Justices assumed office in November 2016, the Constitutional Court was in charge of reviewing the application of the Constitutional Interpretation of same-sex marriage on February 20, 2017. On May 24, 2017, the Constitutional Court issued an 11–3 decision in favour of same-sex marriage, holding that people in Taiwan had a fundamental right to choose their own spouse without excessive government restriction. In so doing, it struck down the provisions of the Civil Code as unconstitutional, making Taiwan the first Asian country to legalize same-sex marriage.

Interpretation No. 748, as it is known, did not ease hostilities between anti- and pro-LGBT forces. Instead, it triggered an all-out confrontation between the two camps. Despite its explicit recognition of the right to same-sex marriage, the Interpretation instructed the Legislative Yuan to either amend the unconstitutional provisions of the Civil Code or enact new legislation consistent with its ruling within two years. While both approaches were intended to guarantee the LGBT community equal rights to form a permanent union of an intimate and exclusive nature, the former would recognize homosexual couples in the same way as heterosexual partners, granting them the right to marriage. By contrast, the latter would continue to define marriage solely as a union between a man and a woman, while establishing a separate civil union – a marriage-like legal status – exclusively for same-sex couples. Acknowledging the difficulty of challenging the Constitutional Court and its ruling (Ho and Huang, Reference Ho, Huang, Schubert and Lee2021), anti-LGBT forces instead sought to appeal directly to the public.

Before 2018, because referenda required both more than 50% voter turnout and approval by a majority of valid ballots, none of the six proposals introduced since the Referendum Act was enacted in 2003 had passed. In December 2017, however, the Act was amended, drastically lowering the threshold: a proposal would pass if the number of valid ‘yes’ votes exceeded the number of ‘no’ votes and if the ‘yes’ votes constituted at least one-quarter of all eligible voters. This reform not only transformed referenda into a new tool for mobilizing supporters but also provided conservative anti-LGBT groups with a powerful weapon to challenge Interpretation No. 748. On January 24, 2018, the Coalition for the Happiness of Our Next Generation (CHONG) proposed referendums on three anti-LGBT issues: overturning Interpretation No. 748, excluding same-sex couples from civil marriage, and prohibiting gender equality education in elementary and middle schools. Given the Tsai administration’s low approval ratings, anti-LGBT groups further aligned themselves with the KMT’s local election campaigns and four KMT-related referenda, including measures on reducing air pollution, opposing coal-fired power plants, restricting food imports from Fukushima, and repealing the provision mandating the phase-out of nuclear power plants.

To counter anti-LGBT groups, the pro-LGBT community launched two referenda aimed at safeguarding the rights to same-sex marriage and gender equality education. At the same time, pro-independence advocates also introduced a referendum calling for Taiwan’s national team to compete in the Tokyo Olympics under the name ‘Taiwan’. The ten referendum proposals, as shown in Table 2, were held in conjunction with the nine-in-one local elections on November 24, 2018, effectively turning the election into a full-scale ‘referendum war’. Consequently, anti-LGBT activists were more successful than their opponents in mobilizing supporters across partisan divides. All three anti-LGBT and gender equality education referenda not only passed the validation threshold by majority vote – thereby constraining the DPP government’s policies on same-sex marriage and gender equality education – but also led to the rejection of the two gender equality referenda by nearly seven million voters, reflecting prevailing views against homosexuality.

Table 2.

Ten referenda in the 2018 Taiwanese referendum

Source: Author’s compilation.

3. Strategies of the anti-LGBT groups

In Taiwan, churches – like their counterparts elsewhere – were on the front lines in the fight against same-sex marriage. However, as mentioned above, Christians account for only 7.62% of the country’s total population (Fang et al., Reference Fang, Kaplan, Chiang and Cheng2020), far below the threshold required to pass a referendum. To evoke public sympathy from non-Christians and persuade them to vote against same-sex marriage, Christian conservatives allied with other conservative religious groups, such as the Unification Church, Taoists, and Yiguandao, as well as minor political parties, forming the League of Taiwanese Religious Groups Caring for Family in 2013. To further attract support from adherents of Chinese folk religions, who comprise nearly 50% of the population, religious conservatives emphasized traditional Confucian gender role distinctions.

4. Data and variables

To investigate whether and how gender role distinctions, concerns about the next generation, and anti-DPP sentiments correlate with the surge in discrimination against homosexuality, this study applies a multivariate regression model to data from the Taiwan WVS Wave 7. The Taiwanese WVS survey was conducted in March 2019, four months after the 2018 local elections and referendum. It included 1223 respondents randomly drawn from the Taiwanese population and provides an accurate reflection of public attitudes toward homosexuality during the referendum controversy (see Appendix for descriptive statistics).

5. Analysis and results

Figure 1 presents a coefficient plot of the OLS estimates, using the justification of homosexuality as the dependent variable. We begin by examining how gender stereotypes are associated with public attitudes toward homosexuality. As illustrated in Figure 1, GNI is positively associated with opposition to LGBT rights: the stronger the belief in traditional gender distinctions, the lower the support for justifying homosexuality. Our findings align with recent scholarship on public attitudes toward homosexuality, suggesting that conceptions of gender roles are a significant determinant in shaping individuals’ acceptance of homosexual behaviour.

Figure 1.

Regression results of the public attitudes toward homosexuality in Taiwan.

Note: The figure presents the OLS estimates obtained from our model. The x-axis displays the estimated coefficients (blue dots) with 95% confidence interval of corresponding variables on y-axis.

We next examine the relationship between family values and attitudes toward the legitimacy of same-sex marriage. As noted earlier, when controlling for other variables, respondents’ emphasis on family values is not significantly associated with attitudes toward homosexuality. This finding is consistent with Chiang and Su (Reference Chiang and Su2024), who argue that both anti- and pro-LGBT groups employed family values as central campaign strategies during the 2018 referenda. To provide a more nuanced analysis, we further consider two specific dimensions of traditional family values: reproduction and filial duty. These variables capture the core expectations of family life and allow us to assess how they relate to our research question. The results show that the more strongly individuals endorsed having children and caring for the elderly as social obligations, the more likely they were to regard homosexuality as unjustified.

Why do the notions of filial piety and procreation shape attitudes toward homosexuality? Yeh and Bedford conceptualize filial piety in Confucian thought as a reciprocal social norm that prescribes mutual obligations between parents and children: parents are expected to provide appropriate care for their children during childhood, while children, in turn, are obligated to ensure the well-being of their parents in old age. Within this framework, the failure to produce descendants not only disrupts this reciprocal mechanism but also jeopardizes the continuity of the family lineage and is therefore regarded in Confucian tradition as one of the gravest forms of unfilial conduct. From this perspective, because same-sex couples are unable to produce biological offspring, they are perceived as violating the filial duty of lineage continuation, while simultaneously undermining the gender roles that sustain the Confucian family system. Combined with the emphasis on patrilineal succession in East Asian societies, this renders sexual diversity particularly vulnerable to cultural stigmatization (Adamczyk and Cheng, Reference Adamczyk and Cheng2015; Liao, Reference Liao2020).

We further investigate the relationship between partisan affiliation and attitudes toward homosexuality by estimating interaction effects between party identification and respondents’ emphasis on social order. Specifically, given our expectation that, unlike the KMT which adopted a consistently oppositional stance toward same-sex marriage – the DPP exhibited internal heterogeneity on this issue and may therefore be more responsive to anti-LGBT appeals grounded in social order, we categorize respondents into three groups: KMT supporters, DPP supporters, and independents/minor party supporters. We then construct multiplicative interaction terms by multiplying respondents’ party affiliation with their emphasis on social order. The model is estimated using ordinary least squares (OLS) regression with robust standard errors and controls for key sociodemographic covariates. Following Brambor, Clark, and Golder, we note that interpreting such models requires attention to the marginal effects rather than simply the statistical significance of coefficients. Accordingly, we compute predicted values from the regression results reported in Figure 1 and present them graphically in Figure 2, which allows us to assess how the effect of emphasizing social order on attitudes toward homosexuality varies across partisan groups.

Figure 2.

The interactive effects of partisanship and emphasizing social order on homosexual attitudes.

Note: The figure presents the OLS estimates derived from the multiplicative interaction terms in our model. The x-axis displays the estimated coefficients with 95% confidence interval, where red dots represent partisans who prioritized social order and blue dots represent those who did not. The y-axis indicates the corresponding categories of partisan affiliation.

Figure 2 shows that, controlling for other factors, KMT supporters display the highest level of opposition to homosexuality ( ${{\hat Y}_{KMT}} = 6.9$ , s.e = 0.12), followed by independents and minor-party supporters ( ${{\hat Y}_{Nonpartisan}} = 6.67$ ，s.e = 0.11)，with DPP supporters exhibiting the lowest level ( ${{\hat Y}_{DPP}} = 6.23$ , s.e = 0.14). However, when we further examine how the emphasis on social order conditions partisan differences in homosexual attitudes, the patterns diverge sharply. KMT supporters and non-partisan voters display consistent attitudes regardless of whether they emphasize social order, with no statistically significant differences between the two groups. By contrast, substantial within-party heterogeneity emerges among DPP supporters. As illustrated in Figure 2, holding all else constant, the predicted mean level of homosexual prejudice among liberal DPP supporters who do not emphasize social order ( ${{\hat Y}_{DP{P_{Lib}}}}$ ) is 5.70, whereas for those who place a high value on social order ( ${{\hat Y}_{DP{P_{con}}}}$ ) it rises to 6.44, substantially higher than the liberal DPPs ( ${{\hat Y}_{DP{P_{con}}}} - {{\hat Y}_{DP{P_{Lib}}}}$ 0.61, $s.e = 0.37$ ) and is comparable to the levels observed among KMT supporters as well as independents and minor-party voters. In other words, the finding above demonstrates how anti-LGBT groups not only rode the anti-DPP tide by convincing nonpartisans and KMT supporters to adopt their position on the referenda but also how they successfully split DPP supporters into pro- and anti-LGBT groups by highlighting the impact of same-sex marriage on social order.

In addition, the associations between demographic factors and homosexual attitudes derived from Figure 1 are similar to the findings of previous studies. The younger, female, and highly educated were more likely to hold liberal attitudes toward the LGBT community than people who were older, male, and less educated. As for religious beliefs, our finding, as expected, is also in alignment with previous studies, revealing that Christianity played a powerful role in shaping attitudes toward homosexuality. In Taiwan, Christians are disproportionately well-educated professionals, for example white-collar employees in the high-tech industry and government or military officials who enjoy high social status (Zhai and Woodberry, Reference Zhai and Woodberry2011). These high-status Christians tend to endorse zero-sum benefits and perceive same-sex marriage legalization as blasphemous to their religious beliefs and a tremendous loss for religious minorities like them (Wilkins et al., Reference Wilkins, Wellman, Toosi, Miller, Lisnek and Martin2022). Therefore, they waged the referenda war against the LGBT communities in hopes of overturning the Constitutional Court’s decision by appealing to public opinion. In addition to Christians, Taiwanese folk believers, who account for a majority of the population, also tended to be less tolerant of homosexuality than followers of other religions besides Christianity. This shows that Christian churches had successfully aroused Taiwanese people’s antipathy toward homosexuality and convinced them to vote against same-sex marriage and gender equality education in the referendum campaign.